Transfer Learning of resnet#
Notebook by:
Royi Avital RoyiAvital@fixelalgorithms.com
Revision History#
Version |
Date |
User |
Content / Changes |
|---|---|---|---|
1.0.000 |
29/05/2024 |
Royi Avital |
First version |
# Import Packages
# General Tools
import numpy as np
import scipy as sp
import pandas as pd
# Machine Learning
from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay
from sklearn.model_selection import ParameterGrid
# Deep Learning
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
from torch.optim.lr_scheduler import LRScheduler
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter
import torchinfo
from torchmetrics.classification import MulticlassAccuracy
import torchvision
from torchvision.transforms import v2 as TorchVisionTrns
# Miscellaneous
import copy
from enum import auto, Enum, unique
import math
import os
from platform import python_version
import random
import shutil
import time
# Typing
from typing import Callable, Dict, Generator, List, Optional, Self, Set, Tuple, Union
# Visualization
import matplotlib as mpl
import matplotlib.pyplot as plt
import seaborn as sns
# Jupyter
from IPython import get_ipython
from IPython.display import HTML, Image
from IPython.display import display
from ipywidgets import Dropdown, FloatSlider, interact, IntSlider, Layout, SelectionSlider
from ipywidgets import interact
2024-06-14 01:49:13.932980: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: SSE4.1 SSE4.2 AVX AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
Notations#
(?) Question to answer interactively.
(!) Simple task to add code for the notebook.
(@) Optional / Extra self practice.
(#) Note / Useful resource / Food for thought.
Code Notations:
someVar = 2; #<! Notation for a variable
vVector = np.random.rand(4) #<! Notation for 1D array
mMatrix = np.random.rand(4, 3) #<! Notation for 2D array
tTensor = np.random.rand(4, 3, 2, 3) #<! Notation for nD array (Tensor)
tuTuple = (1, 2, 3) #<! Notation for a tuple
lList = [1, 2, 3] #<! Notation for a list
dDict = {1: 3, 2: 2, 3: 1} #<! Notation for a dictionary
oObj = MyClass() #<! Notation for an object
dfData = pd.DataFrame() #<! Notation for a data frame
dsData = pd.Series() #<! Notation for a series
hObj = plt.Axes() #<! Notation for an object / handler / function handler
Code Exercise#
Single line fill
vallToFill = ???
Multi Line to Fill (At least one)
# You need to start writing
????
Section to Fill
#===========================Fill This===========================#
# 1. Explanation about what to do.
# !! Remarks to follow / take under consideration.
mX = ???
???
#===============================================================#
# Configuration
# %matplotlib inline
seedNum = 512
np.random.seed(seedNum)
random.seed(seedNum)
# Matplotlib default color palette
lMatPltLibclr = ['#1f77b4', '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2', '#7f7f7f', '#bcbd22', '#17becf']
# sns.set_theme() #>! Apply SeaBorn theme
runInGoogleColab = 'google.colab' in str(get_ipython())
# Improve performance by benchmarking
torch.backends.cudnn.benchmark = True
# Reproducibility (Per PyTorch Version on the same device)
# torch.manual_seed(seedNum)
# torch.backends.cudnn.deterministic = True
# torch.backends.cudnn.benchmark = False #<! Makes things slower
# Constants
FIG_SIZE_DEF = (8, 8)
ELM_SIZE_DEF = 50
CLASS_COLOR = ('b', 'r')
EDGE_COLOR = 'k'
MARKER_SIZE_DEF = 10
LINE_WIDTH_DEF = 2
DATA_SET_FILE_NAME = 'archive.zip'
DATA_SET_FOLDER_NAME = 'IntelImgCls'
D_CLASSES = {0: 'Buildings', 1: 'Forest', 2: 'Glacier', 3: 'Mountain', 4: 'Sea', 5: 'Street'}
L_CLASSES = ['Buildings', 'Forest', 'Glacier', 'Mountain', 'Sea', 'Street']
T_IMG_SIZE = (150, 150, 3)
DATA_FOLDER_PATH = 'Data'
TENSOR_BOARD_BASE = 'TB'
# Download Auxiliary Modules for Google Colab
if runInGoogleColab:
!wget https://raw.githubusercontent.com/FixelAlgorithmsTeam/FixelCourses/master/AIProgram/2024_02/DataManipulation.py
!wget https://raw.githubusercontent.com/FixelAlgorithmsTeam/FixelCourses/master/AIProgram/2024_02/DataVisualization.py
!wget https://raw.githubusercontent.com/FixelAlgorithmsTeam/FixelCourses/master/AIProgram/2024_02/DeepLearningPyTorch.py
# Courses Packages
import sys
sys.path.append('/home/vlad/utils')
from DataVisualization import PlotLabelsHistogram
from DeepLearningPyTorch import TBLogger, TestDataSet
from DeepLearningPyTorch import TrainModel, TrainModelSch
(!) Go through
TestDataSet’s code.
# General Auxiliary Functions
def GenResNetModel( trainedModel: bool, numCls: int, resNetDepth: int = 18 ) -> nn.Module:
# Read on the API change at: How to Train State of the Art Models Using TorchVision’s Latest Primitives
# https://pytorch.org/blog/how-to-train-state-of-the-art-models-using-torchvision-latest-primitives
if (resNetDepth == 18):
modelFun = torchvision.models.resnet18
modelWeights = torchvision.models.ResNet18_Weights.IMAGENET1K_V1
elif (resNetDepth == 34):
modelFun = torchvision.models.resnet34
modelWeights = torchvision.models.ResNet34_Weights.IMAGENET1K_V1
elif (resNetDepth == 50):
modelFun = torchvision.models.resnet50
modelWeights = torchvision.models.ResNet50_Weights.IMAGENET1K_V1
else:
raise ValueError(f'The `resNetDepth`: {resNetDepth} is invalid!')
if trainedModel:
oModel = modelFun(weights = modelWeights)
numFeaturesIn = oModel.fc.in_features
# Assuming numCls << 100
oModel.fc = nn.Sequential(
nn.Linear(numFeaturesIn, 128), nn.ReLU(),
nn.Linear(128, numCls),
)
else:
oModel = modelFun(weights = None, num_classes = numCls)
return oModel
Transfer Learning#
The ResNet model is considered to be one of the most successful architectures.
Its main novelty is the Skip Connection which improved the performance greatly.
By hand waiving the contribution of the skip connection can be explained as:
Ensemble of model.
Skip vanishing
This notebook presents the basics of Transfer Learning in the context of image classification:
Loading a pretrained model on a classification task.
Adjusting its structure to the new classification task.
Finetuning the model.
Evaluating the model.
(#) A great recap on Model Fine Tuning is given in the book Dive into Deep Learning: Computer Vision - Fine Tuning.
# Parameters
# Data
# Model
dropP = 0.5 #<! Dropout Layer
# Training
batchSize = 128
numWorkers = 4 #<! Number of workers
numEpochs = 10
# Visualization
numImg = 3
Generate / Load Data#
This notebook use the Intel Image Classification Data Set.
The data set is composed of 6 classes: Buildings, Forest, Glacier, Mountain, Sea, Street.
Download the Zip file
archive.zipfrom Intel Image Classification Data Set.Copy / Move the file into
AIProgram/<YYYY_MM>/Datafolder.
The following code will arrange the data in a manner compatible with PyTorch’s ImageFolder.
(#) The data set originally appeared on Analytics Vidhya - Practice Problem: Intel Scene Classification Challenge.
(#) Some of the images are not
150x150x3hence they should be handled.(#) Some of the images are not labeled correctly (See discussions on Kaggle).
# Arrange Data for Image Folder
# Assumes `archive.zip` in `./Data`
dataSetPath = os.path.join(DATA_FOLDER_PATH, DATA_SET_FOLDER_NAME)
if not os.path.isdir(dataSetPath):
os.mkdir(dataSetPath)
lFiles = os.listdir(dataSetPath)
if '.processed' not in lFiles: #<! Run only once
os.makedirs(os.path.join(dataSetPath, 'TMP'), exist_ok = True)
os.makedirs(os.path.join(dataSetPath, 'Test'), exist_ok = True)
for clsName in L_CLASSES:
os.makedirs(os.path.join(dataSetPath, 'Train', clsName), exist_ok = True)
os.makedirs(os.path.join(dataSetPath, 'Validation', clsName), exist_ok = True)
shutil.unpack_archive(os.path.join(DATA_FOLDER_PATH, DATA_SET_FILE_NAME), os.path.join(dataSetPath, 'TMP'))
for dirPath, lSubDir, lF in os.walk(os.path.join(dataSetPath, 'TMP')):
if len(lF) > 0:
if 'test' in dirPath:
dstPath = os.path.join(dataSetPath, 'Validation')
elif 'train' in dirPath:
dstPath = os.path.join(dataSetPath, 'Train')
else:
dstPath = os.path.join(dataSetPath, 'Test')
if 'buildings' in dirPath:
for fileName in lF:
shutil.move(os.path.join(dirPath, fileName), os.path.join(dstPath, 'Buildings'))
elif 'forest' in dirPath:
for fileName in lF:
shutil.move(os.path.join(dirPath, fileName), os.path.join(dstPath, 'Forest'))
elif 'glacier' in dirPath:
for fileName in lF:
shutil.move(os.path.join(dirPath, fileName), os.path.join(dstPath, 'Glacier'))
elif 'mountain' in dirPath:
for fileName in lF:
shutil.move(os.path.join(dirPath, fileName), os.path.join(dstPath, 'Mountain'))
elif 'sea' in dirPath:
for fileName in lF:
shutil.move(os.path.join(dirPath, fileName), os.path.join(dstPath, 'Sea'))
elif 'street' in dirPath:
for fileName in lF:
shutil.move(os.path.join(dirPath, fileName), os.path.join(dstPath, 'Street'))
else:
for fileName in lF:
shutil.move(os.path.join(dirPath, fileName), dstPath)
shutil.rmtree(os.path.join(dataSetPath, 'TMP'))
hFile = open(os.path.join(dataSetPath, '.processed'), 'w')
hFile.close()
# Load Data
dsTrain = torchvision.datasets.ImageFolder(os.path.join(DATA_FOLDER_PATH, DATA_SET_FOLDER_NAME, 'Train'), transform = torchvision.transforms.ToTensor())
dsVal = torchvision.datasets.ImageFolder(os.path.join(DATA_FOLDER_PATH, DATA_SET_FOLDER_NAME, 'Validation'), transform = torchvision.transforms.ToTensor())
dsTest = TestDataSet(os.path.join(DATA_FOLDER_PATH, DATA_SET_FOLDER_NAME, 'Test'), transform = torchvision.transforms.ToTensor()) #<! Does not return a label
lClass = dsTrain.classes
numSamples = len(dsTrain)
print(f'The data set number of samples (Train): {numSamples}')
print(f'The data set number of samples (Validation): {len(dsVal)}')
print(f'The data set number of samples (Test): {len(dsTest)}')
print(f'The unique values of the labels: {np.unique(lClass)}')
The data set number of samples (Train): 14034
The data set number of samples (Validation): 3000
The data set number of samples (Test): 7301
The unique values of the labels: ['Buildings' 'Forest' 'Glacier' 'Mountain' 'Sea' 'Street']
(#) The dataset is indexible (Subscriptable). It returns a tuple of the features and the label.
(#) While data is arranged as
H x W x Cthe transformer, when accessing the data, will convert it intoC x H x W.
# Element of the Data Set
mX, valY = dsTrain[0]
print(f'The features shape: {mX.shape}')
print(f'The label value: {valY}')
The features shape: torch.Size([3, 150, 150])
The label value: 0
Plot the Data#
# Plot Data
vIdx = np.random.choice(numSamples, size = 9)
hF, vHa = plt.subplots(nrows = 3, ncols = 3, figsize = (10, 10))
vHa = vHa.flat
for ii, hA in enumerate(vHa):
hA.imshow(dsTrain[vIdx[ii]][0].permute((1, 2, 0)).numpy())
hA.tick_params(axis = 'both', left = False, top = False, right = False, bottom = False,
labelleft = False, labeltop = False, labelright = False, labelbottom = False)
hA.grid(False)
hA.set_title(f'Index = {vIdx[ii]}, Label = {L_CLASSES[dsTrain[vIdx[ii]][1]]}')
plt.show()
(?) If data is converted into grayscale, how would it effect the performance of the classifier? Explain.
You may assume the conversion is done using the mean value of the RGB pixel.
Pre Process Data#
This section:
Normalizes the data in a predefined manner.
Takes a sub set of the data.
Since the model is “borrowed” by Transfer Learning one must:
Use the statistics from the original training set.
Adapt the input dimensions to match the original training set.
(#) The values in training are specified in documentation.
As an example, seeResNet50Weights.
# The Standardization Parameters
# ImageNet statistics
vMean = np.array([0.485, 0.456, 0.406])
vStd = np.array([0.229, 0.224, 0.225])
print('µ =', vMean)
print('σ =', vStd)
µ = [0.485 0.456 0.406]
σ = [0.229 0.224 0.225]
# Check Image Dimensions (Run Only Once)
# Verifies all images have the same size: 3 x 150 x 150.
# for ii in range(len(dsTrain)):
# xx, yy = dsTrain[ii]
# imgH = xx.shape[1]
# imgW = xx.shape[2]
# if ((imgH != 150) or (imgW != 150)):
# print(f'The image {dsTrain.imgs[ii][0]} has incorrect size')
# Update Transforms
# Using v2 Transforms.
# Taking care of the different dimensions of some images.
# Matching the input size of ImageNet.
oDataTrnsTrain = TorchVisionTrns.Compose([
TorchVisionTrns.ToImage(),
TorchVisionTrns.ToDtype(torch.float32, scale = True),
TorchVisionTrns.Resize(224),
TorchVisionTrns.CenterCrop(224), #<! Ensures size is 150 (Pads if needed)
## !!!!!!!!!!!!!!!!!!!
TorchVisionTrns.RandomHorizontalFlip(p = 0.5),
## !!!!!!!!!!!!!!!!!!!
TorchVisionTrns.Normalize(mean = vMean, std = vStd),
])
oDataTrnsVal = TorchVisionTrns.Compose([
TorchVisionTrns.ToImage(),
TorchVisionTrns.ToDtype(torch.float32, scale = True),
TorchVisionTrns.Resize(224),
TorchVisionTrns.CenterCrop(224), #<! Ensures size is 150 (Pads if needed)
TorchVisionTrns.Normalize(mean = vMean, std = vStd),
])
# Using V1
# oDataTrnsTrain = torchvision.transforms.Compose([
# torchvision.transforms.Resize(224),
# torchvision.transforms.CenterCrop(224),
# torchvision.transforms.RandomHorizontalFlip(0.5),
# torchvision.transforms.ToTensor(),
# torchvision.transforms.Normalize(mean = vMean, std = vStd),
# ])
# oDataTrnsVal = torchvision.transforms.Compose([
# torchvision.transforms.Resize(224),
# torchvision.transforms.CenterCrop(224),
# torchvision.transforms.ToTensor(),
# torchvision.transforms.Normalize(mean = vMean, std = vStd),
# ])
# Update the DS transformer
dsTrain.transform = oDataTrnsTrain
dsVal.transform = oDataTrnsVal
(?) What does
RandomHorizontalFlipdo? Why can it be used?
# "Normalized" Image
mX, valY = dsTrain[5]
hF, hA = plt.subplots()
hImg = hA.imshow(np.transpose(mX, (1, 2, 0)))
hF.colorbar(hImg)
plt.show()
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
(?) How can one get the original image from
mX?
inverse of the regulation ;
Data Loaders#
This section defines the data loaded.
# Data Loader
dlTrain = torch.utils.data.DataLoader(dsTrain, shuffle = True, batch_size = 1 * batchSize, num_workers = numWorkers, drop_last = True, persistent_workers = True)
dlVal = torch.utils.data.DataLoader(dsVal, shuffle = False, batch_size = 2 * batchSize, num_workers = numWorkers, persistent_workers = True)
(!) Plot the histogram of labels of the data. Is it balanced?
# Iterate on the Loader
# The first batch.
tX, vY = next(iter(dlTrain)) #<! PyTorch Tensors
print(f'The batch features dimensions: {tX.shape}')
print(f'The batch labels dimensions: {vY.shape}')
The batch features dimensions: torch.Size([128, 3, 224, 224])
The batch labels dimensions: torch.Size([128])
# Looping
for ii, (tX, vY) in zip(range(1), dlVal): #<! https://stackoverflow.com/questions/36106712
print(f'The batch features dimensions: {tX.shape}')
print(f'The batch labels dimensions: {vY.shape}')
The batch features dimensions: torch.Size([256, 3, 224, 224])
The batch labels dimensions: torch.Size([256])
Load the Model#
This section loads the model.
The number of outputs is adjusted to match the number of classes in the data.
# Loading a Pre Defined Model
oModelPreDef = GenResNetModel(trainedModel = False, numCls = len(L_CLASSES))
(!) Go through
GenResNetModel()’s code.
# Model Information - Pre Defined
# Pay attention to the layers name.
torchinfo.summary(oModelPreDef, tX.shape, col_names = ['kernel_size', 'output_size', 'num_params'], device = 'cpu', row_settings = ['depth', 'var_names'])
========================================================================================================================
Layer (type (var_name):depth-idx) Kernel Shape Output Shape Param #
========================================================================================================================
ResNet (ResNet) -- [256, 6] --
├─Conv2d (conv1): 1-1 [7, 7] [256, 64, 112, 112] 9,408
├─BatchNorm2d (bn1): 1-2 -- [256, 64, 112, 112] 128
├─ReLU (relu): 1-3 -- [256, 64, 112, 112] --
├─MaxPool2d (maxpool): 1-4 3 [256, 64, 56, 56] --
├─Sequential (layer1): 1-5 -- [256, 64, 56, 56] --
│ └─BasicBlock (0): 2-1 -- [256, 64, 56, 56] --
│ │ └─Conv2d (conv1): 3-1 [3, 3] [256, 64, 56, 56] 36,864
│ │ └─BatchNorm2d (bn1): 3-2 -- [256, 64, 56, 56] 128
│ │ └─ReLU (relu): 3-3 -- [256, 64, 56, 56] --
│ │ └─Conv2d (conv2): 3-4 [3, 3] [256, 64, 56, 56] 36,864
│ │ └─BatchNorm2d (bn2): 3-5 -- [256, 64, 56, 56] 128
│ │ └─ReLU (relu): 3-6 -- [256, 64, 56, 56] --
│ └─BasicBlock (1): 2-2 -- [256, 64, 56, 56] --
│ │ └─Conv2d (conv1): 3-7 [3, 3] [256, 64, 56, 56] 36,864
│ │ └─BatchNorm2d (bn1): 3-8 -- [256, 64, 56, 56] 128
│ │ └─ReLU (relu): 3-9 -- [256, 64, 56, 56] --
│ │ └─Conv2d (conv2): 3-10 [3, 3] [256, 64, 56, 56] 36,864
│ │ └─BatchNorm2d (bn2): 3-11 -- [256, 64, 56, 56] 128
│ │ └─ReLU (relu): 3-12 -- [256, 64, 56, 56] --
├─Sequential (layer2): 1-6 -- [256, 128, 28, 28] --
│ └─BasicBlock (0): 2-3 -- [256, 128, 28, 28] --
│ │ └─Conv2d (conv1): 3-13 [3, 3] [256, 128, 28, 28] 73,728
│ │ └─BatchNorm2d (bn1): 3-14 -- [256, 128, 28, 28] 256
│ │ └─ReLU (relu): 3-15 -- [256, 128, 28, 28] --
│ │ └─Conv2d (conv2): 3-16 [3, 3] [256, 128, 28, 28] 147,456
│ │ └─BatchNorm2d (bn2): 3-17 -- [256, 128, 28, 28] 256
│ │ └─Sequential (downsample): 3-18 -- [256, 128, 28, 28] 8,448
│ │ └─ReLU (relu): 3-19 -- [256, 128, 28, 28] --
│ └─BasicBlock (1): 2-4 -- [256, 128, 28, 28] --
│ │ └─Conv2d (conv1): 3-20 [3, 3] [256, 128, 28, 28] 147,456
│ │ └─BatchNorm2d (bn1): 3-21 -- [256, 128, 28, 28] 256
│ │ └─ReLU (relu): 3-22 -- [256, 128, 28, 28] --
│ │ └─Conv2d (conv2): 3-23 [3, 3] [256, 128, 28, 28] 147,456
│ │ └─BatchNorm2d (bn2): 3-24 -- [256, 128, 28, 28] 256
│ │ └─ReLU (relu): 3-25 -- [256, 128, 28, 28] --
├─Sequential (layer3): 1-7 -- [256, 256, 14, 14] --
│ └─BasicBlock (0): 2-5 -- [256, 256, 14, 14] --
│ │ └─Conv2d (conv1): 3-26 [3, 3] [256, 256, 14, 14] 294,912
│ │ └─BatchNorm2d (bn1): 3-27 -- [256, 256, 14, 14] 512
│ │ └─ReLU (relu): 3-28 -- [256, 256, 14, 14] --
│ │ └─Conv2d (conv2): 3-29 [3, 3] [256, 256, 14, 14] 589,824
│ │ └─BatchNorm2d (bn2): 3-30 -- [256, 256, 14, 14] 512
│ │ └─Sequential (downsample): 3-31 -- [256, 256, 14, 14] 33,280
│ │ └─ReLU (relu): 3-32 -- [256, 256, 14, 14] --
│ └─BasicBlock (1): 2-6 -- [256, 256, 14, 14] --
│ │ └─Conv2d (conv1): 3-33 [3, 3] [256, 256, 14, 14] 589,824
│ │ └─BatchNorm2d (bn1): 3-34 -- [256, 256, 14, 14] 512
│ │ └─ReLU (relu): 3-35 -- [256, 256, 14, 14] --
│ │ └─Conv2d (conv2): 3-36 [3, 3] [256, 256, 14, 14] 589,824
│ │ └─BatchNorm2d (bn2): 3-37 -- [256, 256, 14, 14] 512
│ │ └─ReLU (relu): 3-38 -- [256, 256, 14, 14] --
├─Sequential (layer4): 1-8 -- [256, 512, 7, 7] --
│ └─BasicBlock (0): 2-7 -- [256, 512, 7, 7] --
│ │ └─Conv2d (conv1): 3-39 [3, 3] [256, 512, 7, 7] 1,179,648
│ │ └─BatchNorm2d (bn1): 3-40 -- [256, 512, 7, 7] 1,024
│ │ └─ReLU (relu): 3-41 -- [256, 512, 7, 7] --
│ │ └─Conv2d (conv2): 3-42 [3, 3] [256, 512, 7, 7] 2,359,296
│ │ └─BatchNorm2d (bn2): 3-43 -- [256, 512, 7, 7] 1,024
│ │ └─Sequential (downsample): 3-44 -- [256, 512, 7, 7] 132,096
│ │ └─ReLU (relu): 3-45 -- [256, 512, 7, 7] --
│ └─BasicBlock (1): 2-8 -- [256, 512, 7, 7] --
│ │ └─Conv2d (conv1): 3-46 [3, 3] [256, 512, 7, 7] 2,359,296
│ │ └─BatchNorm2d (bn1): 3-47 -- [256, 512, 7, 7] 1,024
│ │ └─ReLU (relu): 3-48 -- [256, 512, 7, 7] --
│ │ └─Conv2d (conv2): 3-49 [3, 3] [256, 512, 7, 7] 2,359,296
│ │ └─BatchNorm2d (bn2): 3-50 -- [256, 512, 7, 7] 1,024
│ │ └─ReLU (relu): 3-51 -- [256, 512, 7, 7] --
├─AdaptiveAvgPool2d (avgpool): 1-9 -- [256, 512, 1, 1] --
├─Linear (fc): 1-10 -- [256, 6] 3,078
========================================================================================================================
Total params: 11,179,590
Trainable params: 11,179,590
Non-trainable params: 0
Total mult-adds (Units.GIGABYTES): 464.27
========================================================================================================================
Input size (MB): 154.14
Forward/backward pass size (MB): 10173.30
Params size (MB): 44.72
Estimated Total Size (MB): 10372.16
========================================================================================================================
row_settings = [‘depth’, ‘var_names’]) show tha name:
which can help do directly to layer by name:
oModelPreDef.maxpool
MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
list(oModelPreDef.children())
[Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False),
BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True),
ReLU(inplace=True),
MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False),
Sequential(
(0): BasicBlock(
(conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
(1): BasicBlock(
(conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
),
Sequential(
(0): BasicBlock(
(conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(downsample): Sequential(
(0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
),
Sequential(
(0): BasicBlock(
(conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(downsample): Sequential(
(0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
),
Sequential(
(0): BasicBlock(
(conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(downsample): Sequential(
(0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
),
AdaptiveAvgPool2d(output_size=(1, 1)),
Linear(in_features=512, out_features=6, bias=True)]
len(list(oModelPreDef.children()))
10
len(list(oModelPreDef.modules()))
68
(?) Which layer should be adapted?
changed only the FC;
(?) Does the last (Head) dense layer includes a bias? Explain.
├─AdaptiveAvgPool2d (avgpool): 1-9 -- [256, 512, 1, 1] --
├─Linear (fc): 1-10 -- [256, 6] 3,078
512 * 6 = 3072 + 6 = 3078 ; include bias
# Model Information - Pre Trained
# Pay attention to the layers name.
oModelPreTrn = GenResNetModel(trainedModel = True, numCls = len(L_CLASSES))
# Model Information
# Pay attention to the variable name
torchinfo.summary(oModelPreTrn, tX.shape, col_names = ['kernel_size', 'output_size', 'num_params'], device = 'cpu', row_settings = ['depth', 'var_names'])
========================================================================================================================
Layer (type (var_name):depth-idx) Kernel Shape Output Shape Param #
========================================================================================================================
ResNet (ResNet) -- [256, 6] --
├─Conv2d (conv1): 1-1 [7, 7] [256, 64, 112, 112] 9,408
├─BatchNorm2d (bn1): 1-2 -- [256, 64, 112, 112] 128
├─ReLU (relu): 1-3 -- [256, 64, 112, 112] --
├─MaxPool2d (maxpool): 1-4 3 [256, 64, 56, 56] --
├─Sequential (layer1): 1-5 -- [256, 64, 56, 56] --
│ └─BasicBlock (0): 2-1 -- [256, 64, 56, 56] --
│ │ └─Conv2d (conv1): 3-1 [3, 3] [256, 64, 56, 56] 36,864
│ │ └─BatchNorm2d (bn1): 3-2 -- [256, 64, 56, 56] 128
│ │ └─ReLU (relu): 3-3 -- [256, 64, 56, 56] --
│ │ └─Conv2d (conv2): 3-4 [3, 3] [256, 64, 56, 56] 36,864
│ │ └─BatchNorm2d (bn2): 3-5 -- [256, 64, 56, 56] 128
│ │ └─ReLU (relu): 3-6 -- [256, 64, 56, 56] --
│ └─BasicBlock (1): 2-2 -- [256, 64, 56, 56] --
│ │ └─Conv2d (conv1): 3-7 [3, 3] [256, 64, 56, 56] 36,864
│ │ └─BatchNorm2d (bn1): 3-8 -- [256, 64, 56, 56] 128
│ │ └─ReLU (relu): 3-9 -- [256, 64, 56, 56] --
│ │ └─Conv2d (conv2): 3-10 [3, 3] [256, 64, 56, 56] 36,864
│ │ └─BatchNorm2d (bn2): 3-11 -- [256, 64, 56, 56] 128
│ │ └─ReLU (relu): 3-12 -- [256, 64, 56, 56] --
├─Sequential (layer2): 1-6 -- [256, 128, 28, 28] --
│ └─BasicBlock (0): 2-3 -- [256, 128, 28, 28] --
│ │ └─Conv2d (conv1): 3-13 [3, 3] [256, 128, 28, 28] 73,728
│ │ └─BatchNorm2d (bn1): 3-14 -- [256, 128, 28, 28] 256
│ │ └─ReLU (relu): 3-15 -- [256, 128, 28, 28] --
│ │ └─Conv2d (conv2): 3-16 [3, 3] [256, 128, 28, 28] 147,456
│ │ └─BatchNorm2d (bn2): 3-17 -- [256, 128, 28, 28] 256
│ │ └─Sequential (downsample): 3-18 -- [256, 128, 28, 28] 8,448
│ │ └─ReLU (relu): 3-19 -- [256, 128, 28, 28] --
│ └─BasicBlock (1): 2-4 -- [256, 128, 28, 28] --
│ │ └─Conv2d (conv1): 3-20 [3, 3] [256, 128, 28, 28] 147,456
│ │ └─BatchNorm2d (bn1): 3-21 -- [256, 128, 28, 28] 256
│ │ └─ReLU (relu): 3-22 -- [256, 128, 28, 28] --
│ │ └─Conv2d (conv2): 3-23 [3, 3] [256, 128, 28, 28] 147,456
│ │ └─BatchNorm2d (bn2): 3-24 -- [256, 128, 28, 28] 256
│ │ └─ReLU (relu): 3-25 -- [256, 128, 28, 28] --
├─Sequential (layer3): 1-7 -- [256, 256, 14, 14] --
│ └─BasicBlock (0): 2-5 -- [256, 256, 14, 14] --
│ │ └─Conv2d (conv1): 3-26 [3, 3] [256, 256, 14, 14] 294,912
│ │ └─BatchNorm2d (bn1): 3-27 -- [256, 256, 14, 14] 512
│ │ └─ReLU (relu): 3-28 -- [256, 256, 14, 14] --
│ │ └─Conv2d (conv2): 3-29 [3, 3] [256, 256, 14, 14] 589,824
│ │ └─BatchNorm2d (bn2): 3-30 -- [256, 256, 14, 14] 512
│ │ └─Sequential (downsample): 3-31 -- [256, 256, 14, 14] 33,280
│ │ └─ReLU (relu): 3-32 -- [256, 256, 14, 14] --
│ └─BasicBlock (1): 2-6 -- [256, 256, 14, 14] --
│ │ └─Conv2d (conv1): 3-33 [3, 3] [256, 256, 14, 14] 589,824
│ │ └─BatchNorm2d (bn1): 3-34 -- [256, 256, 14, 14] 512
│ │ └─ReLU (relu): 3-35 -- [256, 256, 14, 14] --
│ │ └─Conv2d (conv2): 3-36 [3, 3] [256, 256, 14, 14] 589,824
│ │ └─BatchNorm2d (bn2): 3-37 -- [256, 256, 14, 14] 512
│ │ └─ReLU (relu): 3-38 -- [256, 256, 14, 14] --
├─Sequential (layer4): 1-8 -- [256, 512, 7, 7] --
│ └─BasicBlock (0): 2-7 -- [256, 512, 7, 7] --
│ │ └─Conv2d (conv1): 3-39 [3, 3] [256, 512, 7, 7] 1,179,648
│ │ └─BatchNorm2d (bn1): 3-40 -- [256, 512, 7, 7] 1,024
│ │ └─ReLU (relu): 3-41 -- [256, 512, 7, 7] --
│ │ └─Conv2d (conv2): 3-42 [3, 3] [256, 512, 7, 7] 2,359,296
│ │ └─BatchNorm2d (bn2): 3-43 -- [256, 512, 7, 7] 1,024
│ │ └─Sequential (downsample): 3-44 -- [256, 512, 7, 7] 132,096
│ │ └─ReLU (relu): 3-45 -- [256, 512, 7, 7] --
│ └─BasicBlock (1): 2-8 -- [256, 512, 7, 7] --
│ │ └─Conv2d (conv1): 3-46 [3, 3] [256, 512, 7, 7] 2,359,296
│ │ └─BatchNorm2d (bn1): 3-47 -- [256, 512, 7, 7] 1,024
│ │ └─ReLU (relu): 3-48 -- [256, 512, 7, 7] --
│ │ └─Conv2d (conv2): 3-49 [3, 3] [256, 512, 7, 7] 2,359,296
│ │ └─BatchNorm2d (bn2): 3-50 -- [256, 512, 7, 7] 1,024
│ │ └─ReLU (relu): 3-51 -- [256, 512, 7, 7] --
├─AdaptiveAvgPool2d (avgpool): 1-9 -- [256, 512, 1, 1] --
├─Sequential (fc): 1-10 -- [256, 6] --
│ └─Linear (0): 2-9 -- [256, 128] 65,664
│ └─ReLU (1): 2-10 -- [256, 128] --
│ └─Linear (2): 2-11 -- [256, 6] 774
========================================================================================================================
Total params: 11,242,950
Trainable params: 11,242,950
Non-trainable params: 0
Total mult-adds (Units.GIGABYTES): 464.29
========================================================================================================================
Input size (MB): 154.14
Forward/backward pass size (MB): 10173.56
Params size (MB): 44.97
Estimated Total Size (MB): 10372.67
========================================================================================================================
re train change our last layer to:
├─Sequential (fc): 1-10 -- [256, 6] --
│ └─Linear (0): 2-9 -- [256, 128] 65,664
│ └─ReLU (1): 2-10 -- [256, 128] --
│ └─Linear (2): 2-11
Train the Model#
This section trains the model.
It compares pre trained model with pre defined model using the same number of epochs.
Transfer Learning Fine Tuning#
The training of the model on the new data is often called fine tuning (See Fine Tuning vs. Transfer Learning vs. Learning from Scratch for a discussion on the semantic).
The concept is training the new layers of the model with the new data while keeping most of the “knowledge” of the model from its original training.
The balance is done by the adaptation of the learning per layer with the extreme of zero learning rate for some layers (Freezing).
The most used combinations are:
Freeze Layers
Freeze (Zero learning rate) the pre trained layers by disabling the gradient (requires_grad).Smaller Learning Rate
Set a smaller learning rate to the pre trained layers.Small Learning
Use small learning rate to the whole process.
In some cases, the policy used is a combination of 2 (Freeze at the beginning / end, the release, etc..).
(#) Freezing is also a regularization as its assists in preventing over fitting.
(#) See Dive into Deep Learning - Computer Vision - Fine Tuning.
(#) How to Freeze Model Weights in PyTorch for Transfer Learning: Step by Step Tutorial.
# Freeze Layers
# Iterating over the net, see https://stackoverflow.com/questions/54203451
for paramName, oPrm in oModelPreTrn.named_parameters():
if not ('fc' in paramName):
oPrm.requires_grad = False
(!) Exclude Batch Norm layers as well.
# Run Device
runDevice = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu') #<! The 1st CUDA device
# Models
lModel = [('Pre Defined Model', oModelPreDef), ('Pre Trained Model', oModelPreTrn)]
# Loss and Score Function
hL = nn.CrossEntropyLoss()
hS = MulticlassAccuracy(num_classes = len(lClass), average = 'micro')
hL = hL.to(runDevice) #<! Not required!
hS = hS.to(runDevice)
(#) The averaging mode
macroaverages samples per class and average the result of each class.(#) The averaging mode
microaverages all samples.(?) Check results with
average = 'micro'. Explain howshuffle - Falsein the validation data loader affects the results.
parameters#
list(oModelPreTrn.parameters())
[Parameter containing:
tensor([[[[-1.0419e-02, -6.1356e-03, -1.8098e-03, ..., 5.6615e-02,
1.7083e-02, -1.2694e-02],
[ 1.1083e-02, 9.5276e-03, -1.0993e-01, ..., -2.7124e-01,
-1.2907e-01, 3.7424e-03],
[-6.9434e-03, 5.9089e-02, 2.9548e-01, ..., 5.1972e-01,
2.5632e-01, 6.3573e-02],
...,
[-2.7535e-02, 1.6045e-02, 7.2595e-02, ..., -3.3285e-01,
-4.2058e-01, -2.5781e-01],
[ 3.0613e-02, 4.0960e-02, 6.2850e-02, ..., 4.1384e-01,
3.9359e-01, 1.6606e-01],
[-1.3736e-02, -3.6746e-03, -2.4084e-02, ..., -1.5070e-01,
-8.2230e-02, -5.7828e-03]],
[[-1.1397e-02, -2.6619e-02, -3.4641e-02, ..., 3.2521e-02,
6.6221e-04, -2.5743e-02],
[ 4.5687e-02, 3.3603e-02, -1.0453e-01, ..., -3.1253e-01,
-1.6051e-01, -1.2826e-03],
[-8.3730e-04, 9.8420e-02, 4.0210e-01, ..., 7.0789e-01,
3.6887e-01, 1.2455e-01],
...,
[-5.5926e-02, -5.2239e-03, 2.7081e-02, ..., -4.6178e-01,
-5.7080e-01, -3.6552e-01],
[ 3.2860e-02, 5.5574e-02, 9.9670e-02, ..., 5.4636e-01,
4.8276e-01, 1.9867e-01],
[ 5.3051e-03, 6.6938e-03, -1.7254e-02, ..., -1.4822e-01,
-7.7248e-02, 7.2183e-04]],
[[-2.0315e-03, -9.1617e-03, 2.1209e-02, ..., 8.9177e-02,
3.3655e-02, -2.0102e-02],
[ 1.5398e-02, -1.8648e-02, -1.2591e-01, ..., -2.5342e-01,
-1.2980e-01, -2.7975e-02],
[ 9.8454e-03, 4.9047e-02, 2.1699e-01, ..., 3.4872e-01,
1.0433e-01, 1.8413e-02],
...,
[-2.8356e-02, 1.8404e-02, 9.8647e-02, ..., -1.1740e-01,
-2.5760e-01, -1.5451e-01],
[ 2.0766e-02, -2.6286e-03, -3.7825e-02, ..., 2.4141e-01,
2.4345e-01, 1.1796e-01],
[ 7.4684e-04, 7.7677e-04, -1.0050e-02, ..., -1.4865e-01,
-1.1754e-01, -3.8350e-02]]],
[[[-4.4154e-03, -4.0645e-03, 3.1589e-03, ..., -3.7026e-02,
-2.5158e-02, -4.7945e-02],
[ 5.1310e-02, 5.3402e-02, 8.0436e-02, ..., 1.4480e-01,
1.4287e-01, 1.2312e-01],
[-7.3337e-03, 2.1755e-03, 3.7580e-02, ..., 6.1517e-02,
8.0324e-02, 1.1715e-01],
...,
[-2.6754e-02, -1.2297e-01, -1.3653e-01, ..., -1.4068e-01,
-1.1155e-01, -4.9556e-02],
[ 2.3524e-02, -1.7288e-02, -1.1122e-02, ..., -1.8826e-02,
-2.3320e-02, -2.9474e-02],
[ 2.8689e-02, 2.1659e-02, 4.7888e-02, ..., 2.5498e-02,
3.5346e-02, 1.1280e-02]],
[[ 4.6919e-04, 1.2153e-02, 4.2035e-02, ..., 4.6403e-02,
4.0423e-02, -1.4439e-02],
[ 4.3463e-02, 6.8779e-02, 1.3268e-01, ..., 2.8606e-01,
2.6905e-01, 2.0935e-01],
[-5.7621e-02, -2.2642e-02, 3.0547e-02, ..., 1.3763e-01,
1.6538e-01, 1.7946e-01],
...,
[-1.0816e-01, -2.5227e-01, -2.9742e-01, ..., -2.8503e-01,
-2.1493e-01, -1.0320e-01],
[ 4.0709e-02, -3.2771e-02, -6.3450e-02, ..., -9.2360e-02,
-6.9876e-02, -4.9841e-02],
[ 8.2942e-02, 8.7580e-02, 1.0111e-01, ..., 5.2714e-02,
6.0968e-02, 4.1198e-02]],
[[-1.6391e-02, -1.3870e-02, 5.2810e-03, ..., 4.3698e-02,
2.2707e-02, -4.5983e-02],
[ 3.3202e-02, 4.2014e-02, 9.3500e-02, ..., 2.6162e-01,
2.2970e-01, 1.6694e-01],
[-4.5987e-02, -1.6365e-02, 2.6811e-02, ..., 1.4951e-01,
1.3216e-01, 1.3579e-01],
...,
[-7.2129e-02, -1.8902e-01, -2.3389e-01, ..., -1.9038e-01,
-1.5609e-01, -7.5974e-02],
[ 5.1161e-02, -2.5815e-02, -6.9357e-02, ..., -5.8999e-02,
-6.1550e-02, -4.4555e-02],
[ 1.1174e-01, 7.8979e-02, 6.5849e-02, ..., 3.1617e-02,
2.5221e-02, 7.4257e-03]]],
[[[-7.0826e-08, -6.4306e-08, -7.3806e-08, ..., -9.8000e-08,
-1.0905e-07, -8.3421e-08],
[-6.1125e-09, 2.0613e-09, -8.0922e-09, ..., -4.9840e-08,
-4.3836e-08, -3.0538e-09],
[ 7.1953e-08, 7.5616e-08, 5.9282e-08, ..., -9.7509e-09,
-1.0951e-09, 4.2442e-08],
...,
[ 9.5889e-08, 1.0039e-07, 7.9817e-08, ..., -1.7491e-08,
-4.7666e-08, -1.3265e-08],
[ 1.2904e-07, 1.4762e-07, 1.7477e-07, ..., 1.3233e-07,
1.0628e-07, 9.3316e-08],
[ 1.2558e-07, 1.3644e-07, 1.8431e-07, ..., 2.1399e-07,
1.7710e-07, 1.7166e-07]],
[[-1.2690e-07, -9.6139e-08, -1.0372e-07, ..., -1.1808e-07,
-1.3309e-07, -1.0820e-07],
[-5.7412e-08, -2.5055e-08, -3.0115e-08, ..., -7.2922e-08,
-6.7022e-08, -2.2574e-08],
[ 2.1813e-08, 4.8608e-08, 3.1222e-08, ..., -1.8694e-08,
-7.9591e-09, 3.9750e-08],
...,
[ 5.6013e-08, 7.5526e-08, 4.4496e-08, ..., -4.4128e-08,
-5.9930e-08, -1.8247e-08],
[ 7.7614e-08, 9.8348e-08, 1.0455e-07, ..., 6.3272e-08,
4.1781e-08, 4.5901e-08],
[ 5.9834e-08, 7.1006e-08, 9.0437e-08, ..., 1.1654e-07,
8.7550e-08, 9.8837e-08]],
[[-4.3810e-08, 1.3270e-08, 7.8275e-09, ..., -5.8804e-09,
-2.6217e-08, -1.5649e-08],
[ 4.1700e-08, 1.0778e-07, 1.0946e-07, ..., 7.6403e-08,
7.1450e-08, 9.7615e-08],
[ 1.0436e-07, 1.6586e-07, 1.5933e-07, ..., 1.3517e-07,
1.3487e-07, 1.6449e-07],
...,
[ 9.8763e-08, 1.5072e-07, 1.2547e-07, ..., 6.8316e-08,
6.8382e-08, 1.1367e-07],
[ 9.1435e-08, 1.3576e-07, 1.3793e-07, ..., 1.1678e-07,
1.1723e-07, 1.4394e-07],
[ 6.2183e-08, 8.8184e-08, 1.0456e-07, ..., 1.3941e-07,
1.3333e-07, 1.5844e-07]]],
...,
[[[-6.1896e-02, -3.0206e-02, 1.9225e-02, ..., 4.3665e-02,
-2.2114e-02, -4.2214e-02],
[-3.8061e-02, 6.0774e-03, 4.5797e-02, ..., 9.6029e-02,
5.9254e-02, 2.9958e-02],
[-2.9672e-02, 2.7766e-03, 2.0457e-02, ..., 5.9828e-02,
4.1422e-02, 2.3134e-02],
...,
[ 1.1916e-02, 4.5701e-02, 4.4892e-02, ..., 4.7419e-02,
2.2274e-02, -5.4993e-03],
[-3.2468e-02, -1.2210e-02, 2.2023e-02, ..., 5.8061e-02,
-7.5033e-03, -5.9736e-02],
[-4.3314e-02, -2.8162e-02, -5.9126e-03, ..., 8.8460e-02,
8.4406e-03, -5.0019e-02]],
[[-6.1292e-02, -1.4004e-02, 1.7229e-02, ..., 1.8349e-02,
-3.2708e-02, -4.1060e-02],
[-3.1506e-02, 2.4460e-02, 4.5516e-02, ..., 6.6806e-02,
4.6687e-02, 3.3248e-02],
[-3.2216e-02, 2.0718e-02, 2.3343e-02, ..., 3.5265e-02,
3.6478e-02, 3.1291e-02],
...,
[ 1.7739e-02, 6.1040e-02, 4.8247e-02, ..., 3.7785e-02,
2.8894e-02, 1.3984e-02],
[-1.0890e-02, 2.2079e-02, 4.2737e-02, ..., 6.0247e-02,
1.6197e-02, -1.2493e-02],
[-2.2284e-02, 1.3220e-02, 3.0897e-02, ..., 1.0403e-01,
4.0119e-02, -5.3310e-03]],
[[-8.5322e-02, -4.2603e-02, 6.8145e-03, ..., 3.0751e-02,
-3.4818e-02, -4.9945e-02],
[-2.9215e-02, 1.8165e-02, 5.1092e-02, ..., 9.0200e-02,
5.3438e-02, 4.0169e-02],
[-3.9932e-02, -1.1100e-03, 9.6176e-03, ..., 2.4114e-02,
2.6298e-02, 2.5489e-02],
...,
[-3.1890e-03, 3.0454e-02, 1.6316e-02, ..., 5.5054e-03,
-6.2689e-03, -8.4638e-03],
[-2.2995e-02, -2.8211e-03, 2.3203e-02, ..., 3.5888e-02,
-1.4296e-02, -3.2419e-02],
[-9.8894e-03, 7.0542e-03, 1.0659e-02, ..., 7.0495e-02,
1.2996e-02, -8.3417e-03]]],
[[[-7.8699e-03, 1.9911e-02, 3.4208e-02, ..., 2.8694e-02,
1.2820e-02, 1.8142e-02],
[ 8.7942e-03, -3.2875e-02, -3.5713e-02, ..., 7.2533e-02,
4.5889e-02, 5.2383e-02],
[-3.6122e-02, -1.1878e-01, -1.3767e-01, ..., 3.3811e-02,
3.7806e-02, 2.6944e-02],
...,
[ 1.7322e-02, 3.9589e-03, -8.2269e-03, ..., 2.7543e-03,
1.8313e-02, 1.6057e-02],
[-9.5007e-04, 1.6428e-02, 1.7156e-02, ..., 3.3672e-03,
2.2857e-02, 6.5783e-04],
[ 6.1727e-03, 2.7145e-02, 1.4340e-02, ..., 7.5867e-03,
1.8770e-02, 1.5624e-02]],
[[-1.3423e-02, -5.0696e-04, 8.0959e-03, ..., -6.0963e-03,
9.2341e-03, 1.5751e-02],
[-1.8343e-02, -6.7982e-02, -7.0685e-02, ..., 2.9855e-02,
2.6264e-02, 2.3773e-02],
[-5.4359e-02, -1.4663e-01, -1.6211e-01, ..., 1.1781e-02,
3.2477e-02, 1.1980e-02],
...,
[ 8.3686e-04, -1.7564e-02, -1.9535e-02, ..., -4.1382e-03,
2.4658e-02, 1.2893e-02],
[-6.3183e-04, 1.1788e-02, 2.4810e-02, ..., 6.1105e-03,
3.9210e-02, 9.6696e-03],
[-7.1831e-03, 6.6918e-03, 5.2723e-03, ..., -7.6077e-03,
2.7253e-02, 1.7735e-02]],
[[-2.3753e-04, -4.9343e-03, 2.2991e-03, ..., -4.7958e-02,
-2.6154e-02, -2.3525e-02],
[-3.3053e-04, -5.1502e-02, -5.9977e-02, ..., -1.7369e-02,
-2.3337e-02, -3.7312e-02],
[-2.2674e-02, -9.9412e-02, -1.1176e-01, ..., -1.1725e-02,
-8.3744e-03, -4.0615e-02],
...,
[ 1.1437e-02, -8.0313e-03, -1.4955e-03, ..., -3.4133e-02,
-8.7267e-03, -2.3526e-02],
[ 2.9522e-03, 6.7770e-04, 1.9933e-02, ..., -2.2002e-02,
1.4814e-02, -1.4487e-02],
[-1.9085e-02, -2.9430e-02, -2.3284e-02, ..., -4.8587e-02,
-1.3049e-02, -2.4368e-02]]],
[[[-3.6296e-02, 7.1996e-03, 1.9100e-02, ..., 1.9602e-02,
1.4870e-02, -1.7298e-02],
[-1.1061e-02, 8.5665e-02, 1.2667e-01, ..., 1.3744e-02,
-5.5036e-05, -3.0162e-02],
[ 1.1322e-01, 1.8634e-01, 5.0658e-02, ..., -1.7333e-01,
-7.2041e-02, -6.2474e-02],
...,
[-5.3062e-02, -2.5781e-01, -2.6747e-01, ..., 2.6781e-01,
1.4344e-01, 5.5145e-02],
[-2.1009e-02, -2.9969e-02, 1.0245e-01, ..., 2.0843e-01,
-4.1518e-03, -3.8118e-02],
[-2.2155e-02, 1.2380e-02, 8.4302e-02, ..., -4.4992e-02,
-1.4687e-01, -9.0890e-02]],
[[-5.3969e-03, 3.2799e-02, 1.5486e-02, ..., -7.7451e-03,
3.0229e-03, 1.1216e-03],
[ 6.1723e-02, 1.4899e-01, 1.4645e-01, ..., -2.8897e-02,
-2.0227e-02, -9.1878e-03],
[ 1.6146e-01, 2.0886e-01, -2.5589e-02, ..., -2.7278e-01,
-1.0735e-01, -6.2971e-02],
...,
[-1.3723e-01, -4.0863e-01, -3.8551e-01, ..., 4.0846e-01,
2.6202e-01, 1.3491e-01],
[-5.9388e-02, -6.1187e-02, 1.4197e-01, ..., 3.5780e-01,
9.0893e-02, -1.7392e-03],
[ 7.8613e-03, 5.8403e-02, 1.5339e-01, ..., 4.7045e-02,
-1.0095e-01, -9.7920e-02]],
[[-5.6799e-03, 1.3425e-02, -2.6461e-02, ..., 4.4881e-03,
2.0666e-03, 1.3902e-02],
[ 6.5943e-03, 4.5181e-02, 6.0260e-02, ..., 1.4368e-02,
-5.0725e-03, 4.0505e-03],
[ 5.5257e-02, 1.2397e-01, 4.3193e-02, ..., -1.4486e-01,
-7.4489e-02, -5.7533e-02],
...,
[-3.1513e-02, -1.6334e-01, -1.5795e-01, ..., 2.2904e-01,
1.2017e-01, 7.1998e-02],
[-1.0456e-02, -1.1248e-03, 8.4582e-02, ..., 1.5748e-01,
2.2142e-02, -1.0083e-02],
[-4.8639e-03, -5.0065e-03, 3.6341e-02, ..., -2.4361e-02,
-7.1195e-02, -6.6788e-02]]]], device='cuda:0'),
Parameter containing:
tensor([ 2.3487e-01, 2.6626e-01, -5.1096e-08, 5.1870e-01, 3.4404e-09,
2.2239e-01, 4.2289e-01, 1.3153e-07, 2.5093e-01, 1.5152e-06,
3.1687e-01, 2.5049e-01, 3.7893e-01, 1.0862e-05, 2.7526e-01,
2.3674e-01, 2.4202e-01, 3.9531e-01, 4.6935e-01, 2.9090e-01,
2.7268e-01, 2.7803e-01, 2.9069e-01, 2.0693e-01, 2.5899e-01,
2.7871e-01, 2.9115e-01, 3.1601e-01, 3.8889e-01, 3.0411e-01,
2.6776e-01, 2.1093e-01, 2.8708e-01, 3.3243e-01, 4.2673e-01,
3.7326e-01, 7.4804e-08, 1.9068e-01, 1.4740e-08, 2.2303e-01,
1.7908e-01, 2.4860e-01, 2.7400e-01, 2.5923e-01, 2.9420e-01,
2.9924e-01, 2.2369e-01, 2.6280e-01, 2.2001e-08, 2.6610e-01,
2.2089e-01, 2.8429e-01, 3.3072e-01, 2.2681e-01, 3.6538e-01,
2.1230e-01, 2.3965e-01, 2.4950e-01, 5.2583e-01, 2.4825e-01,
2.9565e-01, 2.5878e-01, 4.8326e-01, 2.6670e-01], device='cuda:0'),
Parameter containing:
tensor([ 2.3072e-01, 2.5382e-01, -1.0543e-06, -6.6439e-01, -1.6571e-08,
1.6152e-01, 4.5450e-01, -4.3020e-07, 3.0051e-01, -8.0052e-06,
3.4942e-01, 3.1148e-01, -2.4953e-01, -3.4749e-05, 1.0773e-01,
2.1897e-01, 3.8141e-01, -5.2988e-01, -6.2864e-01, 5.7140e-01,
2.9985e-01, 5.8430e-01, 4.8202e-01, 3.2853e-01, 1.9672e-01,
1.9496e-01, 1.5215e-01, 8.5522e-02, 5.1314e-01, 1.5237e-02,
1.6644e-01, 3.3239e-01, 2.4921e-01, 4.4337e-01, -2.8017e-01,
-2.0385e-02, -2.4507e-07, 3.2134e-01, -4.9152e-08, 2.3777e-01,
2.3291e-01, 3.1527e-01, 4.2776e-01, 2.9313e-01, 2.6379e-01,
6.7598e-01, 4.2910e-01, 3.4566e-01, -8.6909e-08, 2.4729e-01,
3.0316e-01, 6.1577e-01, 3.9835e-01, 3.3207e-01, -4.1219e-01,
3.7807e-01, 1.7895e-01, 2.5748e-01, -4.4908e-01, 2.1306e-01,
5.6934e-01, 5.7274e-01, -4.0238e-01, 2.3406e-01], device='cuda:0'),
Parameter containing:
tensor([[[[ 5.7593e-02, -9.5114e-02, -2.0272e-02],
[-7.4556e-02, -7.9931e-01, -2.1284e-01],
[ 6.5571e-02, -9.6534e-02, -1.2111e-02]],
[[-6.9944e-03, 1.4266e-02, 5.5824e-04],
[ 4.1238e-02, -1.6125e-01, -2.3208e-02],
[ 3.2887e-03, 7.1779e-03, 7.1686e-02]],
[[-2.3627e-09, -3.9270e-08, -3.2971e-08],
[ 2.1737e-08, 8.3299e-09, 1.2543e-08],
[ 1.1382e-08, 8.8096e-09, 1.5506e-08]],
...,
[[-3.6921e-02, 1.8294e-02, -2.9358e-02],
[-9.8615e-02, -4.3645e-02, -5.2717e-02],
[-7.9635e-02, 2.9396e-02, 4.1479e-03]],
[[ 1.6948e-02, 1.3978e-02, 9.6727e-03],
[ 1.4297e-02, -6.6985e-04, -2.2077e-02],
[ 1.2398e-02, 3.5454e-02, -2.2320e-02]],
[[-2.2600e-02, -2.5331e-02, -2.3548e-02],
[ 6.0860e-02, -9.6779e-02, 2.4057e-02],
[-1.2750e-02, 9.2237e-02, 4.0152e-03]]],
[[[ 2.2160e-02, 4.2177e-02, -1.6428e-02],
[-2.9667e-02, 5.6865e-02, 2.5486e-02],
[ 4.3847e-03, 5.1188e-02, 1.0436e-02]],
[[ 2.5342e-02, 5.4374e-02, 5.3888e-02],
[-2.8334e-02, -2.0139e-01, -5.6358e-02],
[ 5.6774e-02, 7.4188e-02, 2.1585e-02]],
[[-3.1458e-08, 3.5335e-08, 5.3791e-08],
[-2.6896e-08, 5.1530e-08, 5.4480e-08],
[-3.8487e-08, -1.1234e-08, -7.5787e-09]],
...,
[[-1.2754e-01, 4.3552e-02, -6.5607e-02],
[-6.0462e-02, 1.5989e-01, -7.7070e-03],
[-9.4202e-02, 5.0750e-02, -7.8154e-02]],
[[-3.3309e-02, 1.6631e-03, -8.8497e-03],
[ 1.5553e-02, -5.8277e-02, -2.7437e-02],
[ 1.3126e-02, -3.0268e-02, -2.1661e-03]],
[[-4.2313e-03, 3.4517e-02, 3.8193e-03],
[ 5.4317e-02, -1.2457e-02, 3.2900e-02],
[ 2.2000e-04, 1.6040e-02, 1.2764e-01]]],
[[[-3.5247e-02, 8.0748e-03, 2.0353e-02],
[ 1.7344e-02, -2.4320e-02, -1.5511e-04],
[-2.7634e-04, 2.8024e-02, -2.3777e-03]],
[[-2.3741e-02, -3.2057e-03, -5.7059e-03],
[-1.1582e-02, 1.7200e-03, 2.1067e-02],
[ 4.3606e-03, -4.6459e-02, -7.2954e-02]],
[[ 3.1002e-08, 5.3568e-08, 3.1873e-08],
[-1.6063e-08, -1.8072e-08, -1.9508e-09],
[-5.8339e-08, -4.5366e-08, -1.2395e-08]],
...,
[[-1.9689e-03, -2.6809e-02, -4.3760e-02],
[ 2.4518e-02, -2.8396e-02, -3.5896e-02],
[-1.7883e-04, -2.4661e-02, -2.0085e-02]],
[[ 2.1551e-02, 2.2789e-03, -2.5823e-02],
[ 2.3272e-02, -7.9333e-03, -2.0814e-03],
[-5.7062e-03, -2.6934e-02, -1.4421e-02]],
[[-1.9674e-02, 2.7914e-02, -2.0025e-02],
[ 6.3222e-02, -3.9077e-02, -3.3220e-03],
[-2.7434e-02, 1.1390e-02, -3.1608e-03]]],
...,
[[[ 4.3440e-03, -7.6970e-03, -6.4950e-02],
[ 1.3846e-02, -2.2803e-02, -4.6478e-02],
[ 2.7776e-02, 1.6080e-02, -1.3363e-02]],
[[ 4.7379e-02, -2.4982e-02, -2.7605e-02],
[ 7.0091e-02, 4.2084e-03, -1.0805e-01],
[ 1.7526e-02, 4.5647e-02, 7.8810e-03]],
[[ 2.6680e-09, 2.7671e-08, 2.4702e-08],
[ 6.3905e-09, 4.1020e-08, 3.3631e-08],
[ 5.8335e-09, 1.3334e-08, 9.6604e-09]],
...,
[[ 4.5900e-03, 4.7084e-02, -8.6949e-03],
[-6.3011e-03, 5.9585e-02, 5.8667e-03],
[-2.0255e-02, 4.3285e-02, 4.5094e-03]],
[[ 1.1253e-03, -5.7461e-03, -6.8411e-03],
[ 6.0616e-03, 7.3295e-03, -1.1784e-02],
[-1.1455e-03, 5.1868e-03, -1.9867e-02]],
[[ 1.7529e-02, 4.4606e-02, -2.6595e-02],
[ 2.2102e-02, 4.5857e-02, 2.3347e-02],
[ 1.8052e-02, 5.9689e-02, 1.7129e-02]]],
[[[-2.9112e-02, 3.4242e-03, -1.7523e-02],
[-2.3682e-02, 2.2716e-02, -3.8301e-02],
[-1.0308e-02, -4.3802e-03, -2.3582e-02]],
[[-4.9607e-02, -3.2724e-03, -1.5345e-02],
[-1.3524e-02, 5.4842e-02, 1.1187e-02],
[-2.3549e-02, -2.8495e-02, -6.6371e-02]],
[[-4.9804e-08, -2.8211e-08, -2.0583e-08],
[-5.2389e-08, -2.8522e-08, -3.5099e-08],
[-3.2171e-08, -3.4110e-08, -4.3153e-08]],
...,
[[ 3.4487e-03, 2.6532e-02, -1.1202e-02],
[ 7.0925e-03, 3.7903e-02, -3.2481e-02],
[ 4.1381e-02, 3.2329e-02, 2.8309e-03]],
[[-6.5955e-03, 1.6476e-02, 2.1810e-02],
[-1.2293e-02, 2.2310e-02, 1.2645e-02],
[-8.9897e-03, 1.1948e-03, -5.2390e-03]],
[[-2.5295e-03, 7.2689e-02, -7.8046e-03],
[-4.2221e-02, 7.9756e-02, -2.7738e-02],
[ 4.6716e-03, -5.6596e-02, -8.2261e-02]]],
[[[ 5.2235e-02, 3.5231e-03, -3.3131e-02],
[ 3.1048e-02, 1.6193e-02, 1.7283e-02],
[ 1.4446e-02, 2.4302e-02, -1.9689e-03]],
[[-2.4717e-02, 8.3009e-03, -6.1336e-02],
[-1.6134e-02, 5.5323e-02, -6.5029e-02],
[-2.4715e-02, 1.0030e-03, 3.2437e-02]],
[[ 1.8496e-08, 5.2798e-09, 4.1820e-08],
[ 3.7489e-08, 2.5450e-08, 3.0419e-08],
[ 1.1246e-08, -5.6956e-09, -2.0008e-08]],
...,
[[ 7.1194e-03, -4.1052e-02, -1.0002e-02],
[ 2.5924e-02, -6.3819e-02, 1.3366e-02],
[ 2.9751e-02, -7.9476e-03, 1.4007e-02]],
[[-2.5166e-03, 2.2051e-02, -1.9967e-02],
[-5.9436e-02, 4.3872e-02, 2.6832e-02],
[-1.7509e-02, 2.4625e-02, 2.4822e-02]],
[[ 3.5832e-02, -7.0357e-02, 3.9452e-03],
[-2.9835e-02, 9.2727e-02, 1.9336e-02],
[-2.9145e-02, -9.7087e-03, -7.3388e-02]]]], device='cuda:0'),
Parameter containing:
tensor([0.3090, 0.2147, 0.2366, 0.4259, 0.5137, 0.2181, 0.2204, 0.2300, 0.2640,
0.2695, 0.2138, 0.4602, 0.2661, 0.2319, 0.3900, 0.2389, 0.2660, 0.3634,
0.3474, 0.2477, 0.3285, 0.5349, 0.6440, 0.2275, 0.4482, 0.3078, 0.2604,
0.4651, 0.2179, 0.2858, 0.3426, 0.4420, 0.4450, 0.4500, 0.5516, 0.5092,
0.2564, 0.2634, 0.5664, 0.6410, 0.2228, 0.1986, 0.2460, 0.2242, 0.2143,
0.1982, 0.6368, 0.3106, 0.5049, 0.2403, 0.3065, 0.3760, 0.3794, 0.4281,
0.2991, 0.3326, 0.2596, 0.3345, 0.2006, 0.4351, 0.1683, 0.5149, 0.2629,
0.3254], device='cuda:0'),
Parameter containing:
tensor([ 0.1657, 0.2420, 0.1780, -0.0431, -0.2053, 0.1598, 0.2929, 0.0912,
0.1116, 0.0884, 0.1104, -0.2035, 0.1539, 0.0857, -0.1094, 0.0654,
0.0766, -0.2067, -0.0212, 0.1396, 0.0401, -0.2827, -0.3257, -0.0035,
-0.4373, -0.1248, 0.1282, -0.0874, 0.1199, -0.0829, -0.5315, -0.0780,
-0.3876, -0.0547, -0.1816, -0.1888, 0.1320, 0.0031, -0.2697, -0.2984,
0.1394, 0.2597, 0.1372, 0.0053, 0.0132, 0.3295, -0.2715, -0.0187,
-0.2467, 0.1579, 0.0165, -0.0890, -0.1903, -0.0787, 0.1700, -0.4832,
0.0619, -0.0677, 0.3125, -0.5064, 0.3138, -0.2617, -0.1545, 0.0063],
device='cuda:0'),
Parameter containing:
tensor([[[[ 2.5947e-02, -1.0458e-01, -4.7712e-03],
[-8.6223e-02, -3.3021e-01, -1.0275e-01],
[-5.7426e-02, -1.9074e-01, -5.4646e-02]],
[[-1.6951e-02, 2.1384e-02, -2.1074e-03],
[-3.2983e-03, 4.5014e-02, -1.1510e-02],
[-5.9602e-02, 6.4942e-03, 2.9080e-03]],
[[-4.4903e-03, 1.9637e-02, 1.3167e-02],
[ 1.3050e-02, -7.7471e-03, 1.1931e-02],
[ 1.3454e-02, 1.1103e-02, 5.5145e-03]],
...,
[[ 1.2706e-03, -7.7438e-03, 2.0753e-02],
[-4.0024e-02, -4.0383e-02, -3.4821e-02],
[-2.0251e-02, -9.5164e-03, 1.3954e-02]],
[[-2.3430e-03, 3.2303e-02, -4.3342e-03],
[ 8.6194e-03, 1.0553e-02, 1.8074e-03],
[-1.2760e-02, -1.0232e-02, 4.5711e-03]],
[[ 1.5302e-02, 2.1361e-02, -7.0908e-03],
[-1.4221e-02, 4.5979e-02, 2.1369e-02],
[ 3.1312e-02, 6.6428e-02, 2.1465e-02]]],
[[[ 5.3422e-02, 4.0515e-02, 9.6680e-03],
[ 3.2884e-02, -2.3474e-02, 3.4642e-02],
[-1.2861e-02, 5.0066e-02, 5.4579e-02]],
[[ 2.8764e-02, 4.3431e-02, 2.8258e-02],
[ 2.8734e-02, -3.5459e-02, -5.2788e-02],
[-5.5119e-02, -7.1813e-02, -8.2970e-02]],
[[ 9.5293e-02, 1.2549e-01, -6.4001e-02],
[-4.1166e-02, -9.0480e-04, 5.1387e-02],
[-1.1311e-01, -7.9823e-02, 1.4373e-01]],
...,
[[-7.6924e-03, 2.0647e-02, 1.9521e-02],
[-6.7352e-03, 1.2601e-04, 4.8309e-03],
[-6.2405e-03, -9.2119e-03, -2.5806e-04]],
[[-2.6153e-02, -2.4641e-02, 4.0970e-02],
[-1.9164e-02, -1.0160e-02, 3.3163e-02],
[ 5.4200e-03, 9.0485e-04, 6.7799e-04]],
[[ 7.7762e-03, 2.6447e-02, 6.3650e-02],
[-3.0608e-02, 2.4959e-02, 1.2951e-02],
[-2.0938e-02, -7.7342e-03, -3.8790e-02]]],
[[[ 1.0893e-02, -1.4409e-02, 1.5730e-02],
[ 1.6655e-02, 4.4535e-02, 6.3212e-02],
[ 3.4121e-02, 7.3135e-02, 5.9203e-02]],
[[ 2.3195e-03, 7.7598e-03, 2.0308e-02],
[ 2.0457e-02, 4.0029e-02, 3.4744e-02],
[-4.7356e-02, -3.7286e-02, 1.4542e-02]],
[[-2.2742e-02, -1.9000e-02, -8.4317e-03],
[-9.8759e-04, 2.1510e-02, 6.3959e-03],
[-9.4558e-03, 2.6833e-03, -3.1136e-02]],
...,
[[-7.5787e-03, -1.6056e-02, -6.4204e-04],
[-5.5104e-03, 1.4252e-02, 4.5000e-02],
[-9.2800e-03, 2.2351e-02, 4.1728e-02]],
[[ 2.5705e-02, 4.8207e-02, 7.9145e-02],
[-4.4350e-03, 3.8872e-03, 4.1694e-02],
[ 8.0536e-04, -1.0601e-02, 9.2706e-03]],
[[-3.3892e-02, 9.3543e-03, 4.1746e-02],
[-1.6470e-02, 3.9542e-03, 6.2438e-02],
[-3.1055e-02, -3.6302e-03, 7.0817e-02]]],
...,
[[[-7.1044e-05, -9.0020e-03, -2.6998e-03],
[ 3.0072e-03, 1.1579e-02, 1.5214e-02],
[ 3.4832e-03, 1.1353e-05, 1.6320e-02]],
[[-2.6334e-02, 2.1967e-02, -6.0039e-02],
[ 4.4519e-02, 1.3203e-01, -9.1163e-03],
[ 5.4242e-02, 1.3726e-01, 2.7454e-02]],
[[ 1.7122e-02, 3.7646e-03, 1.4872e-02],
[ 1.2092e-02, 1.1319e-02, 3.4667e-02],
[ 8.1790e-03, -2.0805e-02, 2.7143e-02]],
...,
[[-1.0111e-02, -1.0526e-02, 2.8394e-02],
[-2.5112e-02, -2.2196e-02, 3.7229e-02],
[-3.8220e-02, -4.6644e-02, 1.5660e-02]],
[[-2.5913e-03, -2.4307e-02, 1.0611e-02],
[-2.1730e-02, -4.3938e-02, -7.1536e-03],
[-2.5171e-02, -5.9467e-02, -2.5577e-02]],
[[ 2.8652e-02, 2.5850e-04, 1.1416e-03],
[ 3.7812e-02, -1.1271e-03, 9.6027e-03],
[ 3.9350e-02, 1.0134e-02, 1.0449e-02]]],
[[[-7.9305e-03, 7.0872e-03, 2.1412e-02],
[-6.0065e-02, 1.4147e-03, 9.7281e-02],
[-6.0130e-02, -2.1725e-02, 3.6863e-02]],
[[ 2.8024e-02, 2.6183e-02, -2.3027e-02],
[ 5.1900e-02, -2.0588e-03, -1.0940e-01],
[-3.2729e-02, -6.2752e-03, 8.0630e-03]],
[[-1.8062e-02, -1.9510e-02, 4.3163e-02],
[ 4.6080e-02, 2.9494e-02, 4.0844e-02],
[ 5.9607e-03, -6.5891e-03, -6.4623e-03]],
...,
[[ 2.2193e-02, 8.4653e-03, 3.6764e-03],
[ 1.7549e-02, 2.1971e-02, -4.5108e-03],
[ 2.1124e-02, 3.4591e-02, -1.6310e-02]],
[[ 3.8144e-02, 4.8395e-02, -9.5556e-02],
[ 1.8923e-02, 1.1341e-02, -7.6311e-02],
[ 4.7358e-03, 3.2138e-02, -7.4777e-02]],
[[-1.9031e-02, -3.2568e-02, -3.8251e-02],
[ 1.0705e-02, 2.3121e-03, -7.5078e-02],
[ 3.3316e-02, 3.5515e-02, -2.1023e-03]]],
[[[-1.3330e-01, 7.4683e-02, -3.8624e-03],
[ 9.1377e-02, 8.2415e-02, 3.9469e-02],
[-1.8265e-02, -5.9943e-02, 8.9354e-02]],
[[ 1.5566e-02, -4.1716e-02, 1.0633e-02],
[ 7.2644e-03, 3.1934e-02, 1.2732e-03],
[-2.0851e-02, -3.7593e-03, -7.0170e-02]],
[[-6.6139e-02, 1.0627e-01, 1.9590e-02],
[ 5.4987e-02, -1.5552e-01, -1.8819e-02],
[-4.2554e-03, 4.4964e-02, -2.4632e-02]],
...,
[[-6.1691e-02, -4.5531e-02, -9.1721e-03],
[ 4.3995e-02, 4.5703e-02, -7.0108e-02],
[ 1.1388e-02, 4.4678e-02, -4.5953e-02]],
[[ 4.3432e-03, 2.3194e-02, -2.1895e-02],
[-8.0216e-02, -5.7606e-02, -9.8455e-03],
[-3.3285e-02, -1.1468e-01, -2.3779e-02]],
[[-6.3785e-02, -2.4485e-02, -4.9061e-02],
[-6.1594e-02, 1.0328e-01, 5.9685e-03],
[ 8.1863e-02, -3.0314e-02, -4.6373e-03]]]], device='cuda:0'),
Parameter containing:
tensor([0.2496, 0.2198, 0.2756, 0.6073, 0.2654, 0.2942, 0.1136, 0.4425, 0.2868,
0.2974, 0.2506, 0.4103, 0.4855, 0.3383, 0.4670, 0.1772, 0.2171, 0.5025,
0.2263, 0.3667, 0.4867, 0.4586, 0.4652, 0.2200, 0.1510, 0.2761, 0.3813,
0.2803, 0.2382, 0.3953, 0.3032, 0.3163, 0.2025, 0.2323, 0.2003, 0.1661,
0.4690, 0.3476, 0.3414, 0.2274, 0.2485, 0.2356, 0.2726, 0.4657, 0.3429,
0.2465, 0.4674, 0.2812, 0.6241, 0.4152, 0.3403, 0.4218, 0.1152, 0.2985,
0.5802, 0.2795, 0.4706, 0.4517, 0.4303, 0.2749, 0.3427, 0.1137, 0.5069,
0.4370], device='cuda:0'),
Parameter containing:
tensor([ 2.2752e-01, 8.6747e-03, -6.7346e-02, -6.8779e-02, 3.5977e-01,
-2.0167e-01, -4.8431e-05, 2.3735e-02, 3.9549e-01, 3.7079e-02,
6.8793e-03, 2.7578e-01, -7.0272e-02, -2.3970e-01, -8.1753e-02,
-9.4132e-02, -1.4544e-01, 3.7301e-02, -3.6174e-01, -3.9561e-01,
-4.0789e-01, 3.5559e-03, -2.7878e-01, -3.5299e-02, -7.0281e-02,
2.1005e-01, -4.6362e-03, -1.9665e-01, -2.8066e-01, -1.6540e-02,
2.6452e-01, -8.9359e-02, -2.1046e-01, -1.3026e-01, 1.7215e-01,
5.3403e-02, -2.2295e-01, -4.8033e-02, 2.4572e-01, 2.0950e-01,
1.6220e-01, 1.1370e-01, 1.1457e-01, -1.4870e-01, -3.2150e-02,
-3.0549e-01, 4.9125e-01, 1.0873e-01, 1.2779e-02, 1.0044e-01,
4.1553e-01, -1.4710e-02, 2.3922e-02, 9.9812e-02, -1.7273e-01,
1.0078e-01, -1.4564e-01, -2.2735e-01, 1.3637e-01, 2.0127e-01,
-5.7430e-02, 2.3530e-01, -1.1299e-01, 3.0933e-01], device='cuda:0'),
Parameter containing:
tensor([[[[ 1.9712e-02, -5.2562e-03, -3.7619e-03],
[-1.9635e-02, -1.2336e-02, -3.5196e-02],
[ 5.0761e-02, 7.5668e-02, 4.3344e-02]],
[[ 1.4160e-02, -8.6094e-03, -1.0541e-02],
[-4.2586e-02, -2.3814e-02, -5.4694e-02],
[-1.4018e-03, 4.6720e-02, 5.0898e-02]],
[[ 2.1559e-02, 4.1633e-03, -9.7118e-03],
[-9.3201e-03, -2.5432e-02, -2.8274e-02],
[-3.0107e-02, -4.8230e-02, -2.6001e-02]],
...,
[[ 5.4300e-03, 9.1875e-02, 3.1938e-03],
[-1.7945e-02, 5.7266e-02, -8.4098e-03],
[-3.4961e-02, -2.3296e-02, -3.5089e-02]],
[[ 2.5603e-02, -3.1689e-02, -5.4160e-02],
[ 6.9736e-02, -1.0716e-02, -6.8034e-02],
[ 3.5578e-02, 3.4749e-02, -1.9334e-02]],
[[-6.5420e-02, -4.6427e-03, -2.3362e-02],
[ 7.5833e-02, 9.1174e-03, -4.9701e-02],
[ 6.2944e-02, -9.8735e-02, 3.3158e-02]]],
[[[-9.0557e-03, -3.0753e-02, 1.1953e-02],
[-3.2539e-02, -6.2846e-03, -2.0235e-02],
[ 4.7996e-03, -2.1462e-02, -4.1557e-03]],
[[ 1.7163e-02, -2.3303e-03, 7.3972e-02],
[-3.2105e-02, -7.7536e-02, -1.2648e-02],
[ 3.8985e-02, -4.3170e-02, 1.0904e-02]],
[[-2.9643e-02, -5.8534e-02, -5.9736e-02],
[-2.9437e-02, -3.6441e-02, -1.2380e-02],
[-2.2775e-02, -2.4485e-03, -1.6124e-03]],
...,
[[ 2.6830e-02, 1.4267e-02, 6.2658e-02],
[ 3.0585e-04, -5.3241e-03, 3.2786e-03],
[ 2.1097e-02, -2.3189e-02, 1.2102e-02]],
[[-6.1182e-02, -2.9227e-02, 2.0036e-02],
[-7.6089e-02, -7.7057e-02, 8.6544e-02],
[-3.9228e-02, -3.2361e-02, -8.8970e-02]],
[[-1.3372e-01, 8.8362e-02, 8.3836e-02],
[-1.1688e-02, 4.3156e-01, -3.3629e-03],
[-2.3925e-02, -1.0092e-01, -1.0184e-01]]],
[[[ 8.0165e-02, 4.3042e-02, 2.7325e-03],
[ 3.5269e-02, -1.5504e-02, -3.5011e-02],
[-1.7164e-02, -2.6827e-02, -3.3946e-02]],
[[ 4.5439e-02, 5.1585e-02, 1.8321e-02],
[-3.9647e-02, 2.3956e-02, -2.6609e-02],
[-3.0358e-02, -6.4729e-02, 2.5834e-02]],
[[ 3.8105e-02, 4.0986e-02, 4.1005e-02],
[ 1.7584e-02, -1.6494e-02, -3.2716e-02],
[ 5.5886e-03, -1.7068e-02, -3.0605e-02]],
...,
[[-1.3694e-01, -1.4074e-01, 5.1423e-02],
[-1.2521e-01, -1.3128e-01, 7.5733e-02],
[-4.5032e-02, -1.7081e-02, 7.1252e-02]],
[[ 6.3381e-02, 1.5874e-02, -2.7322e-02],
[ 8.0356e-02, 3.6104e-02, -2.8506e-02],
[ 2.6638e-02, 2.2021e-02, 3.2345e-02]],
[[-1.2068e-03, -4.6179e-02, -1.5351e-02],
[-1.1276e-02, 1.9200e-02, 3.4336e-02],
[ 1.6540e-02, -7.8592e-03, -2.5392e-02]]],
...,
[[[ 3.3384e-02, 6.9963e-02, 1.0745e-02],
[-1.7518e-02, -5.3524e-02, -6.4960e-02],
[ 3.4248e-04, -4.5557e-02, -4.7336e-02]],
[[-5.1031e-03, 7.9784e-03, -8.6553e-04],
[-1.6557e-03, 1.4661e-02, 5.3365e-03],
[-3.1784e-02, -6.6940e-02, -4.6889e-02]],
[[-1.1775e-02, 7.2759e-03, 7.6622e-03],
[-6.1288e-02, -5.2078e-02, -4.5152e-02],
[-8.6584e-02, -9.7381e-02, -1.0405e-01]],
...,
[[ 2.1243e-02, 6.2456e-02, 2.5188e-02],
[-2.2911e-02, -2.1100e-03, -2.7573e-02],
[ 4.6557e-02, 6.4980e-02, 3.1879e-02]],
[[ 6.2867e-03, 2.4255e-02, 8.9674e-02],
[-7.7718e-03, -5.4311e-02, -4.6843e-02],
[-6.7499e-03, -6.6857e-02, -4.9842e-02]],
[[ 4.7326e-03, -3.9533e-02, 1.1500e-03],
[-2.7957e-02, -1.3466e-01, -6.0753e-02],
[-3.2010e-03, 7.2213e-02, 1.1009e-01]]],
[[[ 2.3763e-02, -1.7876e-02, -7.4843e-03],
[ 1.6239e-02, 5.4479e-04, -3.3735e-02],
[-2.2854e-02, -1.4316e-03, 1.1010e-02]],
[[ 5.2277e-03, -2.5941e-03, 5.9594e-03],
[-2.9058e-03, -7.3409e-03, 3.0652e-02],
[ 7.5540e-02, 6.6445e-03, 2.5518e-03]],
[[-6.5970e-02, -4.1286e-02, -3.0278e-02],
[-3.5108e-02, -3.9099e-02, -1.6818e-02],
[-1.0224e-02, -8.6995e-03, -5.9939e-04]],
...,
[[ 2.1233e-02, -2.4559e-02, -7.4436e-03],
[-4.3734e-03, -3.2864e-02, -3.3453e-02],
[ 8.9269e-03, -1.7646e-02, 3.8375e-04]],
[[-7.8930e-02, -7.2940e-02, -6.7911e-02],
[-8.4146e-02, -8.3657e-02, 5.3666e-02],
[-3.5577e-02, -3.6835e-02, 5.8987e-03]],
[[ 8.3767e-02, 8.0476e-05, 7.2164e-02],
[-6.4219e-02, -1.2661e-01, 4.6026e-02],
[ 9.3033e-02, -4.7521e-02, 3.6777e-02]]],
[[[ 4.1012e-02, 1.3361e-03, -5.8616e-02],
[ 4.2461e-02, 2.9437e-03, -2.0445e-02],
[ 7.6097e-02, 5.2504e-02, -5.5636e-03]],
[[ 2.2046e-02, 4.0888e-03, 1.4645e-02],
[-7.7532e-02, -1.1912e-01, -7.0892e-02],
[-1.0618e-02, -3.2121e-02, -2.3969e-02]],
[[-2.1612e-02, -2.6110e-03, -3.1664e-02],
[-3.2892e-02, -3.9771e-02, -5.1463e-02],
[-2.6150e-02, -3.6554e-02, -2.3315e-02]],
...,
[[ 4.4600e-03, 8.4181e-02, 2.3199e-02],
[ 5.7595e-02, 1.3036e-01, 3.2172e-02],
[-2.2774e-03, 4.2065e-02, -4.8619e-02]],
[[ 3.1533e-02, -4.3655e-02, 2.0361e-02],
[ 3.9973e-03, -5.1430e-02, -6.3839e-02],
[ 6.4002e-03, 4.5347e-02, 4.7346e-02]],
[[-9.1818e-02, 1.0264e-02, 9.6565e-02],
[-2.1635e-03, -2.3452e-02, -5.9038e-02],
[ 1.9402e-02, 2.8854e-02, -9.6113e-02]]]], device='cuda:0'),
Parameter containing:
tensor([0.3910, 0.4375, 0.3746, 0.3990, 0.3404, 0.3503, 0.2618, 0.2707, 0.2865,
0.4308, 0.1895, 0.3041, 0.3837, 0.2944, 0.2105, 0.3304, 0.2943, 0.2887,
0.2060, 0.4627, 0.2335, 0.1831, 0.4489, 0.2830, 0.3389, 0.2997, 0.3503,
0.2735, 0.3908, 0.2817, 0.2636, 0.4462, 0.3282, 0.3776, 0.4471, 0.3878,
0.2516, 0.3172, 0.3661, 0.3166, 0.3818, 0.3128, 0.2274, 0.3627, 0.2902,
0.2381, 0.2988, 0.2469, 0.3840, 0.2886, 0.3197, 0.2879, 0.3218, 0.4559,
0.3500, 0.2420, 0.3396, 0.3519, 0.3839, 0.3806, 0.4039, 0.2826, 0.4594,
0.3342], device='cuda:0'),
Parameter containing:
tensor([-0.0997, -0.4755, -0.0474, -0.2698, -0.0834, -0.0072, 0.0474, 0.1022,
-0.0170, -0.1471, 0.2307, 0.1447, -0.1775, 0.0273, 0.1559, -0.1836,
0.1238, -0.1522, 0.0554, -0.2881, -0.2606, 0.2316, -0.3242, -0.0219,
-0.2645, 0.0576, -0.2465, 0.0481, -0.3530, 0.0950, -0.1862, -0.1707,
-0.0161, -0.2604, -0.3145, -0.1083, 0.0659, -0.1427, -0.0570, -0.0076,
-0.3006, -0.0744, -0.0683, -0.1104, 0.0253, 0.0489, -0.2515, 0.1150,
-0.3783, 0.0846, -0.0368, 0.1439, -0.0468, -0.3087, -0.0240, 0.1397,
-0.0908, -0.1795, -0.1129, -0.0793, -0.1491, 0.0594, -0.4433, -0.0138],
device='cuda:0'),
Parameter containing:
tensor([[[[-2.1574e-02, -4.5688e-03, 4.5483e-03],
[-8.1870e-03, 4.1740e-02, 2.3010e-02],
[-8.9283e-03, 5.7352e-02, 2.9818e-02]],
[[ 5.8627e-02, 4.2864e-02, 4.4912e-02],
[ 2.2281e-02, -1.2969e-02, 7.6099e-03],
[ 4.5373e-02, 3.0712e-02, 3.7700e-02]],
[[-1.5456e-02, -3.8692e-02, -4.6010e-02],
[-2.3123e-02, 2.8293e-02, 4.7790e-03],
[-2.0328e-02, 1.3756e-02, 2.5883e-02]],
...,
[[ 5.1302e-02, 4.2291e-02, 5.7833e-02],
[ 4.5210e-02, 5.5850e-02, 1.4318e-02],
[ 1.4241e-02, 1.7968e-02, 1.4344e-02]],
[[ 4.6012e-03, 1.2566e-02, 4.8931e-02],
[-6.5754e-03, -2.6431e-02, 1.5855e-02],
[ 1.3192e-02, 1.9011e-02, 1.3842e-02]],
[[ 6.1983e-02, 6.9919e-02, 6.1035e-02],
[ 6.1253e-02, 9.9557e-02, 5.9060e-02],
[ 5.8298e-02, 8.1652e-02, 8.1499e-02]]],
[[[-1.0088e-02, -1.2959e-02, 9.7798e-03],
[ 5.5408e-02, 4.3501e-02, 5.6983e-02],
[ 5.3427e-02, 3.5118e-02, 3.6782e-02]],
[[ 2.4442e-03, -3.0207e-02, -1.0377e-02],
[-4.5297e-02, -4.5318e-02, 5.4623e-03],
[-4.4762e-02, -1.5508e-02, 6.9745e-03]],
[[ 3.9658e-02, 3.6838e-02, 5.8796e-03],
[ 2.3207e-02, 3.9240e-03, -2.0887e-02],
[-1.4829e-02, 5.3606e-03, 1.7404e-03]],
...,
[[ 3.2160e-02, 5.9042e-02, 4.8433e-02],
[-2.6464e-02, -8.0667e-03, -1.0359e-02],
[-2.6699e-02, -9.5411e-03, -2.8902e-02]],
[[-2.9235e-02, -3.9078e-02, -4.4955e-02],
[-2.0346e-02, -4.4891e-02, -3.7477e-02],
[ 1.9653e-02, -1.5562e-03, -5.8245e-03]],
[[-5.0696e-02, -4.8902e-02, 9.1631e-03],
[ 5.1668e-03, 2.0509e-02, 6.6874e-02],
[ 2.8934e-02, 4.6717e-02, 2.1371e-02]]],
[[[ 2.1744e-02, -2.8354e-02, -3.2557e-02],
[ 3.0519e-02, 1.8536e-02, 1.5244e-02],
[ 1.3832e-03, 1.7051e-02, 3.2020e-02]],
[[-3.6293e-02, 1.0914e-02, 4.5371e-02],
[ 1.3399e-02, 6.4272e-02, 8.8210e-02],
[ 4.6697e-02, 9.9653e-02, 8.7606e-02]],
[[-2.4336e-02, -2.9627e-02, 1.9537e-02],
[-3.3412e-02, -2.2290e-02, -2.8879e-02],
[ 1.4765e-02, 1.7234e-02, -1.8185e-02]],
...,
[[-3.9859e-02, -7.1075e-02, -5.8546e-02],
[ 2.2902e-02, 1.1184e-02, -2.3654e-02],
[ 8.1897e-02, 1.1996e-01, 9.3242e-02]],
[[ 3.1984e-02, 7.4931e-02, 6.6020e-02],
[ 2.8490e-02, 1.1931e-01, 1.2100e-01],
[ 7.9259e-04, 4.3812e-02, 4.4648e-02]],
[[ 3.2748e-02, 4.1444e-02, -8.1932e-03],
[ 4.5541e-02, 2.9426e-02, -8.5440e-03],
[ 1.1634e-04, 1.8045e-03, 1.4826e-02]]],
...,
[[[-4.4144e-02, -8.3106e-02, -5.3073e-02],
[ 3.2124e-02, 1.0286e-02, 2.4409e-02],
[ 6.1606e-03, -1.9455e-02, 4.0534e-02]],
[[ 5.6026e-04, 9.6961e-03, 2.5010e-03],
[ 7.1679e-03, -1.7535e-02, -2.3857e-02],
[-9.8745e-03, -1.8550e-02, 1.7301e-03]],
[[ 4.3882e-03, 4.2049e-02, 7.5950e-02],
[-6.5610e-02, -3.6130e-02, -1.9404e-02],
[-3.8091e-02, -2.6749e-02, -1.3865e-02]],
...,
[[-4.5593e-02, -4.6050e-02, -2.2809e-02],
[-9.7648e-03, 2.4910e-03, 2.4503e-02],
[ 2.0381e-02, 5.2393e-02, 6.9019e-02]],
[[ 9.3306e-04, 1.2483e-02, -1.1817e-02],
[-1.2627e-02, -1.8756e-02, -1.4144e-03],
[-5.2490e-03, -4.6126e-03, -1.3224e-02]],
[[ 7.4689e-04, -1.0135e-02, -7.8264e-03],
[ 1.2491e-02, -2.5865e-02, 4.0514e-02],
[ 5.8855e-03, 4.5990e-02, 1.0651e-01]]],
[[[ 1.2262e-02, -1.5378e-02, 1.3862e-03],
[ 4.1166e-02, -2.4944e-02, -2.6686e-02],
[-1.7423e-02, 5.2690e-03, -2.1861e-02]],
[[-3.1207e-02, -3.3025e-02, 2.2114e-02],
[-2.4009e-02, 1.2988e-02, 2.2430e-02],
[ 1.0332e-02, 4.3601e-03, 4.7321e-03]],
[[ 2.0182e-02, 6.1569e-02, -2.8771e-02],
[ 5.8231e-02, 4.6767e-02, -2.8417e-05],
[ 3.7545e-02, -4.5886e-02, 1.5849e-02]],
...,
[[ 7.0431e-03, -3.6082e-03, 7.1986e-03],
[ 2.4895e-02, 6.1671e-03, -3.2427e-02],
[ 7.2338e-03, 2.2406e-03, -5.3330e-02]],
[[ 2.8072e-02, -1.0571e-02, -1.3854e-02],
[-1.0879e-02, 6.1929e-03, -5.6713e-03],
[-2.6083e-02, 8.1861e-03, -3.2873e-02]],
[[-3.1032e-02, -6.0485e-02, -2.5583e-02],
[-4.6239e-02, -2.2805e-02, -7.7678e-03],
[-9.4698e-03, 4.0247e-03, -4.8637e-03]]],
[[[ 2.3128e-02, -5.6038e-02, -3.4572e-02],
[ 1.0638e-03, 5.7929e-02, -7.6970e-03],
[-3.0103e-02, 3.5573e-02, -1.8143e-02]],
[[ 9.6840e-02, -1.1186e-01, -7.8766e-02],
[-1.0444e-01, -1.0851e-01, -1.9553e-01],
[-1.1986e-01, -7.1474e-02, 3.6750e-02]],
[[-2.2194e-02, 6.0298e-03, 5.6914e-02],
[-4.8342e-02, 7.8893e-02, -5.1026e-02],
[-5.1294e-02, -5.7434e-02, -1.9178e-02]],
...,
[[-4.4896e-02, -8.1267e-02, 5.1794e-02],
[-8.3985e-02, -5.7778e-02, 6.7891e-02],
[ 2.3837e-02, 3.8954e-02, 4.1141e-02]],
[[ 4.6446e-03, 2.7367e-02, -2.3154e-02],
[ 2.0675e-02, 2.3429e-02, 6.4380e-04],
[-5.2222e-02, -1.4854e-02, -2.5150e-02]],
[[ 2.1291e-02, 1.2736e-02, 8.4553e-03],
[-8.2932e-02, 7.2067e-02, 1.3107e-01],
[ 8.5491e-03, 1.3677e-01, 3.9867e-02]]]], device='cuda:0'),
Parameter containing:
tensor([0.2560, 0.5690, 0.4042, 0.5130, 0.2178, 0.4940, 0.3315, 0.5510, 0.4354,
0.5291, 0.2081, 0.4735, 0.5945, 0.5645, 0.2761, 0.2571, 0.4853, 0.6240,
0.4370, 0.2308, 0.4970, 0.3157, 0.5706, 0.2162, 0.1932, 0.1448, 0.2218,
0.2389, 0.5871, 0.3501, 0.4109, 0.3199, 0.5808, 0.3281, 0.2723, 0.1971,
0.6139, 0.4075, 0.6304, 0.3874, 0.7605, 0.2111, 0.3071, 0.4603, 0.3099,
0.1914, 0.4431, 0.2537, 0.5745, 0.6459, 0.3914, 0.3090, 0.6782, 0.1937,
0.5814, 0.2570, 0.3514, 0.2124, 0.5794, 0.3415, 0.2051, 0.0715, 0.4090,
0.4416], device='cuda:0'),
Parameter containing:
tensor([-0.1778, -0.1287, 0.0349, -0.1452, 0.1864, -0.1413, -0.4201, -0.1334,
0.2183, -0.1912, 0.0311, -0.0235, -0.1724, -0.0274, -0.0295, -0.1031,
0.0047, 0.0828, -0.1521, 0.0183, -0.2418, -0.0831, -0.0491, -0.0688,
-0.2560, 0.1381, -0.0165, 0.2092, -0.0028, -0.0265, -0.0225, 0.0286,
-0.1065, -0.3698, 0.2862, -0.1036, 0.3080, -0.0894, 0.2772, 0.1136,
-0.3157, 0.0423, 0.0567, 0.2369, -0.0727, 0.0465, -0.0536, 0.1309,
0.0282, -0.1371, 0.1464, -0.0717, -0.3237, -0.1583, -0.0424, -0.1278,
-0.1703, 0.0413, 0.0891, 0.0770, -0.0730, 0.0683, -0.0391, 0.0476],
device='cuda:0'),
Parameter containing:
tensor([[[[-7.1555e-02, -1.1031e-01, -1.3711e-01],
[ 7.0593e-02, -1.4782e-02, -1.0053e-01],
[ 1.1938e-01, 8.7330e-02, -8.2206e-03]],
[[-2.3999e-02, -6.3682e-03, 2.4303e-03],
[ 6.1831e-03, 1.8781e-02, 2.5324e-02],
[ 2.3656e-03, -4.0037e-03, -1.1949e-02]],
[[ 6.0344e-03, 6.3784e-03, -1.2247e-02],
[ 7.8854e-03, -1.3464e-02, -4.2702e-02],
[ 1.7380e-02, -1.3862e-02, -4.7145e-02]],
...,
[[ 3.4324e-02, 3.2257e-02, 2.5819e-02],
[ 8.4676e-03, -4.5413e-04, -1.0832e-02],
[-6.7166e-03, -1.5052e-02, -2.6939e-02]],
[[-1.2089e-02, -2.3588e-02, -2.2689e-02],
[ 1.0135e-02, 1.8285e-02, -1.5695e-02],
[ 2.1352e-02, 5.8568e-02, 4.2873e-02]],
[[ 1.4421e-02, -2.8298e-02, -7.0770e-03],
[ 3.0260e-02, -6.6294e-03, -1.6901e-02],
[ 3.9085e-02, 1.4222e-02, 2.2294e-02]]],
[[[-7.7911e-02, -7.3929e-02, -3.6671e-02],
[-3.4903e-02, -6.2355e-02, -3.7793e-02],
[-2.8379e-02, -5.4291e-02, -4.9411e-02]],
[[-1.2970e-02, -2.1825e-02, -2.8767e-04],
[ 7.6444e-03, 1.7653e-02, 1.6660e-02],
[ 3.8337e-02, 2.3006e-02, -1.6620e-03]],
[[-8.7592e-02, -8.4735e-02, -5.5818e-02],
[-7.7731e-02, -8.0311e-02, -3.2554e-02],
[-5.6313e-02, -4.2047e-02, 1.5247e-02]],
...,
[[-3.2377e-02, -4.0018e-02, -2.9523e-02],
[-1.5294e-02, -1.4165e-02, 2.7086e-03],
[ 1.1652e-02, 2.3886e-02, 2.4413e-02]],
[[ 2.0891e-03, -3.0475e-02, -3.3818e-02],
[ 6.7829e-03, 3.8681e-04, -1.4540e-02],
[-3.1306e-03, 6.7689e-03, 8.4524e-03]],
[[ 3.0586e-02, 4.6281e-02, 3.8359e-04],
[ 5.3079e-02, 6.7488e-02, 3.0547e-02],
[ 2.3374e-02, 4.3993e-02, -3.8713e-03]]],
[[[ 1.3878e-02, 3.2724e-02, 4.6584e-02],
[-8.0647e-03, 1.6209e-03, 1.5153e-02],
[-7.0342e-02, -5.3299e-02, -4.5920e-02]],
[[ 4.6035e-02, 3.5400e-02, 3.4941e-02],
[ 5.8351e-02, 5.4640e-02, 2.7162e-02],
[ 2.6799e-02, 4.5056e-02, 6.6886e-03]],
[[-3.3766e-02, -3.8605e-02, -2.4172e-02],
[-1.8285e-03, 1.0888e-02, 1.1425e-02],
[ 2.2282e-02, 1.4024e-02, 3.6332e-03]],
...,
[[-1.6330e-02, -6.9552e-02, -8.9737e-02],
[ 3.9766e-02, 1.5501e-02, -2.2695e-02],
[ 1.0290e-01, 1.2294e-01, 6.3867e-02]],
[[-4.2318e-03, 4.9511e-02, -7.6289e-03],
[-2.7720e-02, 7.0398e-03, -9.4052e-03],
[-6.7008e-02, -6.0542e-02, -2.5967e-02]],
[[-5.8560e-03, -1.7573e-02, -3.8016e-02],
[ 2.8579e-03, -4.1603e-03, 1.0113e-02],
[ 2.6243e-02, 3.5200e-02, 3.1143e-02]]],
...,
[[[-4.4193e-02, -6.5322e-02, -1.7594e-02],
[-9.3970e-02, -5.8291e-02, 1.2093e-02],
[-2.2998e-02, 3.2463e-02, 7.1731e-02]],
[[-4.7220e-03, -3.0125e-03, -1.8075e-02],
[ 1.2667e-02, -8.0509e-03, -1.4605e-02],
[ 7.8220e-03, -1.0720e-02, -2.6515e-02]],
[[-2.5299e-02, -4.9383e-02, -1.2720e-02],
[-5.2206e-02, -4.7233e-02, -4.2470e-03],
[-4.8697e-02, -2.5320e-02, 8.6178e-03]],
...,
[[-3.7617e-03, 7.8398e-03, -5.9525e-03],
[ 4.0277e-03, 7.3575e-03, -1.1667e-02],
[-3.9997e-02, -3.8038e-02, -5.0469e-02]],
[[-3.8949e-03, -6.8965e-03, 3.4102e-02],
[-6.9814e-03, -4.9762e-02, 5.8711e-02],
[ 1.8361e-02, 2.5874e-02, 8.0028e-02]],
[[-3.3014e-02, -2.1510e-02, -2.1509e-03],
[-4.3894e-02, -3.2009e-02, -1.6265e-02],
[-1.1037e-02, 2.8872e-04, 3.0937e-02]]],
[[[-4.9907e-02, -5.0222e-02, -5.0985e-02],
[ 2.2644e-02, -1.4098e-02, -2.4426e-02],
[ 1.9960e-02, 9.6426e-02, 1.0580e-01]],
[[-3.6873e-02, 2.1413e-03, 8.3469e-03],
[-4.0796e-02, -3.3767e-02, -3.4955e-02],
[ 3.9466e-02, 7.0508e-02, 8.6065e-02]],
[[ 1.4842e-02, 6.6914e-03, 1.4324e-02],
[-3.2621e-02, -4.4027e-02, -2.2269e-02],
[ 7.1982e-03, -1.9187e-02, -4.9348e-03]],
...,
[[-4.9938e-03, 1.6018e-02, 1.1242e-02],
[-4.7668e-03, 2.1921e-02, 2.2660e-02],
[-2.6753e-02, 2.6917e-04, -5.6827e-03]],
[[-8.7725e-03, 1.0761e-02, 7.3603e-03],
[-1.8010e-05, -1.7926e-02, 4.8229e-03],
[ 4.2431e-02, -1.5764e-02, 2.3554e-02]],
[[-1.3830e-02, -3.0793e-03, -4.0854e-03],
[ 3.3363e-02, 4.2952e-02, 3.5867e-02],
[-3.9653e-02, -3.0855e-02, -4.3189e-02]]],
[[[-3.8617e-02, -3.1549e-03, 2.5739e-03],
[-1.1592e-02, 9.8761e-03, 7.5235e-03],
[-1.9339e-02, -9.8779e-03, 2.1755e-03]],
[[ 1.6889e-04, 1.8302e-03, -8.9537e-03],
[ 5.8343e-03, 1.7360e-02, -1.9029e-02],
[ 5.8642e-03, -7.4307e-04, 1.4667e-03]],
[[-1.6506e-02, -2.8401e-02, 1.3986e-02],
[-2.2922e-02, -4.3484e-02, 1.0471e-02],
[-2.5801e-03, -4.5258e-02, 7.9791e-03]],
...,
[[-1.5260e-03, -7.6469e-03, 1.3597e-02],
[ 5.5301e-04, -2.9176e-03, 2.2147e-02],
[ 3.2763e-03, -1.0775e-05, 1.3163e-02]],
[[ 5.1756e-03, 1.8495e-02, -8.0268e-03],
[-3.5030e-02, 2.6403e-02, -7.1220e-03],
[-5.2325e-02, -1.1185e-02, 1.9146e-02]],
[[-6.8805e-02, 5.1618e-02, 1.9787e-02],
[ 2.5533e-02, -6.1926e-02, 4.9924e-02],
[ 1.0532e-01, -4.4136e-02, 4.9907e-02]]]], device='cuda:0'),
Parameter containing:
tensor([0.3248, 0.3613, 0.2960, 0.2913, 0.3407, 0.3435, 0.3049, 0.3308, 0.3447,
0.3860, 0.3196, 0.2622, 0.2994, 0.2189, 0.2397, 0.3744, 0.3555, 0.1948,
0.3349, 0.2159, 0.3349, 0.3454, 0.3094, 0.3769, 0.3546, 0.3267, 0.3178,
0.3272, 0.3832, 0.2585, 0.2973, 0.3481, 0.2827, 0.2995, 0.3451, 0.3471,
0.3440, 0.3344, 0.3211, 0.3180, 0.2940, 0.3353, 0.3253, 0.3733, 0.3198,
0.2987, 0.1620, 0.3262, 0.3271, 0.3410, 0.3693, 0.3320, 0.3357, 0.2951,
0.3115, 0.3185, 0.3139, 0.2633, 0.3089, 0.3601, 0.2734, 0.3433, 0.3335,
0.3288, 0.2706, 0.2879, 0.3318, 0.3310, 0.3170, 0.2977, 0.3300, 0.3216,
0.3205, 0.3231, 0.3481, 0.3130, 0.2826, 0.2856, 0.3279, 0.3666, 0.3288,
0.3575, 0.3377, 0.2904, 0.3273, 0.3214, 0.3332, 0.3452, 0.1842, 0.3916,
0.3337, 0.2325, 0.3285, 0.3358, 0.2885, 0.3149, 0.3288, 0.2236, 0.3159,
0.2993, 0.3403, 0.3220, 0.3171, 0.2950, 0.2847, 0.3224, 0.3119, 0.2613,
0.3374, 0.3333, 0.3330, 0.2959, 0.4087, 0.2192, 0.2982, 0.4006, 0.3081,
0.3171, 0.2862, 0.2952, 0.3070, 0.3583, 0.3232, 0.3345, 0.3453, 0.3043,
0.3327, 0.3337], device='cuda:0'),
Parameter containing:
tensor([-0.0589, -0.1686, -0.0206, 0.0027, -0.0955, -0.1048, 0.0349, -0.0885,
-0.2053, -0.1764, -0.1224, -0.0364, -0.0785, 0.2088, -0.0403, -0.1820,
-0.1076, 0.2989, -0.0570, 0.2064, -0.0921, -0.1376, -0.1304, -0.1193,
-0.1006, -0.0380, -0.1108, -0.0477, -0.1087, 0.1581, -0.1123, -0.1584,
0.0976, -0.0430, -0.1349, -0.1189, -0.0986, -0.0479, -0.0837, -0.0720,
-0.0836, -0.2442, -0.3376, -0.2124, -0.0693, -0.0651, 0.4979, -0.0811,
-0.1021, -0.0788, -0.1802, -0.1011, -0.1090, -0.0617, -0.0856, -0.0495,
-0.0370, 0.0023, -0.0508, -0.2430, 0.0009, -0.1525, -0.0963, -0.0516,
-0.0473, 0.0884, -0.1028, -0.0907, -0.1086, -0.0379, -0.1030, -0.1609,
-0.0903, -0.0898, -0.1282, -0.0830, -0.0186, -0.0232, -0.0045, -0.2131,
-0.1431, -0.1391, -0.1303, -0.0568, -0.1862, -0.1209, -0.0340, -0.1181,
0.2298, -0.2085, -0.1335, 0.1418, -0.0891, -0.1273, 0.0107, -0.1029,
-0.1025, 0.1562, -0.0937, -0.0657, -0.1245, -0.0451, -0.0707, -0.0447,
0.0715, -0.0484, -0.0312, -0.0437, -0.0927, -0.1465, -0.1151, -0.0183,
-0.1927, 0.2491, 0.0300, -0.1310, -0.0468, -0.0851, -0.0421, -0.0413,
-0.0457, -0.1433, -0.0981, -0.1046, -0.1315, -0.1249, -0.0982, -0.0961],
device='cuda:0'),
Parameter containing:
tensor([[[[-0.0074, -0.0098, 0.0028],
[-0.0108, 0.0258, 0.0455],
[-0.0272, 0.0053, 0.0132]],
[[ 0.0354, 0.0251, 0.0078],
[ 0.0040, 0.0199, 0.0274],
[ 0.0353, 0.0355, 0.0133]],
[[ 0.0193, -0.0213, -0.0362],
[-0.0196, -0.0189, -0.0595],
[-0.0218, -0.0077, 0.0039]],
...,
[[-0.0068, 0.0108, -0.0037],
[ 0.0135, 0.0114, -0.0013],
[ 0.0081, 0.0002, 0.0006]],
[[ 0.0077, 0.0077, 0.0044],
[-0.0102, -0.0117, -0.0096],
[-0.0039, 0.0181, 0.0133]],
[[ 0.0124, -0.0269, -0.0120],
[ 0.0268, 0.0264, -0.0215],
[ 0.0129, 0.0028, -0.0054]]],
[[[ 0.0128, 0.0185, -0.0027],
[-0.0168, -0.0123, 0.0355],
[-0.0009, 0.0245, 0.0182]],
[[-0.0067, -0.0207, -0.0144],
[-0.0073, 0.0426, 0.0074],
[ 0.0276, 0.0160, 0.0159]],
[[-0.0229, -0.0206, 0.0236],
[-0.0275, -0.0561, -0.0699],
[ 0.0205, 0.0513, 0.0220]],
...,
[[-0.0272, -0.0390, -0.0395],
[-0.0621, -0.0749, -0.0956],
[-0.0605, -0.0777, -0.0746]],
[[ 0.0287, 0.0293, 0.0287],
[ 0.0079, 0.0471, 0.0146],
[-0.0018, 0.0220, 0.0074]],
[[ 0.0016, -0.0168, -0.0046],
[-0.0081, -0.0255, -0.0524],
[-0.0093, -0.0010, -0.0378]]],
[[[ 0.0047, 0.0061, -0.0071],
[ 0.0236, -0.0931, -0.0793],
[-0.0079, -0.0505, -0.0105]],
[[ 0.0148, 0.0162, -0.0515],
[ 0.0086, 0.0081, -0.0429],
[ 0.0908, 0.0654, 0.0435]],
[[-0.0138, -0.0064, 0.0085],
[ 0.0138, -0.0124, 0.0054],
[ 0.0202, -0.0035, 0.0080]],
...,
[[-0.0009, 0.0185, -0.0306],
[-0.0170, 0.0057, -0.0491],
[-0.0328, -0.0374, -0.0459]],
[[-0.0046, 0.0069, -0.0011],
[-0.0079, -0.0499, 0.0421],
[-0.0752, -0.0048, -0.0058]],
[[ 0.0115, -0.0146, 0.0379],
[ 0.0141, 0.0486, 0.0232],
[ 0.0215, -0.0101, 0.0338]]],
...,
[[[ 0.0264, 0.0040, 0.0075],
[ 0.0636, -0.0320, -0.0018],
[ 0.0262, 0.0076, 0.0495]],
[[-0.0287, -0.0227, -0.0513],
[-0.0260, -0.0487, -0.0140],
[-0.0173, -0.0416, -0.0117]],
[[-0.0350, 0.0356, 0.0347],
[ 0.0183, 0.0436, -0.0263],
[ 0.0178, 0.0356, 0.0113]],
...,
[[ 0.0097, -0.0173, -0.0002],
[-0.0199, 0.0820, 0.0129],
[-0.0238, -0.0048, 0.0486]],
[[-0.0244, -0.0258, -0.0353],
[-0.0296, -0.0966, -0.0535],
[-0.0150, 0.0059, -0.0197]],
[[ 0.0068, -0.0368, -0.0255],
[-0.0116, -0.0236, -0.0078],
[ 0.0086, 0.0079, -0.0189]]],
[[[ 0.0329, 0.0297, 0.0111],
[ 0.0390, -0.0090, 0.0226],
[ 0.0078, -0.0280, -0.0230]],
[[ 0.0137, 0.0229, -0.0190],
[ 0.0027, 0.0112, 0.0074],
[ 0.0211, 0.0436, 0.0108]],
[[-0.0237, 0.0221, 0.0004],
[-0.0309, 0.0609, 0.0167],
[-0.0885, -0.0834, -0.0342]],
...,
[[-0.0073, -0.0197, 0.0018],
[ 0.0073, -0.0343, -0.0243],
[-0.0115, -0.0605, -0.0551]],
[[ 0.0030, 0.0026, 0.0171],
[-0.0134, -0.0086, 0.0090],
[ 0.0195, 0.0094, 0.0045]],
[[ 0.0014, 0.0008, 0.0119],
[-0.0024, -0.0107, 0.0126],
[-0.0051, -0.0058, -0.0146]]],
[[[-0.0074, 0.0118, -0.0427],
[ 0.0016, -0.0457, -0.1398],
[-0.0065, -0.0021, -0.0484]],
[[ 0.0075, 0.0527, 0.0388],
[-0.0125, 0.0847, 0.0062],
[ 0.0013, -0.0197, -0.0822]],
[[-0.0249, 0.0166, 0.0169],
[ 0.0087, 0.0214, 0.0117],
[-0.0009, 0.0306, 0.0136]],
...,
[[-0.0078, 0.0220, -0.0175],
[-0.0076, -0.0211, -0.0037],
[ 0.0126, -0.0207, -0.0054]],
[[ 0.0302, -0.0082, -0.0649],
[-0.0238, -0.0954, -0.0530],
[-0.0168, -0.0111, 0.0010]],
[[-0.0245, -0.0847, 0.0251],
[ 0.0106, 0.0387, 0.1400],
[ 0.0155, -0.0095, 0.0041]]]], device='cuda:0'),
Parameter containing:
tensor([0.1454, 0.3270, 0.3113, 0.2538, 0.4086, 0.3937, 0.4400, 0.3108, 0.3406,
0.2168, 0.2170, 0.3857, 0.1971, 0.2692, 0.1663, 0.2454, 0.3232, 0.3686,
0.3893, 0.3264, 0.3875, 0.4707, 0.1958, 0.4717, 0.1673, 0.3938, 0.3044,
0.1929, 0.2175, 0.2119, 0.4230, 0.3683, 0.2455, 0.2229, 0.3370, 0.3229,
0.2688, 0.3557, 0.2581, 0.4031, 0.4492, 0.3642, 0.2599, 0.1881, 0.1359,
0.2958, 0.1913, 0.3065, 0.3981, 0.4102, 0.1874, 0.4516, 0.3340, 0.1628,
0.3599, 0.1624, 0.2886, 0.1358, 0.4491, 0.2694, 0.4823, 0.3393, 0.4764,
0.3155, 0.6005, 0.4654, 0.5264, 0.2991, 0.2992, 0.4621, 0.2614, 0.4247,
0.4662, 0.4249, 0.3345, 0.2655, 0.4048, 0.3605, 0.1782, 0.3833, 0.2823,
0.3843, 0.3307, 0.2151, 0.3317, 0.1458, 0.2771, 0.4917, 0.3199, 0.4222,
0.1559, 0.4884, 0.3267, 0.3440, 0.1608, 0.4855, 0.2677, 0.1616, 0.3221,
0.4243, 0.3661, 0.1893, 0.3400, 0.3648, 0.1779, 0.3544, 0.2852, 0.2437,
0.4472, 0.3011, 0.3997, 0.6173, 0.2794, 0.4867, 0.1502, 0.6021, 0.3604,
0.4696, 0.3711, 0.2388, 0.5347, 0.1509, 0.3213, 0.4394, 0.3229, 0.4329,
0.1489, 0.3702], device='cuda:0'),
Parameter containing:
tensor([ 0.0246, 0.0593, 0.1347, -0.1089, -0.0470, -0.1359, -0.0550, 0.0509,
-0.0613, 0.0916, 0.0031, -0.0274, -0.0539, 0.0177, 0.0432, 0.0074,
0.0548, -0.0321, -0.0224, 0.0142, -0.2150, -0.1160, 0.0486, -0.1141,
0.1066, 0.0355, 0.0140, 0.0177, 0.0781, 0.1331, 0.0139, 0.0447,
0.1063, 0.0528, -0.0539, -0.1160, 0.1055, -0.1591, 0.0100, 0.1197,
0.0170, 0.0929, -0.0675, 0.0987, 0.1034, 0.0501, 0.0297, 0.0281,
-0.0075, -0.0577, -0.0144, -0.1640, 0.1255, 0.0817, 0.0635, 0.0936,
0.0213, 0.0486, -0.1174, 0.0237, -0.2177, 0.0099, -0.1883, 0.0467,
-0.0829, 0.0585, -0.0306, 0.0509, 0.0541, -0.1671, 0.0115, -0.0302,
-0.1393, 0.0115, 0.0428, 0.1189, -0.1289, 0.0479, 0.0474, -0.0625,
0.0009, -0.0144, 0.0909, 0.1342, -0.0338, 0.0560, 0.0848, -0.0467,
0.0228, -0.0097, 0.1360, -0.2625, 0.0088, -0.0553, 0.0383, -0.0720,
0.0907, 0.1612, -0.1076, 0.1011, -0.0519, 0.0838, -0.0704, -0.0806,
-0.0243, 0.0533, 0.1277, 0.1403, -0.0593, -0.0639, -0.0766, -0.1163,
0.0661, -0.1644, 0.0422, -0.2786, -0.1006, -0.0696, -0.0761, 0.0371,
-0.0247, 0.0916, -0.0200, -0.0176, 0.0298, -0.0373, 0.0466, -0.1371],
device='cuda:0'),
Parameter containing:
tensor([[[[ 0.0159]],
[[-0.3109]],
[[ 0.0126]],
...,
[[-0.1672]],
[[ 0.0127]],
[[ 0.0132]]],
[[[ 0.0036]],
[[-0.0011]],
[[-0.0083]],
...,
[[-0.0234]],
[[-0.0756]],
[[-0.0126]]],
[[[-0.0419]],
[[ 0.0079]],
[[-0.1662]],
...,
[[ 0.0319]],
[[-0.0188]],
[[ 0.0645]]],
...,
[[[-0.0287]],
[[ 0.0470]],
[[-0.0523]],
...,
[[-0.0474]],
[[ 0.0586]],
[[ 0.0588]]],
[[[-0.0078]],
[[-0.0203]],
[[ 0.0564]],
...,
[[ 0.7802]],
[[-0.0023]],
[[-0.0259]]],
[[[-0.0283]],
[[-0.0132]],
[[-0.0514]],
...,
[[ 0.0236]],
[[-0.0677]],
[[ 0.0268]]]], device='cuda:0'),
Parameter containing:
tensor([ 0.3334, 0.0581, 0.0715, 0.3442, 0.1756, 0.1509, 0.1568, 0.3100,
0.1927, 0.1516, 0.3044, 0.2238, 0.3706, 0.1739, 0.3051, 0.2610,
0.1575, 0.2015, 0.2933, 0.1010, 0.5871, 0.0676, 0.2499, 0.0929,
0.2443, 0.0495, 0.2449, 0.2750, 0.3071, 0.3025, 0.1818, 0.0688,
0.2223, 0.3766, 0.4661, 0.3284, 0.1035, 0.3400, 0.2325, 0.1514,
0.1753, 0.2269, 0.2606, 0.1831, 0.2894, 0.2590, 0.2208, 0.1399,
0.0643, 0.2833, 0.3451, 0.2017, 0.0696, 0.2722, 0.1127, 0.2917,
0.2358, 0.2703, 0.0911, 0.2591, 0.1302, 0.2261, 0.1967, 0.0539,
0.0697, 0.0524, 0.1050, 0.0861, 0.1173, 0.0957, 0.1862, 0.1642,
0.1336, 0.1065, 0.1312, 0.0888, 0.0793, 0.0475, 0.3049, 0.2325,
0.2908, 0.1292, 0.0778, 0.2263, 0.2379, 0.3405, 0.0914, 0.1936,
0.1223, 0.1400, 0.2953, 0.2360, 0.1681, 0.1338, 0.2666, 0.1495,
0.0761, 0.1674, 0.1784, 0.1720, 0.2318, 0.3753, 0.2103, 0.1922,
0.4002, 0.1718, 0.0593, 0.0742, 0.0686, 0.1931, 0.1386, 0.1111,
0.3055, 0.1205, 0.3443, 0.1633, 0.3673, 0.1534, 0.0742, 0.2088,
0.0394, 0.2594, 0.1385, -0.0051, 0.1905, 0.1275, 0.3071, 0.1682],
device='cuda:0'),
Parameter containing:
tensor([ 0.0246, 0.0593, 0.1347, -0.1089, -0.0470, -0.1359, -0.0550, 0.0509,
-0.0613, 0.0916, 0.0031, -0.0274, -0.0539, 0.0177, 0.0432, 0.0074,
0.0548, -0.0321, -0.0224, 0.0142, -0.2150, -0.1160, 0.0486, -0.1141,
0.1066, 0.0355, 0.0140, 0.0177, 0.0781, 0.1331, 0.0139, 0.0447,
0.1063, 0.0528, -0.0539, -0.1160, 0.1055, -0.1591, 0.0100, 0.1197,
0.0170, 0.0929, -0.0675, 0.0987, 0.1034, 0.0501, 0.0297, 0.0281,
-0.0075, -0.0577, -0.0144, -0.1640, 0.1255, 0.0817, 0.0635, 0.0936,
0.0213, 0.0486, -0.1174, 0.0237, -0.2177, 0.0099, -0.1883, 0.0467,
-0.0829, 0.0585, -0.0306, 0.0509, 0.0541, -0.1671, 0.0115, -0.0302,
-0.1393, 0.0115, 0.0428, 0.1189, -0.1289, 0.0479, 0.0474, -0.0625,
0.0009, -0.0144, 0.0909, 0.1342, -0.0338, 0.0560, 0.0848, -0.0467,
0.0228, -0.0097, 0.1360, -0.2625, 0.0088, -0.0553, 0.0383, -0.0720,
0.0907, 0.1612, -0.1076, 0.1011, -0.0519, 0.0838, -0.0704, -0.0806,
-0.0243, 0.0533, 0.1277, 0.1403, -0.0593, -0.0639, -0.0766, -0.1163,
0.0661, -0.1644, 0.0422, -0.2786, -0.1006, -0.0696, -0.0761, 0.0371,
-0.0247, 0.0916, -0.0200, -0.0176, 0.0298, -0.0373, 0.0466, -0.1371],
device='cuda:0'),
Parameter containing:
tensor([[[[-9.9023e-04, -7.7429e-03, -7.9740e-03],
[ 2.4844e-02, 1.8642e-03, 5.8352e-03],
[ 9.5089e-03, -1.6476e-02, 3.9157e-03]],
[[-2.1488e-02, -1.2330e-03, -1.4281e-02],
[-1.7044e-02, 9.5922e-03, 7.0445e-03],
[ 1.0790e-02, -7.2350e-03, -1.1357e-02]],
[[-1.1126e-03, 3.0388e-02, 2.2247e-02],
[-6.1184e-02, -2.3797e-02, 2.3747e-03],
[ 4.0678e-02, -1.0356e-01, -6.0011e-02]],
...,
[[-8.5833e-03, 1.1438e-02, 2.0800e-02],
[-1.6565e-02, -3.9587e-02, 1.2594e-02],
[-1.4314e-03, -5.4257e-03, 3.6794e-02]],
[[-1.3687e-02, -2.9514e-02, -1.4745e-02],
[ 2.8299e-02, 2.2096e-02, 3.4839e-03],
[-4.3521e-03, -2.6706e-03, 1.2258e-04]],
[[ 7.6403e-03, 2.0666e-02, 3.7429e-02],
[ 6.9478e-03, 4.3983e-02, 1.7538e-02],
[-9.7797e-03, -2.4789e-02, -1.1349e-03]]],
[[[ 8.4439e-02, 8.4827e-02, -5.1478e-02],
[ 3.5253e-02, -1.1375e-03, -1.0331e-01],
[-6.4078e-02, -1.2660e-01, -1.2952e-01]],
[[ 1.0628e-03, -1.4083e-02, 4.7109e-03],
[-2.1059e-02, -2.8778e-02, 9.9708e-03],
[ 1.4074e-02, 1.8691e-02, 5.8192e-02]],
[[ 2.2139e-02, 8.9027e-03, 1.4790e-02],
[-1.7497e-02, -5.3924e-03, 2.7834e-02],
[-1.3855e-02, -1.3346e-02, 1.7668e-02]],
...,
[[-3.8032e-02, -2.3097e-02, -7.1775e-03],
[-3.5089e-02, 1.0861e-02, 1.3640e-02],
[ 6.3449e-04, 9.7476e-03, 7.3670e-03]],
[[-4.4184e-02, -1.6190e-02, 1.2243e-02],
[-4.0349e-02, -1.7894e-02, 2.8911e-02],
[-6.5176e-03, -1.0490e-02, 9.1658e-03]],
[[ 4.3621e-03, 1.3119e-02, 1.8442e-03],
[ 1.1555e-02, -1.3031e-02, -9.5657e-03],
[-2.3314e-02, 1.1609e-03, 2.6771e-03]]],
[[[-2.1180e-02, -6.2213e-03, 1.7609e-03],
[-4.7424e-03, 1.1101e-02, 1.1296e-02],
[-1.4529e-02, 2.9843e-02, 2.4383e-03]],
[[ 6.9183e-03, 9.2937e-03, 3.0078e-02],
[-4.2612e-03, 4.9560e-03, -4.7338e-03],
[ 3.1360e-02, 1.9035e-03, -4.7242e-03]],
[[-3.6726e-02, 5.7285e-03, 1.3919e-01],
[-4.2992e-02, 9.4023e-04, 7.7141e-02],
[-5.0050e-02, -4.9479e-03, 2.4693e-02]],
...,
[[ 3.7203e-02, 7.4712e-03, -4.2659e-02],
[-8.1729e-03, -9.2536e-02, -5.4934e-03],
[-2.5927e-02, 8.3993e-04, 7.4632e-02]],
[[ 1.8076e-02, 4.5272e-03, -1.3757e-02],
[-1.8939e-02, -3.2739e-02, -2.9666e-02],
[-2.0608e-02, -4.6167e-03, 1.3080e-03]],
[[-1.2078e-02, -2.0285e-03, -1.6998e-02],
[-3.4805e-02, -4.9195e-02, -3.1973e-02],
[-2.1021e-02, -5.1164e-03, -4.8522e-03]]],
...,
[[[ 3.1791e-02, 2.2948e-02, 1.0390e-02],
[-1.2628e-02, -2.9320e-03, 4.2645e-03],
[-2.1707e-02, -1.0856e-02, 1.6094e-02]],
[[-1.4525e-03, -1.0131e-02, -4.6862e-04],
[ 2.2130e-02, 2.2736e-02, 5.0183e-03],
[-6.0125e-02, -4.3150e-02, -4.4480e-02]],
[[ 3.0761e-03, 3.4396e-03, 6.0877e-03],
[-1.3683e-02, 4.0576e-03, -2.6544e-02],
[ 6.8231e-02, 6.3474e-02, -9.3660e-03]],
...,
[[ 1.8752e-02, 1.9400e-02, 4.1691e-02],
[ 8.7770e-03, 8.2394e-04, 1.8619e-02],
[ 1.8796e-02, 6.2238e-02, -2.3801e-02]],
[[-2.9788e-02, -3.4598e-02, -2.5225e-02],
[ 8.4234e-03, -2.3222e-02, -9.4612e-03],
[ 6.9035e-03, 6.9737e-02, -1.3359e-02]],
[[ 2.6981e-03, -4.3182e-02, -1.6731e-02],
[ 2.5812e-02, -7.2025e-02, -6.5399e-02],
[ 4.6257e-02, 2.9469e-02, -1.5811e-02]]],
[[[-2.1079e-02, 3.8220e-02, 8.3305e-03],
[-5.9912e-03, 3.5584e-02, -1.7534e-03],
[ 1.8735e-02, 7.0859e-03, -3.5151e-03]],
[[-4.5937e-02, -7.4695e-02, -5.3608e-02],
[-8.6266e-03, 9.0894e-03, -3.0345e-02],
[-2.8158e-02, -2.1204e-02, -8.4730e-03]],
[[-7.1772e-02, -6.8582e-02, 2.5544e-02],
[ 5.0363e-02, 2.5269e-02, 5.6668e-02],
[ 2.6238e-03, 1.3871e-03, -8.4692e-03]],
...,
[[-2.9644e-02, 1.0896e-02, -3.0402e-02],
[ 1.5095e-03, 5.0455e-02, 1.5597e-02],
[-2.1015e-02, -1.0757e-02, -3.4942e-02]],
[[-2.7573e-02, 2.9707e-02, -2.9490e-02],
[ 2.3301e-03, -3.9011e-02, 6.8010e-03],
[ 4.4006e-02, 3.5397e-02, 7.9087e-02]],
[[-2.7480e-02, 5.0337e-02, 1.4290e-02],
[-5.2482e-02, -4.7748e-03, 1.2988e-02],
[-1.8935e-02, -3.0808e-02, -1.7583e-02]]],
[[[ 3.2280e-02, 4.7408e-02, 3.4054e-02],
[ 2.1445e-02, 3.8987e-03, 4.6985e-04],
[ 1.5159e-02, 8.2067e-03, 3.2426e-02]],
[[ 9.2653e-03, 2.3661e-02, 4.2089e-02],
[ 2.1976e-02, 4.6128e-02, 1.1402e-02],
[ 7.2843e-03, 5.2285e-02, 8.6340e-03]],
[[ 1.4022e-02, 1.2800e-02, 3.5398e-02],
[-4.4398e-02, 1.7399e-02, -1.5838e-02],
[ 3.1712e-02, 5.8679e-02, -9.3244e-03]],
...,
[[-4.8399e-03, 7.8628e-03, -5.6169e-04],
[ 8.0402e-03, 1.7392e-02, 7.8734e-03],
[-1.7713e-02, -4.5957e-02, -9.8762e-03]],
[[-9.7569e-03, -7.5795e-03, -2.4627e-02],
[-8.2454e-03, 6.3065e-02, -3.2954e-03],
[-7.7549e-03, -1.3404e-04, -8.1337e-03]],
[[ 1.7664e-02, 1.0114e-02, 4.2687e-03],
[-3.7950e-03, 2.6715e-02, 2.0121e-02],
[ 1.6868e-02, -6.6515e-03, -1.1107e-02]]]], device='cuda:0'),
Parameter containing:
tensor([0.3323, 0.2908, 0.3246, 0.3435, 0.3011, 0.3054, 0.3041, 0.3539, 0.2862,
0.3601, 0.2970, 0.3381, 0.2565, 0.3276, 0.3030, 0.4085, 0.3519, 0.4218,
0.3055, 0.2551, 0.3425, 0.3215, 0.3366, 0.2700, 0.2849, 0.3954, 0.3166,
0.3286, 0.3515, 0.3953, 0.2768, 0.3625, 0.1988, 0.2717, 0.3355, 0.2797,
0.2510, 0.3832, 0.3266, 0.3263, 0.3681, 0.3401, 0.3651, 0.3391, 0.3071,
0.3231, 0.3691, 0.2410, 0.3536, 0.3189, 0.3238, 0.3611, 0.3086, 0.3309,
0.3886, 0.4362, 0.4550, 0.2962, 0.3071, 0.3386, 0.3317, 0.3228, 0.2393,
0.3147, 0.2738, 0.3218, 0.3198, 0.3411, 0.3611, 0.2833, 0.3035, 0.3183,
0.3146, 0.3890, 0.2607, 0.3479, 0.3236, 0.3709, 0.2592, 0.3742, 0.2555,
0.2966, 0.3505, 0.3165, 0.2808, 0.2660, 0.2817, 0.4795, 0.3372, 0.2723,
0.2955, 0.3225, 0.2470, 0.3160, 0.3515, 0.3131, 0.3372, 0.2837, 0.3540,
0.2897, 0.2490, 0.3019, 0.3114, 0.3510, 0.3022, 0.3617, 0.2859, 0.2831,
0.3243, 0.2769, 0.3314, 0.2394, 0.2932, 0.2788, 0.2686, 0.3194, 0.3542,
0.2683, 0.2955, 0.2924, 0.3538, 0.4256, 0.3603, 0.3013, 0.2763, 0.4354,
0.3991, 0.2694], device='cuda:0'),
Parameter containing:
tensor([-0.1735, -0.2337, -0.3383, -0.0806, -0.1920, -0.0621, -0.1885, -0.2830,
-0.1680, -0.1796, -0.2645, -0.1983, -0.1183, -0.2432, -0.1706, -0.3090,
-0.2661, -0.4040, -0.1949, -0.1392, -0.2449, -0.1242, -0.2012, -0.1901,
-0.1014, -0.3468, -0.2245, -0.3272, -0.3057, -0.3289, -0.1532, -0.1967,
-0.0667, -0.3281, -0.1418, -0.1527, -0.0987, -0.3243, -0.2252, -0.3462,
-0.2284, -0.2263, -0.1810, -0.1564, -0.1730, -0.1507, -0.2913, -0.1643,
-0.1998, -0.1532, -0.2211, -0.2247, -0.0913, -0.1563, -0.2453, -0.4854,
-0.4428, -0.1021, -0.1615, -0.2125, -0.2239, -0.1952, -0.0447, -0.1733,
-0.1178, -0.4775, -0.2110, -0.2305, -0.1795, -0.1582, -0.2008, -0.2041,
-0.1974, -0.2750, -0.0395, -0.2161, -0.2786, -0.2626, -0.0997, -0.2953,
-0.1431, -0.1448, -0.1894, -0.1283, -0.1807, -0.1144, -0.1308, -0.4154,
-0.2324, -0.1376, -0.1154, -0.2099, -0.0966, -0.1669, -0.3835, -0.2545,
-0.1603, -0.1904, -0.2420, -0.1658, -0.1133, -0.1498, -0.1213, -0.2318,
-0.2017, -0.3827, -0.1491, -0.1174, -0.1261, -0.2031, -0.1832, -0.2274,
-0.1281, -0.2557, -0.1400, -0.0723, -0.2212, -0.1486, -0.2914, -0.1116,
-0.2194, -0.4898, -0.3693, -0.1437, -0.1232, -0.3723, -0.6794, -0.1536],
device='cuda:0'),
Parameter containing:
tensor([[[[-1.6153e-02, 5.0134e-03, -9.0186e-04],
[-8.8386e-03, -1.9390e-02, -2.4174e-02],
[ 6.3052e-03, 1.0245e-02, -1.3816e-02]],
[[-1.0979e-02, 2.6164e-03, 2.3656e-02],
[-1.7687e-02, 1.9861e-02, 6.4150e-02],
[ 6.0224e-03, 7.6342e-02, 1.0215e-01]],
[[-8.1113e-03, 6.8414e-03, 2.5436e-02],
[-8.0696e-03, 9.2929e-03, 8.2899e-03],
[ 7.7306e-03, 1.2159e-02, 7.1625e-03]],
...,
[[ 1.5175e-02, 6.2196e-03, 2.1798e-02],
[-1.5199e-02, -8.5439e-02, -2.4713e-02],
[-1.8460e-02, -4.9767e-02, -1.6818e-03]],
[[ 3.0728e-02, 3.9962e-02, 3.1253e-02],
[-1.8738e-02, -6.7510e-02, -2.7649e-02],
[ 2.8429e-02, 3.1854e-02, 1.0543e-02]],
[[-1.8320e-02, -1.5854e-02, -1.0685e-02],
[-2.7442e-02, -3.0616e-02, -1.0485e-02],
[-1.5122e-02, -1.0595e-02, -2.5322e-02]]],
[[[ 3.6868e-03, 3.0996e-02, 4.2763e-02],
[ 4.6537e-02, 4.8606e-02, 2.3800e-03],
[ 1.6654e-02, 1.2900e-02, -1.8230e-02]],
[[-1.0441e-02, -1.5934e-03, -1.6128e-02],
[-1.2799e-02, 4.9570e-03, -1.4585e-02],
[-2.3553e-02, -3.7023e-03, -1.4399e-02]],
[[ 1.0338e-02, -1.7560e-02, -3.3046e-02],
[-3.2090e-02, -5.9258e-03, 2.0201e-03],
[-4.1428e-02, 4.9121e-03, 1.6906e-02]],
...,
[[-4.9525e-02, -4.6498e-02, -5.9916e-02],
[-2.6670e-02, -1.9079e-02, -2.9419e-02],
[-3.9683e-03, 1.9405e-02, 7.3317e-03]],
[[ 1.4293e-02, 1.5643e-02, 5.8117e-04],
[ 5.1493e-03, 7.4332e-03, -3.6928e-03],
[-1.3522e-02, -8.5536e-03, -2.1259e-03]],
[[-3.0908e-02, -1.9839e-02, -1.9375e-02],
[-1.0368e-02, -2.4294e-02, 2.4103e-04],
[-1.9275e-02, -2.9707e-02, -1.5623e-02]]],
[[[-4.9212e-02, -2.9588e-02, 8.8023e-02],
[ 4.7453e-03, 4.3564e-02, 9.3115e-02],
[ 7.4083e-02, 4.2868e-02, -5.1033e-02]],
[[ 6.6992e-03, 2.1676e-02, -5.4254e-04],
[ 1.9286e-02, 1.0920e-02, -4.5440e-03],
[ 3.1075e-02, -1.7168e-03, -2.7603e-02]],
[[ 6.0096e-02, -2.9359e-02, -5.8911e-02],
[-1.9133e-02, -8.1624e-02, -2.2553e-02],
[ 1.1597e-02, 2.5092e-02, 1.2130e-02]],
...,
[[ 5.4307e-03, -2.3130e-02, 9.6233e-03],
[-4.3785e-02, -2.6735e-02, 2.1993e-02],
[-3.5919e-02, -4.1009e-02, -2.1860e-02]],
[[ 3.3705e-02, 6.2938e-02, 4.3502e-02],
[ 1.1111e-03, 1.9243e-02, -1.9707e-03],
[-1.1493e-02, -5.3445e-02, -9.6676e-03]],
[[-2.6664e-03, -2.6954e-02, -1.7667e-02],
[-8.3382e-03, 8.9920e-03, 8.1260e-04],
[-2.6832e-02, -3.5991e-02, -4.2495e-02]]],
...,
[[[-1.8876e-03, -2.2728e-02, -4.2991e-03],
[-9.2231e-03, -3.4333e-02, -1.3392e-02],
[-1.2774e-02, -1.1435e-02, 1.5617e-02]],
[[ 1.0703e-02, 1.2792e-02, 2.2662e-02],
[ 7.3185e-03, -1.7847e-02, 1.0674e-02],
[-1.5936e-02, -1.9318e-02, 2.1768e-02]],
[[-7.3009e-03, 3.0234e-02, -1.1899e-02],
[-2.6099e-02, 3.7452e-03, 3.2776e-02],
[-3.3101e-02, -7.1923e-03, 1.6559e-02]],
...,
[[-3.2818e-02, -1.0021e-01, -4.7012e-02],
[ 2.8293e-03, 4.1410e-02, -1.1391e-02],
[-1.1152e-02, -5.5861e-03, 1.9968e-02]],
[[-2.3932e-02, -3.0687e-02, -1.1756e-03],
[ 1.5311e-03, -3.5002e-02, -2.4414e-02],
[-8.7575e-03, -7.7842e-02, -3.8842e-02]],
[[ 2.6107e-02, 1.5406e-02, 1.7569e-02],
[-1.5130e-02, -4.8687e-03, 3.0773e-03],
[-1.3470e-02, -9.3201e-03, -4.8982e-03]]],
[[[-2.0228e-02, -3.0006e-02, -9.8419e-03],
[-3.8676e-02, -3.3481e-02, -7.4265e-03],
[-2.8935e-02, -3.2037e-02, 2.9245e-03]],
[[-1.2900e-02, 3.8046e-03, 1.5940e-02],
[-2.4030e-02, 2.0666e-03, 5.7250e-03],
[ 6.9989e-03, 1.2192e-02, 1.5406e-02]],
[[-1.5018e-02, -9.0988e-03, 2.4450e-02],
[ 1.0039e-02, 1.2561e-02, 2.6997e-02],
[ 2.9556e-02, 1.9463e-02, -2.6584e-03]],
...,
[[-1.8481e-02, 3.9417e-04, 9.9768e-03],
[-4.5447e-03, 1.2307e-02, 3.5507e-02],
[-1.1873e-03, -2.6185e-03, 1.1547e-02]],
[[ 4.6292e-03, -1.3690e-02, -1.0171e-02],
[ 1.2104e-02, 1.6793e-02, 1.3003e-02],
[ 1.3328e-03, 3.4701e-03, 1.7323e-02]],
[[-8.7332e-05, 5.8646e-03, -3.5117e-03],
[ 3.8112e-03, -7.1828e-03, -1.1407e-02],
[ 1.9705e-02, 2.0556e-02, 5.7084e-03]]],
[[[ 3.6998e-02, 3.2616e-02, -9.4535e-04],
[-2.9484e-02, -2.3441e-02, -2.8085e-02],
[-2.5451e-02, 3.9048e-02, 3.6686e-02]],
[[-1.8732e-02, -1.5352e-02, 1.1149e-02],
[-2.1324e-03, -2.3177e-02, 1.7628e-02],
[-4.0012e-03, 1.5463e-02, 9.2496e-03]],
[[-2.9346e-02, 7.7071e-03, -5.6520e-03],
[-2.3611e-02, -1.9390e-03, 2.0221e-02],
[ 8.0955e-03, -2.3268e-02, -2.8827e-02]],
...,
[[-3.3532e-02, -2.9092e-02, -4.0045e-02],
[ 2.6530e-03, -2.0568e-02, 1.3075e-02],
[ 1.6061e-02, -5.5725e-02, -4.9167e-02]],
[[-7.9132e-03, 2.1466e-02, 2.0913e-02],
[-1.7259e-02, -2.5851e-02, 2.7177e-03],
[-4.6532e-02, -2.4846e-02, -1.9911e-02]],
[[-5.0350e-02, -2.5574e-02, 1.7763e-02],
[-3.4474e-02, 5.5247e-03, -2.7754e-02],
[-2.0743e-02, -2.2332e-02, -4.3512e-02]]]], device='cuda:0'),
Parameter containing:
tensor([0.1194, 0.1625, 0.3084, 0.2931, 0.2957, 0.5263, 0.4038, 0.2024, 0.3401,
0.1982, 0.2559, 0.2311, 0.1630, 0.2891, 0.2248, 0.2311, 0.2417, 0.2187,
0.1922, 0.3103, 0.2015, 0.4802, 0.2481, 0.3898, 0.3204, 0.4035, 0.2617,
0.1551, 0.2256, 0.2117, 0.2708, 0.3537, 0.2505, 0.1843, 0.2465, 0.6501,
0.3898, 0.4289, 0.1799, 0.1604, 0.1775, 0.3600, 0.2694, 0.1283, 0.1662,
0.1716, 0.1837, 0.1710, 0.4178, 0.3249, 0.1759, 0.4717, 0.4115, 0.1995,
0.2025, 0.1492, 0.2860, 0.1072, 0.3649, 0.1906, 0.5369, 0.2400, 0.4411,
0.1702, 0.1993, 0.2045, 0.1972, 0.4041, 0.3034, 0.6168, 0.2284, 0.3228,
0.4547, 0.4370, 0.1570, 0.4057, 0.5791, 0.2338, 0.1586, 0.3130, 0.2201,
0.3195, 0.1166, 0.2517, 0.2184, 0.0989, 0.3116, 0.2613, 0.3277, 0.1778,
0.2718, 0.4174, 0.5140, 0.2136, 0.1905, 0.2898, 0.2472, 0.1341, 0.6212,
0.1810, 0.2394, 0.1417, 0.1759, 0.2827, 0.1987, 0.3775, 0.3749, 0.1274,
0.3656, 0.4305, 0.4212, 0.2673, 0.2016, 0.5098, 0.1449, 0.4408, 0.3583,
0.2503, 0.5682, 0.2518, 0.1392, 0.0617, 0.3406, 0.1313, 0.4586, 0.2914,
0.1326, 0.3915], device='cuda:0'),
Parameter containing:
tensor([-0.1403, -0.0889, -0.4147, -0.2264, -0.0737, -0.3534, -0.3379, -0.0752,
-0.1791, 0.0448, -0.2842, -0.1765, -0.1591, -0.0675, -0.1543, -0.1061,
-0.2334, -0.0981, -0.0908, -0.0567, -0.1908, -0.2055, -0.2704, -0.1883,
-0.3570, -0.1125, -0.1632, -0.0211, -0.1687, -0.2124, -0.1713, -0.0872,
-0.2194, -0.1888, -0.2954, -0.4570, -0.0226, -0.0527, 0.0406, -0.0609,
-0.0456, -0.1176, -0.0145, 0.0318, -0.2046, -0.0953, -0.0496, -0.1051,
-0.0793, -0.1933, -0.1467, -0.3215, -0.3257, -0.2287, -0.0356, -0.1869,
-0.1932, -0.0771, 0.2768, -0.0656, -0.0895, -0.2548, -0.2365, 0.0021,
-0.0987, -0.3178, 0.1613, 0.0006, -0.2347, -0.4150, -0.1310, -0.3142,
-0.2582, -0.5400, 0.0772, -0.2546, -0.4454, -0.0262, -0.0937, -0.2201,
-0.2044, -0.0155, -0.0893, -0.2167, 0.1112, -0.0619, -0.1217, -0.1593,
-0.1317, -0.1717, -0.3729, -0.3354, -0.3414, 0.0358, -0.2067, -0.1087,
0.0141, -0.0338, -0.2129, -0.1122, -0.1627, -0.2000, 0.0908, -0.0041,
-0.1313, -0.2942, 0.0160, -0.1065, -0.1289, -0.1699, -0.1721, -0.1809,
-0.2295, -0.3611, -0.1746, -0.3540, -0.1554, -0.2709, -0.2607, 0.0084,
-0.0311, -0.0022, -0.0831, 0.0380, -0.4893, -0.2749, 0.1245, -0.1272],
device='cuda:0'),
Parameter containing:
tensor([[[[-0.0159, -0.0166, -0.0159],
[-0.0053, 0.0151, 0.0099],
[-0.0149, 0.0004, -0.0114]],
[[-0.0095, -0.0186, -0.0061],
[ 0.0098, -0.0123, -0.0053],
[ 0.0071, -0.0161, -0.0071]],
[[-0.0227, -0.0377, -0.0337],
[-0.0316, -0.0580, -0.0391],
[-0.0346, -0.0388, -0.0157]],
...,
[[ 0.0049, 0.0080, 0.0144],
[-0.0015, 0.0242, 0.0056],
[-0.0044, 0.0062, 0.0069]],
[[ 0.0160, -0.0120, -0.0013],
[ 0.0096, 0.0057, 0.0016],
[-0.0099, -0.0136, -0.0064]],
[[ 0.0534, 0.0464, 0.0248],
[ 0.0341, -0.0029, -0.0041],
[-0.0140, -0.0046, -0.0142]]],
[[[-0.0012, -0.0186, -0.0345],
[ 0.0050, -0.0117, -0.0333],
[ 0.0060, -0.0162, -0.0175]],
[[ 0.0107, 0.0140, -0.0192],
[ 0.0029, 0.0126, 0.0073],
[-0.0168, -0.0187, -0.0014]],
[[-0.0219, 0.0063, 0.0159],
[-0.0035, 0.0057, 0.0311],
[ 0.0023, 0.0032, 0.0175]],
...,
[[-0.0162, -0.0078, 0.0077],
[-0.0015, -0.0122, -0.0093],
[ 0.0009, -0.0022, 0.0074]],
[[ 0.0029, -0.0236, 0.0058],
[ 0.0143, -0.0169, -0.0062],
[-0.0077, -0.0323, -0.0337]],
[[ 0.0087, 0.0140, 0.0081],
[-0.0034, 0.0105, 0.0150],
[ 0.0189, 0.0309, 0.0256]]],
[[[-0.0358, -0.0226, -0.0144],
[-0.0075, -0.0221, 0.0111],
[-0.0036, -0.0148, -0.0164]],
[[-0.0146, -0.0307, -0.0204],
[-0.0285, -0.0453, -0.0579],
[ 0.0288, -0.0152, -0.0245]],
[[ 0.0174, 0.0199, -0.0046],
[ 0.0178, 0.0236, 0.0136],
[ 0.0293, 0.0434, 0.0186]],
...,
[[ 0.0079, -0.0122, -0.0158],
[-0.0107, -0.0314, -0.0075],
[-0.0070, -0.0173, -0.0308]],
[[-0.0103, -0.0118, -0.0171],
[-0.0264, -0.0015, 0.0280],
[-0.0089, 0.0054, 0.0097]],
[[ 0.0145, -0.0315, -0.0197],
[-0.0149, -0.0177, 0.0077],
[ 0.0125, 0.0071, -0.0062]]],
...,
[[[ 0.0053, 0.0255, 0.0071],
[ 0.0197, 0.0270, 0.0420],
[ 0.0412, 0.0223, 0.0350]],
[[ 0.0026, 0.0041, 0.0106],
[ 0.0046, 0.0203, 0.0081],
[ 0.0145, -0.0030, -0.0197]],
[[-0.0040, 0.0255, 0.0023],
[-0.0152, 0.0260, 0.0083],
[-0.0021, 0.0269, 0.0032]],
...,
[[ 0.0055, -0.0049, 0.0183],
[ 0.0085, 0.0023, 0.0077],
[ 0.0082, 0.0096, 0.0215]],
[[-0.0177, 0.0099, -0.0129],
[-0.0127, 0.0096, -0.0124],
[ 0.0090, 0.0493, 0.0362]],
[[ 0.0123, 0.0184, -0.0178],
[ 0.0058, -0.0058, -0.0117],
[ 0.0034, -0.0103, -0.0425]]],
[[[ 0.0316, 0.0765, 0.0443],
[ 0.0940, 0.1480, 0.1510],
[ 0.0665, 0.1386, 0.1132]],
[[-0.0282, -0.0041, -0.0200],
[-0.0193, -0.0012, 0.0107],
[-0.0165, -0.0028, 0.0008]],
[[-0.0122, -0.0322, -0.0153],
[-0.0009, -0.0123, -0.0039],
[-0.0193, -0.0107, -0.0211]],
...,
[[-0.0030, 0.0066, -0.0003],
[ 0.0146, -0.0167, 0.0142],
[ 0.0227, 0.0120, 0.0072]],
[[ 0.0039, -0.0197, 0.0111],
[-0.0251, -0.0384, -0.0447],
[-0.0359, -0.0981, -0.0684]],
[[-0.0085, 0.0023, 0.0031],
[ 0.0038, 0.0182, 0.0069],
[ 0.0089, 0.0080, 0.0126]]],
[[[-0.0130, -0.0090, 0.0011],
[-0.0260, -0.0195, -0.0094],
[ 0.0048, 0.0024, 0.0106]],
[[-0.0025, -0.0140, -0.0286],
[-0.0024, 0.0011, -0.0233],
[ 0.0123, 0.0008, 0.0014]],
[[-0.0490, -0.0439, -0.0579],
[-0.0359, -0.0365, -0.0386],
[-0.0410, -0.0333, -0.0137]],
...,
[[-0.0118, -0.0081, -0.0158],
[-0.0272, -0.0285, 0.0075],
[-0.0244, 0.0139, 0.0061]],
[[-0.0273, 0.0197, 0.0222],
[-0.0376, 0.0208, 0.0187],
[-0.0446, -0.0044, -0.0168]],
[[ 0.0199, 0.0268, 0.0120],
[ 0.0203, 0.0215, 0.0025],
[ 0.0002, -0.0076, -0.0196]]]], device='cuda:0'),
Parameter containing:
tensor([0.2856, 0.2425, 0.3032, 0.3168, 0.3011, 0.3475, 0.3076, 0.3105, 0.3646,
0.3255, 0.2195, 0.3167, 0.2674, 0.3104, 0.3026, 0.3443, 0.2915, 0.3379,
0.2887, 0.2996, 0.3588, 0.3164, 0.2882, 0.2917, 0.3492, 0.3749, 0.3587,
0.3166, 0.2756, 0.2978, 0.3364, 0.2893, 0.3106, 0.2506, 0.3460, 0.3621,
0.2570, 0.3695, 0.2935, 0.3286, 0.3243, 0.3188, 0.3093, 0.3314, 0.3550,
0.2978, 0.2737, 0.3023, 0.3179, 0.2831, 0.3065, 0.3390, 0.3053, 0.3099,
0.3017, 0.3472, 0.3034, 0.2935, 0.3352, 0.3676, 0.3163, 0.3404, 0.3078,
0.2819, 0.3794, 0.3083, 0.2778, 0.3363, 0.2284, 0.3259, 0.2790, 0.3072,
0.2975, 0.3847, 0.3372, 0.2253, 0.2827, 0.3737, 0.2796, 0.3485, 0.3879,
0.3288, 0.3340, 0.3335, 0.2756, 0.3500, 0.2897, 0.2798, 0.2907, 0.3220,
0.3824, 0.3522, 0.3278, 0.3689, 0.3147, 0.3600, 0.3123, 0.2519, 0.2355,
0.3211, 0.3203, 0.3345, 0.2768, 0.3341, 0.3153, 0.3175, 0.2224, 0.2956,
0.3206, 0.2658, 0.3662, 0.2715, 0.3655, 0.3427, 0.2820, 0.2754, 0.4669,
0.3090, 0.3468, 0.3144, 0.3220, 0.2765, 0.3301, 0.3219, 0.3152, 0.2813,
0.2497, 0.3514, 0.3264, 0.3014, 0.2734, 0.3522, 0.3831, 0.3028, 0.2940,
0.2825, 0.3099, 0.2373, 0.2705, 0.4189, 0.2985, 0.3841, 0.2754, 0.3091,
0.3169, 0.2824, 0.2749, 0.3493, 0.4018, 0.3108, 0.2176, 0.2821, 0.3199,
0.3358, 0.2468, 0.3332, 0.2876, 0.2964, 0.2385, 0.3451, 0.3081, 0.2760,
0.2533, 0.2576, 0.3092, 0.2950, 0.3089, 0.3113, 0.3475, 0.3172, 0.2474,
0.3371, 0.3450, 0.3189, 0.3150, 0.3008, 0.2694, 0.3730, 0.3235, 0.2988,
0.2812, 0.3245, 0.3630, 0.2843, 0.3533, 0.3451, 0.3244, 0.3524, 0.3118,
0.3429, 0.3215, 0.2748, 0.3287, 0.3656, 0.2901, 0.2523, 0.3284, 0.2523,
0.3426, 0.2851, 0.2918, 0.2497, 0.5159, 0.3026, 0.2743, 0.2379, 0.3524,
0.3394, 0.2264, 0.2652, 0.3759, 0.3777, 0.2459, 0.3046, 0.3067, 0.3775,
0.2976, 0.3552, 0.2696, 0.2649, 0.2872, 0.2985, 0.2867, 0.3676, 0.3494,
0.3823, 0.3246, 0.3567, 0.2662, 0.3357, 0.2935, 0.2987, 0.2664, 0.3019,
0.3175, 0.2436, 0.3274, 0.2764, 0.2466, 0.2876, 0.3060, 0.3157, 0.3329,
0.2984, 0.2961, 0.3309, 0.3729, 0.3238, 0.3491, 0.3342, 0.3037, 0.3578,
0.2849, 0.2827, 0.2809, 0.3249], device='cuda:0'),
Parameter containing:
tensor([-0.0915, 0.0189, -0.1235, -0.0613, -0.1003, -0.1306, -0.1473, -0.1079,
-0.2438, -0.1113, 0.1361, -0.1477, 0.0387, -0.0907, 0.0352, -0.1851,
-0.1319, -0.1746, -0.0815, -0.1004, -0.3394, -0.1712, -0.0807, -0.1228,
-0.2263, -0.1503, -0.2314, -0.2327, -0.0854, -0.0802, -0.0716, -0.0839,
-0.0592, 0.0358, -0.0322, -0.2197, 0.0027, -0.1471, -0.0264, -0.1886,
-0.2417, -0.1494, -0.1904, -0.1089, -0.2657, -0.1362, -0.0487, -0.1340,
-0.0930, -0.0064, -0.1721, -0.1476, -0.1714, 0.0336, -0.1011, -0.1761,
-0.1184, -0.0482, -0.3260, -0.1555, -0.0169, -0.2373, -0.1015, -0.1051,
-0.2738, -0.1917, -0.0503, -0.1098, 0.1484, -0.2282, -0.0700, -0.1427,
-0.1417, -0.3096, -0.2043, 0.0269, -0.0779, -0.0842, -0.0464, -0.1429,
-0.3917, 0.0257, -0.1779, -0.0993, -0.0507, -0.2222, -0.0951, -0.0861,
-0.0743, -0.1666, -0.2054, -0.1782, -0.1150, -0.2525, -0.0694, -0.0536,
-0.0499, -0.0311, 0.1212, -0.0988, -0.1570, -0.3093, -0.0797, -0.0994,
-0.1774, -0.0505, 0.0766, -0.0480, -0.1278, -0.0651, -0.1737, 0.0303,
-0.1334, -0.2435, -0.0746, -0.0365, -0.1843, -0.0887, -0.1924, -0.1110,
-0.1458, -0.0895, -0.0956, -0.2042, -0.1338, -0.0637, -0.0699, -0.1656,
-0.1521, -0.1317, -0.0826, -0.2470, -0.1174, -0.1475, -0.0840, -0.0681,
-0.1789, 0.0288, -0.0362, -0.3005, -0.1441, -0.0812, -0.0492, -0.0657,
-0.1249, -0.1104, 0.0187, -0.1351, -0.1944, -0.0909, 0.2067, -0.1081,
-0.2499, -0.0999, 0.0507, -0.1899, -0.0369, -0.1432, 0.1279, -0.1782,
-0.1172, -0.0099, 0.0785, -0.0681, -0.0365, -0.1596, -0.1606, -0.0922,
-0.1773, -0.1788, 0.0306, -0.1101, -0.1355, -0.2244, -0.0860, -0.1232,
-0.0927, -0.1666, -0.1393, -0.0898, -0.0614, -0.1740, -0.2503, -0.0593,
-0.1272, -0.1422, -0.0743, -0.2208, -0.2207, -0.2742, -0.1302, -0.0916,
-0.1696, -0.2481, -0.1524, 0.0410, -0.1077, 0.0408, -0.1915, -0.0697,
-0.1049, -0.0110, -0.3257, -0.1336, -0.1021, 0.0128, -0.2717, -0.1245,
0.0288, -0.1025, -0.2405, -0.1476, 0.1008, -0.0220, -0.0983, -0.4417,
-0.0774, -0.3207, -0.0272, -0.0726, -0.0608, -0.0430, -0.0872, -0.1280,
-0.1608, -0.1529, -0.1745, -0.1702, -0.0486, -0.1459, -0.0552, -0.0808,
-0.0264, -0.0952, -0.1126, -0.0452, -0.0837, -0.0331, 0.0127, -0.0865,
-0.1446, -0.0732, -0.2160, -0.0952, -0.1297, -0.2008, -0.2135, -0.2204,
-0.2381, -0.1787, -0.1386, -0.1901, -0.0981, -0.0850, -0.0761, -0.0586],
device='cuda:0'),
Parameter containing:
tensor([[[[-0.0093, -0.0339, -0.0119],
[-0.0246, -0.0798, -0.0487],
[-0.0435, -0.0801, -0.0653]],
[[-0.0289, 0.0002, -0.0286],
[ 0.0099, 0.0103, -0.0177],
[-0.0107, 0.0028, -0.0125]],
[[-0.0147, 0.0226, 0.0044],
[ 0.0155, 0.0109, -0.0040],
[-0.0208, -0.0180, -0.0173]],
...,
[[ 0.0003, -0.0041, -0.0008],
[-0.0217, -0.0223, -0.0299],
[ 0.0105, 0.0035, -0.0114]],
[[ 0.0097, 0.0184, 0.0370],
[ 0.0037, 0.0104, 0.0152],
[ 0.0084, 0.0183, 0.0302]],
[[ 0.0014, 0.0084, 0.0097],
[ 0.0265, 0.0415, 0.0553],
[ 0.0169, 0.0610, 0.0563]]],
[[[ 0.0117, 0.0165, 0.0051],
[ 0.0294, 0.0204, 0.0216],
[ 0.0079, 0.0133, 0.0117]],
[[-0.0153, -0.0213, -0.0090],
[-0.0292, -0.0516, -0.0436],
[-0.0045, -0.0372, -0.0420]],
[[ 0.0003, 0.0398, -0.0001],
[ 0.0129, 0.0332, 0.0163],
[-0.0096, -0.0057, -0.0170]],
...,
[[ 0.0447, 0.0082, 0.0169],
[-0.0066, -0.0371, -0.0059],
[-0.0239, -0.0607, -0.0289]],
[[-0.0057, -0.0423, -0.0219],
[-0.0228, -0.0314, -0.0583],
[-0.0196, -0.0530, -0.0485]],
[[ 0.0065, 0.0033, 0.0093],
[ 0.0010, -0.0049, -0.0110],
[-0.0160, -0.0101, -0.0146]]],
[[[-0.0061, -0.0067, -0.0069],
[-0.0052, -0.0089, -0.0143],
[-0.0115, -0.0171, -0.0237]],
[[ 0.0399, 0.0167, 0.0210],
[ 0.0165, -0.0262, -0.0116],
[ 0.0059, -0.0206, -0.0153]],
[[ 0.0060, 0.0242, 0.0207],
[ 0.0050, -0.0062, 0.0148],
[ 0.0099, -0.0279, -0.0054]],
...,
[[-0.0097, -0.0017, -0.0101],
[-0.0006, 0.0434, 0.0381],
[ 0.0038, 0.0450, 0.0392]],
[[ 0.0134, 0.0069, 0.0196],
[ 0.0068, 0.0250, 0.0152],
[ 0.0018, -0.0044, 0.0037]],
[[-0.0174, -0.0163, -0.0240],
[-0.0197, -0.0174, -0.0178],
[-0.0300, -0.0137, -0.0211]]],
...,
[[[ 0.0066, 0.0061, 0.0284],
[-0.0130, -0.0509, -0.0204],
[-0.0154, -0.0149, -0.0181]],
[[ 0.0123, 0.0325, 0.0227],
[-0.0042, -0.0181, 0.0022],
[ 0.0029, -0.0174, 0.0032]],
[[-0.0110, 0.0111, -0.0142],
[ 0.0082, 0.0301, 0.0416],
[ 0.0006, 0.0002, 0.0213]],
...,
[[-0.0008, 0.0266, 0.0118],
[-0.0159, -0.0074, -0.0051],
[ 0.0042, -0.0069, -0.0073]],
[[-0.0070, 0.0010, 0.0019],
[ 0.0061, 0.0510, -0.0035],
[-0.0081, -0.0303, -0.0174]],
[[-0.0100, 0.0003, -0.0011],
[-0.0015, -0.0297, -0.0195],
[-0.0010, 0.0049, 0.0151]]],
[[[ 0.0161, 0.0126, 0.0055],
[-0.0132, -0.0196, -0.0208],
[-0.0077, -0.0230, -0.0202]],
[[-0.0019, 0.0530, 0.0134],
[ 0.0027, 0.0249, 0.0167],
[-0.0192, -0.0188, -0.0200]],
[[ 0.0081, 0.0300, 0.0006],
[ 0.0036, 0.0254, -0.0104],
[-0.0010, -0.0193, -0.0121]],
...,
[[ 0.0163, 0.0251, 0.0290],
[-0.0073, -0.0014, 0.0124],
[-0.0211, -0.0347, -0.0195]],
[[ 0.0165, 0.0519, 0.0494],
[ 0.0058, 0.0199, 0.0363],
[-0.0049, -0.0165, -0.0130]],
[[-0.0102, -0.0308, -0.0340],
[ 0.0055, -0.0109, 0.0005],
[ 0.0382, 0.0209, 0.0326]]],
[[[ 0.0184, -0.0031, 0.0103],
[-0.0072, -0.0013, -0.0070],
[ 0.0217, 0.0011, -0.0033]],
[[-0.0166, 0.0002, -0.0092],
[ 0.0150, 0.0584, 0.0218],
[-0.0136, 0.0181, 0.0164]],
[[ 0.0219, 0.0316, 0.0183],
[-0.0021, -0.0156, 0.0203],
[ 0.0053, 0.0005, 0.0157]],
...,
[[ 0.0045, -0.0025, 0.0017],
[-0.0052, -0.0088, 0.0137],
[-0.0107, 0.0010, 0.0170]],
[[-0.0060, -0.0368, 0.0031],
[ 0.0102, 0.0291, -0.0007],
[ 0.0099, 0.0600, 0.0272]],
[[-0.0092, -0.0183, 0.0062],
[-0.0319, -0.0294, -0.0149],
[-0.0148, -0.0123, -0.0236]]]], device='cuda:0'),
Parameter containing:
tensor([0.3212, 0.2124, 0.2661, 0.3594, 0.2785, 0.2582, 0.3108, 0.3096, 0.3348,
0.2992, 0.2545, 0.2458, 0.3133, 0.4159, 0.2997, 0.3070, 0.3135, 0.4418,
0.3743, 0.2570, 0.2943, 0.3078, 0.2738, 0.3948, 0.2928, 0.3572, 0.3435,
0.5379, 0.4243, 0.3908, 0.2745, 0.2798, 0.3217, 0.1956, 0.2751, 0.3187,
0.3507, 0.2751, 0.1919, 0.3307, 0.2850, 0.3038, 0.2179, 0.2652, 0.2944,
0.2138, 0.2184, 0.2948, 0.3262, 0.3759, 0.2557, 0.3796, 0.2950, 0.3386,
0.3243, 0.3070, 0.3331, 0.2302, 0.3036, 0.3377, 0.2922, 0.2204, 0.3267,
0.3198, 0.4023, 0.2987, 0.4860, 0.2854, 0.2716, 0.4341, 0.2834, 0.2296,
0.2507, 0.3120, 0.3673, 0.3244, 0.3380, 0.3272, 0.2868, 0.2877, 0.3210,
0.2332, 0.3379, 0.2767, 0.2942, 0.2672, 0.4401, 0.2908, 0.3771, 0.2789,
0.3056, 0.3276, 0.3871, 0.2453, 0.2559, 0.2783, 0.3168, 0.3410, 0.2318,
0.3577, 0.5036, 0.3557, 0.2475, 0.1852, 0.2273, 0.3602, 0.2919, 0.3928,
0.4423, 0.2052, 0.2524, 0.2189, 0.4113, 0.3611, 0.4284, 0.2333, 0.3504,
0.7001, 0.3754, 0.2874, 0.3702, 0.3174, 0.3640, 0.2889, 0.4155, 0.2479,
0.2898, 0.3740, 0.4926, 0.2808, 0.2388, 0.3473, 0.1868, 0.2837, 0.3090,
0.3614, 0.2797, 0.6871, 0.2854, 0.2937, 0.3128, 0.4863, 0.2193, 0.2871,
0.2554, 0.4175, 0.3044, 0.3230, 0.3343, 0.4947, 0.3924, 0.2264, 0.2657,
0.4193, 0.3483, 0.3551, 0.2877, 0.2559, 0.2459, 0.2775, 0.3842, 0.2949,
0.3510, 0.1926, 0.3101, 0.3417, 0.3931, 0.3918, 0.3239, 0.2851, 0.4583,
0.2669, 0.2663, 0.4433, 0.3221, 0.3655, 0.3336, 0.4393, 0.3970, 0.3727,
0.3523, 0.3586, 0.3286, 0.4181, 0.2955, 0.3050, 0.2988, 0.4320, 0.2309,
0.3826, 0.2270, 0.2228, 0.3206, 0.3273, 0.2627, 0.3087, 0.2920, 0.2328,
0.4144, 0.4075, 0.3264, 0.3583, 0.3014, 0.3150, 0.4438, 0.4042, 0.2028,
0.3855, 0.2570, 0.2361, 0.2343, 0.3312, 0.2303, 0.3744, 0.4727, 0.3601,
0.2754, 0.1987, 0.3027, 0.3427, 0.2994, 0.2533, 0.2639, 0.3460, 0.3847,
0.4368, 0.3786, 0.3123, 0.2591, 0.3979, 0.2577, 0.3131, 0.2934, 0.3027,
0.2942, 0.2266, 0.2806, 0.2977, 0.1858, 0.2788, 0.2504, 0.3948, 0.3496,
0.2429, 0.2155, 0.2683, 0.4100, 0.3495, 0.4243, 0.2627, 0.3329, 0.2849,
0.3924, 0.3728, 0.2655, 0.3338], device='cuda:0'),
Parameter containing:
tensor([-2.6385e-02, 9.9545e-02, -6.8362e-03, -8.7678e-02, 7.8322e-03,
4.0653e-02, -3.0738e-02, 5.9561e-03, 1.6938e-03, 4.7769e-02,
6.3004e-02, 3.5760e-02, -5.0405e-02, 2.1372e-02, -9.0327e-03,
-3.3702e-02, -4.5529e-02, -1.9238e-01, -6.7608e-02, 7.7464e-02,
-3.3957e-02, -7.9947e-02, 1.3138e-01, -1.2730e-01, -6.2825e-02,
-5.4973e-03, -9.1520e-02, -1.7570e-01, -8.2958e-03, -9.4468e-02,
2.4915e-03, -3.1894e-02, -1.5755e-02, 1.4372e-01, -3.4770e-03,
1.0756e-02, -5.1074e-02, 3.5848e-02, 8.7819e-02, -4.5240e-02,
-4.5785e-02, 1.4690e-02, 6.8685e-02, 1.6806e-02, -4.7715e-02,
5.6814e-02, 4.5985e-02, -5.0668e-02, 5.8914e-03, -1.0335e-01,
1.0281e-02, -1.0520e-01, -1.6557e-02, -1.9213e-02, -3.4522e-02,
2.0138e-02, -1.3621e-01, 3.9602e-02, -8.8277e-03, -1.0837e-02,
-2.9803e-02, 7.2119e-02, -6.6893e-02, -9.4287e-03, -3.0956e-02,
-2.6700e-02, -1.4184e-01, 1.1900e-01, 6.6929e-02, -2.1375e-01,
4.2720e-02, 4.7782e-02, 3.3905e-02, 1.3857e-04, -1.4822e-01,
-2.3684e-02, -7.4337e-02, -6.8427e-02, -2.0076e-02, 1.4697e-02,
-3.9620e-02, 1.9393e-02, -6.9640e-02, -5.5779e-02, 7.9664e-03,
2.3639e-02, -2.5781e-01, 6.3564e-03, -1.0041e-01, 2.8047e-02,
1.5223e-02, -4.8413e-02, -1.5355e-01, 1.0493e-01, 4.9892e-02,
6.5689e-02, -5.4139e-02, 7.7265e-03, 9.4090e-02, -1.9978e-02,
-2.3561e-01, -6.2284e-02, 3.3363e-02, 1.1017e-01, 7.6999e-02,
-3.2478e-02, 4.8089e-02, -1.4995e-01, -1.6503e-01, 1.2304e-01,
7.1157e-02, 5.8895e-02, -4.8155e-02, -9.7220e-02, -1.8604e-01,
8.5326e-02, -5.1602e-02, -3.0799e-01, -6.0410e-02, -7.7131e-02,
-2.7282e-01, 2.8950e-02, -1.3280e-01, 1.7264e-02, -3.9184e-02,
5.4153e-02, -3.7219e-02, -1.5283e-01, -1.7663e-01, 8.3937e-02,
6.9329e-02, -8.2645e-02, 1.1178e-01, -5.0772e-02, -4.4772e-02,
-3.7455e-02, 3.0361e-02, -3.7824e-01, 1.4857e-02, 6.8245e-03,
-5.2090e-02, -2.9501e-01, 8.9860e-02, 2.9646e-02, 1.9930e-02,
-8.3483e-02, -9.6416e-02, -2.3830e-02, 3.4911e-02, -2.6626e-01,
-1.6184e-01, 7.3552e-02, 2.7568e-02, -1.1087e-01, -1.0322e-02,
-9.7499e-02, 1.3995e-02, 1.0802e-02, 7.8444e-02, 1.3057e-02,
-3.9476e-02, 2.4805e-02, -7.7445e-02, -2.8419e-02, 1.0412e-02,
-4.2330e-02, -1.6634e-01, -9.4908e-02, -3.4328e-02, 4.5544e-02,
-3.0001e-01, -6.9376e-03, 1.4103e-02, -2.6155e-01, -7.3598e-02,
-1.0627e-01, -1.0544e-02, -7.1160e-02, -1.0339e-01, -2.9787e-02,
-1.4277e-01, -5.1710e-02, -5.7126e-02, -5.4392e-02, -4.2260e-02,
-8.5120e-03, 1.5887e-02, -6.5358e-02, -6.1346e-02, -1.4502e-01,
3.9939e-02, 8.1640e-02, -7.7597e-03, -3.4062e-02, 3.1969e-02,
-4.4756e-02, -7.0306e-02, 1.0213e-01, -1.7993e-01, -2.1173e-01,
-5.9797e-02, -1.1596e-01, 3.9271e-02, -4.5443e-02, -1.8446e-01,
-1.0848e-01, 5.5781e-02, -6.3649e-02, 1.6825e-02, 1.6623e-04,
7.9866e-02, -6.7240e-02, 7.9827e-02, -3.9905e-03, -1.9016e-01,
2.0026e-02, 7.3181e-02, 1.0323e-01, -2.6431e-02, 2.3963e-02,
-4.4247e-02, 2.2896e-02, 2.3444e-02, -2.3481e-02, 1.0516e-02,
-2.1494e-01, -1.2810e-01, -1.8338e-02, -6.0221e-04, -5.1624e-02,
5.6575e-02, -5.4297e-02, 1.4146e-02, -4.9852e-02, 6.7255e-02,
5.1713e-02, -4.0266e-03, 3.5057e-02, 8.2781e-02, 1.0034e-02,
5.9230e-02, -2.0429e-01, -7.6169e-02, 4.1432e-02, 7.7517e-02,
7.6005e-02, -1.5919e-01, -8.3613e-02, -1.6625e-01, 2.2968e-03,
-6.8489e-02, 3.8119e-02, -9.8731e-02, -2.0290e-02, 1.5394e-02,
-1.0548e-01], device='cuda:0'),
Parameter containing:
tensor([[[[ 0.0081]],
[[-0.0192]],
[[-0.0173]],
...,
[[-0.0128]],
[[ 0.0025]],
[[ 0.0054]]],
[[[-0.0143]],
[[-0.0554]],
[[-0.0346]],
...,
[[ 0.0275]],
[[ 0.0360]],
[[ 0.0240]]],
[[[ 0.0076]],
[[ 0.0207]],
[[-0.0101]],
...,
[[-0.0278]],
[[ 0.0064]],
[[-0.0022]]],
...,
[[[ 0.0336]],
[[-0.0424]],
[[ 0.0226]],
...,
[[-0.0330]],
[[-0.0009]],
[[-0.0177]]],
[[[-0.0114]],
[[-0.0183]],
[[ 0.0076]],
...,
[[ 0.0151]],
[[ 0.0332]],
[[ 0.0002]]],
[[[ 0.0063]],
[[-0.0200]],
[[ 0.0010]],
...,
[[-0.0191]],
[[ 0.0455]],
[[ 0.0078]]]], device='cuda:0'),
Parameter containing:
tensor([0.0674, 0.0514, 0.0385, 0.1692, 0.0604, 0.0460, 0.1209, 0.1110, 0.0418,
0.0387, 0.0442, 0.0707, 0.0790, 0.1094, 0.0959, 0.0544, 0.1032, 0.2190,
0.0459, 0.0372, 0.1410, 0.0587, 0.0360, 0.0955, 0.1657, 0.1024, 0.1417,
0.0580, 0.0536, 0.0716, 0.0865, 0.1110, 0.0511, 0.0515, 0.0809, 0.1154,
0.0777, 0.0449, 0.0490, 0.1056, 0.1457, 0.0744, 0.0530, 0.0600, 0.1026,
0.0486, 0.0408, 0.1312, 0.0639, 0.1062, 0.0915, 0.1476, 0.0900, 0.0742,
0.1069, 0.0776, 0.1423, 0.0495, 0.0974, 0.0661, 0.1292, 0.0548, 0.1145,
0.0950, 0.0921, 0.1579, 0.0496, 0.0236, 0.0398, 0.0935, 0.0291, 0.0653,
0.0885, 0.1190, 0.1692, 0.0692, 0.1316, 0.0606, 0.0480, 0.0654, 0.1082,
0.0624, 0.1103, 0.1106, 0.1076, 0.0400, 0.0723, 0.0947, 0.0662, 0.0464,
0.0444, 0.1727, 0.0921, 0.0345, 0.0451, 0.0374, 0.0940, 0.0818, 0.0397,
0.0452, 0.0985, 0.1095, 0.1072, 0.0506, 0.0444, 0.0755, 0.0420, 0.1046,
0.1172, 0.0447, 0.0459, 0.0409, 0.0539, 0.1036, 0.0741, 0.0311, 0.1086,
0.1746, 0.0777, 0.0689, 0.1100, 0.0489, 0.1048, 0.1097, 0.1025, 0.0448,
0.0675, 0.0707, 0.1364, 0.0438, 0.0346, 0.1769, 0.0667, 0.1155, 0.0628,
0.0873, 0.0406, 0.2890, 0.0703, 0.0428, 0.1173, 0.1049, 0.0611, 0.0469,
0.0400, 0.0744, 0.1003, 0.1012, 0.0599, 0.1078, 0.1512, 0.0322, 0.0430,
0.0977, 0.0951, 0.0838, 0.0958, 0.0448, 0.0263, 0.0425, 0.1154, 0.0771,
0.1781, 0.0300, 0.0699, 0.0724, 0.1600, 0.0893, 0.1130, 0.0534, 0.1359,
0.0375, 0.0809, 0.1145, 0.1232, 0.0942, 0.0880, 0.0346, 0.0996, 0.0461,
0.0694, 0.0630, 0.1590, 0.0509, 0.1254, 0.0590, 0.0744, 0.1084, 0.0514,
0.0931, 0.0848, 0.0240, 0.0279, 0.0993, 0.0612, 0.0599, 0.1095, 0.0508,
0.0658, 0.1162, 0.0833, 0.1651, 0.0505, 0.1231, 0.1228, 0.1038, 0.0369,
0.0756, 0.0415, 0.1192, 0.0292, 0.0839, 0.0577, 0.0951, 0.0944, 0.0309,
0.0390, 0.0604, 0.0672, 0.0501, 0.0383, 0.0946, 0.0958, 0.0501, 0.0243,
0.1074, 0.1908, 0.0693, 0.1376, 0.1151, 0.0329, 0.0647, 0.0616, 0.1106,
0.0358, 0.0721, 0.0851, 0.0375, 0.0368, 0.0947, 0.0464, 0.1666, 0.1049,
0.0755, 0.0398, 0.0249, 0.1528, 0.1167, 0.0886, 0.0540, 0.0726, 0.0736,
0.0797, 0.0854, 0.0609, 0.1263], device='cuda:0'),
Parameter containing:
tensor([-2.6385e-02, 9.9545e-02, -6.8362e-03, -8.7678e-02, 7.8322e-03,
4.0653e-02, -3.0738e-02, 5.9561e-03, 1.6938e-03, 4.7769e-02,
6.3004e-02, 3.5760e-02, -5.0405e-02, 2.1372e-02, -9.0327e-03,
-3.3702e-02, -4.5529e-02, -1.9238e-01, -6.7608e-02, 7.7464e-02,
-3.3957e-02, -7.9947e-02, 1.3138e-01, -1.2730e-01, -6.2825e-02,
-5.4973e-03, -9.1520e-02, -1.7570e-01, -8.2958e-03, -9.4468e-02,
2.4915e-03, -3.1894e-02, -1.5755e-02, 1.4372e-01, -3.4770e-03,
1.0756e-02, -5.1074e-02, 3.5848e-02, 8.7819e-02, -4.5240e-02,
-4.5785e-02, 1.4690e-02, 6.8685e-02, 1.6806e-02, -4.7715e-02,
5.6814e-02, 4.5985e-02, -5.0668e-02, 5.8914e-03, -1.0335e-01,
1.0281e-02, -1.0520e-01, -1.6557e-02, -1.9213e-02, -3.4522e-02,
2.0138e-02, -1.3621e-01, 3.9602e-02, -8.8277e-03, -1.0837e-02,
-2.9803e-02, 7.2119e-02, -6.6893e-02, -9.4287e-03, -3.0956e-02,
-2.6700e-02, -1.4184e-01, 1.1900e-01, 6.6929e-02, -2.1375e-01,
4.2720e-02, 4.7782e-02, 3.3905e-02, 1.3857e-04, -1.4822e-01,
-2.3684e-02, -7.4337e-02, -6.8427e-02, -2.0076e-02, 1.4697e-02,
-3.9620e-02, 1.9393e-02, -6.9640e-02, -5.5779e-02, 7.9664e-03,
2.3639e-02, -2.5781e-01, 6.3564e-03, -1.0041e-01, 2.8047e-02,
1.5223e-02, -4.8413e-02, -1.5355e-01, 1.0493e-01, 4.9892e-02,
6.5689e-02, -5.4139e-02, 7.7265e-03, 9.4090e-02, -1.9978e-02,
-2.3561e-01, -6.2284e-02, 3.3363e-02, 1.1017e-01, 7.6999e-02,
-3.2478e-02, 4.8089e-02, -1.4995e-01, -1.6503e-01, 1.2304e-01,
7.1157e-02, 5.8895e-02, -4.8155e-02, -9.7220e-02, -1.8604e-01,
8.5326e-02, -5.1602e-02, -3.0799e-01, -6.0410e-02, -7.7131e-02,
-2.7282e-01, 2.8950e-02, -1.3280e-01, 1.7264e-02, -3.9184e-02,
5.4153e-02, -3.7219e-02, -1.5283e-01, -1.7663e-01, 8.3937e-02,
6.9329e-02, -8.2645e-02, 1.1178e-01, -5.0772e-02, -4.4772e-02,
-3.7455e-02, 3.0361e-02, -3.7824e-01, 1.4857e-02, 6.8245e-03,
-5.2090e-02, -2.9501e-01, 8.9860e-02, 2.9646e-02, 1.9930e-02,
-8.3483e-02, -9.6416e-02, -2.3830e-02, 3.4911e-02, -2.6626e-01,
-1.6184e-01, 7.3552e-02, 2.7568e-02, -1.1087e-01, -1.0322e-02,
-9.7499e-02, 1.3995e-02, 1.0802e-02, 7.8444e-02, 1.3057e-02,
-3.9476e-02, 2.4805e-02, -7.7445e-02, -2.8419e-02, 1.0412e-02,
-4.2330e-02, -1.6634e-01, -9.4908e-02, -3.4328e-02, 4.5544e-02,
-3.0001e-01, -6.9376e-03, 1.4103e-02, -2.6155e-01, -7.3598e-02,
-1.0627e-01, -1.0544e-02, -7.1160e-02, -1.0339e-01, -2.9787e-02,
-1.4277e-01, -5.1710e-02, -5.7126e-02, -5.4392e-02, -4.2260e-02,
-8.5120e-03, 1.5887e-02, -6.5358e-02, -6.1346e-02, -1.4502e-01,
3.9939e-02, 8.1640e-02, -7.7597e-03, -3.4062e-02, 3.1969e-02,
-4.4756e-02, -7.0306e-02, 1.0213e-01, -1.7993e-01, -2.1173e-01,
-5.9797e-02, -1.1596e-01, 3.9271e-02, -4.5443e-02, -1.8446e-01,
-1.0848e-01, 5.5781e-02, -6.3649e-02, 1.6825e-02, 1.6623e-04,
7.9866e-02, -6.7240e-02, 7.9827e-02, -3.9905e-03, -1.9016e-01,
2.0026e-02, 7.3181e-02, 1.0323e-01, -2.6431e-02, 2.3963e-02,
-4.4247e-02, 2.2896e-02, 2.3444e-02, -2.3481e-02, 1.0516e-02,
-2.1494e-01, -1.2810e-01, -1.8338e-02, -6.0221e-04, -5.1624e-02,
5.6575e-02, -5.4297e-02, 1.4146e-02, -4.9852e-02, 6.7255e-02,
5.1713e-02, -4.0266e-03, 3.5057e-02, 8.2781e-02, 1.0034e-02,
5.9230e-02, -2.0429e-01, -7.6169e-02, 4.1432e-02, 7.7517e-02,
7.6005e-02, -1.5919e-01, -8.3613e-02, -1.6625e-01, 2.2968e-03,
-6.8489e-02, 3.8119e-02, -9.8731e-02, -2.0290e-02, 1.5394e-02,
-1.0548e-01], device='cuda:0'),
Parameter containing:
tensor([[[[ 4.8367e-02, 4.8045e-02, 3.8471e-02],
[ 4.9888e-02, 5.5208e-02, 5.6701e-02],
[ 2.4192e-02, 1.3436e-02, 2.4655e-02]],
[[-3.6542e-03, -3.1100e-03, 4.9227e-03],
[-1.2114e-03, 3.4020e-03, 1.9846e-02],
[-2.1704e-02, -2.1158e-02, -2.8686e-03]],
[[-1.2536e-02, -2.0486e-02, -2.3154e-02],
[-1.3515e-02, -2.3781e-02, -2.5515e-02],
[ 1.0584e-02, 7.2999e-03, -5.2329e-03]],
...,
[[-4.3596e-02, -1.8328e-02, -5.0577e-02],
[ 1.6590e-02, 5.0719e-02, 2.1919e-02],
[-1.9203e-02, -8.8315e-03, -2.0335e-02]],
[[-7.6949e-03, -1.5848e-02, 1.5841e-03],
[-6.2470e-03, -1.3135e-02, 6.9092e-03],
[-3.3791e-03, 1.7889e-03, 3.7373e-03]],
[[-6.6310e-03, 5.8503e-03, -5.8571e-04],
[-2.4600e-02, -8.9747e-03, -7.2466e-03],
[-1.7566e-02, -8.5829e-03, -7.5220e-03]]],
[[[-2.3679e-02, -9.4399e-03, -1.1688e-02],
[-2.4777e-02, -1.7326e-02, -3.1489e-02],
[-3.3683e-03, 9.7571e-03, -5.1527e-03]],
[[-3.0809e-02, -4.0685e-02, -2.2731e-02],
[-5.1065e-03, -1.6457e-02, -1.8804e-02],
[ 5.0382e-02, 5.2054e-02, 3.9185e-02]],
[[-3.7790e-02, -4.2234e-02, -2.9703e-02],
[-6.4766e-03, 2.6967e-03, -8.1736e-03],
[ 3.7747e-02, 5.5416e-02, 2.5806e-02]],
...,
[[-2.7275e-02, -4.5364e-02, -3.9567e-02],
[ 8.9827e-03, 1.6150e-02, 1.1675e-02],
[-9.7209e-03, -3.6449e-02, -1.6842e-02]],
[[ 1.7824e-02, 1.5013e-02, 1.0225e-02],
[ 5.4044e-03, 1.1664e-02, 6.4623e-03],
[ 2.1803e-02, 4.1795e-02, 1.9234e-02]],
[[-2.6730e-04, 1.5218e-03, -5.0352e-03],
[ 2.5761e-02, 2.7110e-02, -9.3395e-04],
[-1.1949e-02, -7.5204e-03, -3.9370e-03]]],
[[[-1.7447e-02, -1.8358e-02, -2.6020e-02],
[-1.4074e-02, -1.1302e-02, -1.4814e-02],
[-3.1460e-03, -1.8674e-02, -9.3350e-03]],
[[-5.1125e-03, -4.8036e-03, 1.8139e-02],
[-1.0524e-02, -1.5152e-02, 2.3904e-03],
[ 8.7093e-03, 9.3810e-03, 2.4203e-03]],
[[-7.6392e-03, -8.1496e-03, -1.5331e-02],
[-8.0622e-03, -1.3383e-02, -1.3938e-02],
[-1.6904e-02, -3.0059e-02, -1.8659e-02]],
...,
[[ 1.8390e-02, -2.6080e-03, 9.3782e-03],
[-6.4662e-04, -1.3146e-02, 1.0045e-02],
[-2.2293e-03, -1.4097e-02, 1.7385e-02]],
[[ 3.0293e-04, 2.9622e-03, 1.0030e-02],
[-5.7588e-03, -1.6943e-03, 6.9988e-03],
[ 9.8134e-03, 1.4197e-02, 5.9742e-03]],
[[ 2.8753e-03, -1.7814e-03, 1.0873e-02],
[ 1.5230e-02, 4.5867e-03, 1.6860e-02],
[ 1.9536e-03, 1.9503e-02, 1.2168e-02]]],
...,
[[[ 1.3983e-02, 2.4598e-03, -7.4604e-03],
[-2.2250e-02, -1.2757e-02, -2.8846e-03],
[-1.0911e-02, 7.5499e-03, 8.6910e-03]],
[[-4.8463e-03, -8.3250e-03, 1.3420e-02],
[-6.2502e-03, -7.3982e-03, 1.1153e-02],
[ 4.0391e-03, -9.0354e-03, -7.5441e-03]],
[[-5.1627e-03, -8.9529e-03, -1.2414e-02],
[-4.9261e-03, -3.5488e-03, 2.1501e-03],
[-1.1709e-02, -1.4984e-02, -1.9216e-03]],
...,
[[ 1.5428e-02, -7.6036e-04, -1.3522e-03],
[-3.4856e-02, -7.4478e-04, -6.5064e-03],
[-9.1655e-03, -2.8467e-02, -4.8924e-02]],
[[ 1.2207e-02, 1.0519e-02, -8.4421e-03],
[-2.5495e-02, 2.8140e-03, 1.6165e-03],
[-1.8831e-02, 1.2268e-02, 1.5439e-02]],
[[-1.3684e-02, -4.1732e-03, 1.2609e-02],
[-6.8834e-04, 5.9757e-03, -1.0183e-02],
[ 2.1559e-04, -1.3462e-02, -3.0114e-02]]],
[[[-1.6186e-02, -6.4926e-02, -4.3146e-02],
[-2.1790e-02, -4.9106e-02, -3.4568e-02],
[ 4.0506e-02, 4.2449e-02, 6.1562e-02]],
[[ 3.5715e-03, -1.0916e-02, -2.2922e-02],
[-2.4831e-03, 6.4555e-03, -1.1316e-02],
[ 1.6662e-03, -1.9145e-02, -2.3007e-02]],
[[-7.1243e-03, -4.2783e-05, 4.9363e-03],
[-1.5832e-02, 4.0474e-03, 4.5135e-04],
[-4.7967e-03, -7.2164e-04, -1.7230e-02]],
...,
[[ 1.1589e-02, 7.7814e-04, 6.3205e-03],
[ 1.1360e-02, -6.2076e-03, -2.7689e-02],
[ 2.6392e-02, 2.3775e-03, -1.4937e-02]],
[[-1.1237e-02, -2.6285e-03, 9.1537e-03],
[-8.2120e-03, -2.2236e-02, 3.2917e-04],
[ 5.5909e-03, -1.3858e-03, 6.8947e-03]],
[[-1.4783e-02, -1.0367e-02, -2.7472e-02],
[-4.1090e-02, -3.8532e-02, -3.9202e-02],
[-2.1614e-02, -3.4340e-02, -1.8542e-02]]],
[[[-1.9492e-02, -1.6098e-02, -3.1792e-02],
[ 2.5374e-02, 4.6815e-02, 2.7513e-02],
[ 3.5903e-02, 3.1892e-02, 2.6156e-02]],
[[ 1.6856e-02, 1.5645e-02, 1.4189e-02],
[ 2.2550e-02, 3.0456e-02, 1.6739e-02],
[-2.3615e-04, -7.9501e-03, -1.9666e-03]],
[[-7.9060e-03, -4.7390e-03, 1.6030e-03],
[ 1.3802e-03, -8.5837e-03, 6.9451e-03],
[ 1.1407e-02, -5.9877e-03, 1.3759e-02]],
...,
[[ 4.0124e-03, 2.9951e-02, 1.1915e-02],
[-4.3412e-02, -3.1776e-03, -2.7705e-02],
[-1.6183e-02, -1.1247e-02, -3.5084e-02]],
[[ 2.9837e-02, 5.9935e-02, 2.4631e-02],
[-1.9571e-03, 2.2415e-02, -1.5499e-02],
[ 1.6075e-02, 1.7850e-02, -1.8412e-02]],
[[-4.3712e-03, -4.9032e-02, -2.1335e-02],
[-5.2598e-03, -2.8579e-02, -2.2090e-02],
[ 8.5126e-03, 2.0862e-03, 2.3301e-02]]]], device='cuda:0'),
Parameter containing:
tensor([0.2480, 0.1972, 0.2279, 0.2709, 0.3296, 0.2640, 0.2710, 0.3475, 0.2388,
0.2904, 0.2769, 0.3045, 0.2268, 0.2634, 0.2999, 0.2397, 0.2724, 0.2723,
0.2133, 0.3806, 0.2767, 0.2403, 0.2406, 0.2917, 0.2675, 0.2305, 0.2394,
0.3123, 0.2984, 0.3353, 0.2234, 0.1919, 0.3168, 0.2626, 0.2901, 0.2918,
0.3455, 0.2561, 0.2434, 0.2298, 0.3318, 0.3481, 0.2032, 0.2478, 0.2478,
0.2483, 0.3252, 0.2567, 0.2685, 0.1977, 0.2541, 0.4079, 0.2480, 0.2076,
0.2276, 0.2683, 0.2098, 0.2056, 0.2010, 0.3560, 0.2384, 0.3284, 0.1952,
0.2445, 0.2848, 0.3742, 0.2746, 0.2117, 0.3859, 0.4785, 0.3005, 0.2848,
0.3762, 0.2903, 0.2126, 0.1776, 0.2778, 0.3878, 0.3123, 0.1974, 0.2679,
0.2300, 0.2474, 0.2320, 0.2635, 0.2819, 0.2296, 0.3194, 0.3814, 0.2503,
0.2269, 0.2676, 0.3431, 0.3799, 0.3787, 0.2968, 0.3021, 0.2575, 0.3007,
0.1939, 0.1950, 0.3217, 0.3623, 0.2171, 0.2486, 0.2266, 0.2133, 0.2851,
0.2715, 0.2720, 0.3107, 0.2174, 0.2675, 0.2387, 0.3434, 0.2761, 0.2084,
0.2975, 0.3178, 0.2818, 0.2858, 0.3498, 0.2675, 0.2638, 0.3159, 0.2879,
0.1873, 0.2986, 0.3584, 0.2570, 0.1815, 0.2758, 0.2640, 0.2486, 0.2567,
0.2252, 0.3420, 0.2910, 0.2898, 0.2902, 0.2404, 0.2381, 0.3633, 0.2690,
0.3810, 0.2947, 0.2743, 0.4644, 0.3133, 0.2444, 0.3477, 0.3001, 0.1977,
0.2301, 0.2513, 0.2660, 0.3271, 0.1622, 0.2274, 0.2225, 0.3596, 0.3215,
0.1997, 0.2215, 0.2706, 0.2831, 0.2621, 0.3710, 0.2730, 0.2903, 0.1893,
0.2140, 0.2460, 0.3141, 0.2424, 0.3699, 0.2364, 0.2420, 0.2948, 0.2497,
0.2760, 0.2686, 0.2895, 0.3857, 0.1398, 0.2832, 0.3362, 0.2522, 0.2823,
0.2381, 0.2311, 0.3274, 0.4078, 0.2648, 0.2525, 0.3388, 0.3251, 0.2420,
0.2856, 0.3605, 0.2603, 0.2294, 0.2483, 0.2171, 0.2353, 0.4117, 0.2588,
0.2888, 0.1972, 0.2408, 0.2755, 0.3031, 0.2457, 0.2744, 0.3564, 0.2546,
0.3673, 0.2883, 0.2590, 0.3021, 0.2890, 0.3505, 0.2092, 0.2953, 0.3222,
0.2925, 0.2574, 0.3012, 0.3893, 0.2211, 0.2226, 0.3258, 0.3205, 0.2975,
0.2323, 0.3323, 0.2812, 0.2702, 0.2300, 0.2846, 0.3318, 0.2292, 0.3498,
0.2622, 0.3581, 0.4003, 0.2924, 0.3049, 0.3478, 0.2845, 0.2742, 0.2019,
0.2466, 0.2988, 0.2044, 0.2691], device='cuda:0'),
Parameter containing:
tensor([-0.1332, -0.0644, -0.3239, -0.2390, -0.3262, -0.1796, -0.2087, -0.3208,
-0.1874, -0.2988, -0.2099, -0.2283, -0.2141, -0.2460, -0.2768, -0.1351,
-0.2498, -0.2393, -0.1223, -0.4590, -0.2172, -0.1220, -0.2101, -0.1779,
-0.2426, -0.1546, -0.1549, -0.3716, -0.2817, -0.3886, -0.1545, -0.0687,
-0.3412, -0.2261, -0.1961, -0.2242, -0.2984, -0.1381, -0.2251, -0.1658,
-0.4534, -0.3226, -0.0977, -0.1349, -0.2619, -0.1428, -0.3960, -0.1633,
-0.2101, -0.1161, -0.1448, -0.5502, -0.2179, -0.1246, 0.0502, -0.1902,
-0.1047, -0.1000, -0.1411, -0.3124, -0.2190, -0.3062, -0.1247, -0.1557,
-0.2973, -0.3825, -0.1951, -0.1381, -0.5761, -0.3879, -0.2808, -0.2542,
-0.3470, -0.2460, -0.1091, -0.0562, -0.1833, -0.4956, -0.3059, -0.0988,
-0.2255, -0.1958, -0.1320, -0.1738, -0.2287, -0.1926, -0.0924, -0.3427,
-0.5489, -0.2431, -0.1935, -0.1641, -0.2503, -0.3274, -0.4008, -0.2824,
-0.2694, -0.1939, -0.2413, -0.0309, -0.0880, -0.3421, -0.3104, -0.1102,
-0.1539, -0.1233, -0.1780, -0.2715, -0.2005, -0.1846, -0.2843, -0.1117,
-0.1816, -0.2119, -0.3304, -0.2267, -0.1413, -0.3376, -0.2674, -0.2524,
-0.2554, -0.4735, -0.2342, -0.2130, -0.3282, -0.1966, -0.1063, -0.2615,
-0.4234, -0.1374, -0.0811, -0.3069, -0.1538, -0.1453, -0.1612, -0.1631,
-0.3759, -0.2608, -0.2382, -0.2499, -0.1485, -0.1487, -0.4328, -0.1377,
-0.2781, -0.2259, -0.2072, -0.4165, -0.3582, -0.1382, -0.3598, -0.2672,
-0.2090, -0.0177, -0.1279, -0.2812, -0.3621, 0.0476, -0.2232, -0.1272,
-0.3237, -0.3008, -0.1119, -0.0839, -0.2426, -0.2000, -0.1873, -0.4685,
-0.2000, -0.3462, -0.0706, -0.1973, -0.3548, -0.1975, -0.3537, -0.3546,
-0.1433, -0.2052, -0.2722, -0.1528, -0.2798, -0.1945, -0.2474, -0.4910,
0.1322, -0.2378, -0.5166, -0.3959, -0.2354, -0.1266, -0.0810, -0.4132,
-0.5576, -0.2238, -0.1563, -0.3950, -0.3283, -0.0846, -0.3103, -0.3130,
-0.1498, -0.1396, -0.0972, -0.1620, -0.1631, -0.6364, -0.1350, -0.2345,
-0.1049, -0.1625, -0.2878, -0.2450, -0.1468, -0.2035, -0.5358, -0.1683,
-0.5524, -0.2511, -0.1230, -0.2305, -0.1925, -0.3759, -0.1014, -0.1697,
-0.4002, -0.2980, -0.3035, -0.1563, -0.4660, -0.1155, -0.1665, -0.3382,
-0.2935, -0.3122, -0.3015, -0.3261, -0.2542, -0.2037, -0.0955, -0.2070,
-0.4370, -0.2051, -0.4205, -0.3125, -0.4845, -0.3528, -0.2624, -0.2894,
-0.3976, -0.2107, -0.1791, -0.1075, -0.1213, -0.3022, 0.0516, -0.1928],
device='cuda:0'),
Parameter containing:
tensor([[[[-4.2568e-02, -2.6148e-02, -2.2019e-02],
[-1.7334e-02, -7.5950e-03, -7.2384e-03],
[-1.7876e-03, 2.3800e-02, 1.4873e-02]],
[[-2.8277e-03, -5.0644e-03, -4.9442e-03],
[ 1.2117e-03, 1.4908e-02, 1.6013e-02],
[ 1.4391e-02, 3.3109e-02, 5.0061e-02]],
[[-3.4891e-03, -4.4437e-03, 2.6589e-03],
[ 1.5105e-02, 2.6303e-02, 2.6802e-02],
[ 3.9232e-02, 5.0057e-02, 4.6637e-02]],
...,
[[ 2.2877e-02, 1.5454e-02, -2.4483e-02],
[ 3.1145e-02, 3.4944e-02, 1.3296e-02],
[-1.7674e-04, 7.3297e-03, -5.7174e-03]],
[[-2.1781e-02, -3.7379e-02, -1.3382e-02],
[ 1.8976e-02, 1.4155e-02, -6.5395e-03],
[ 2.6831e-02, 3.6354e-02, 1.1450e-02]],
[[ 3.1603e-02, 3.3933e-02, 3.1575e-02],
[-1.0098e-02, -1.2657e-02, 1.1674e-02],
[ 1.0325e-02, 7.9424e-05, 1.5911e-02]]],
[[[ 2.5937e-02, 6.2590e-03, 6.0798e-03],
[-4.5745e-03, -3.5188e-02, -2.9249e-02],
[ 2.1366e-02, 2.0480e-03, 6.2699e-03]],
[[-2.9549e-03, -1.3679e-03, -8.6876e-03],
[ 7.9988e-03, 1.2888e-03, -5.9629e-03],
[-9.1481e-03, -2.1914e-02, -4.1572e-02]],
[[ 7.6390e-03, 3.0253e-03, 2.7817e-04],
[ 7.0329e-03, 1.1914e-02, -2.4419e-03],
[-8.2131e-03, -9.7848e-05, -1.9223e-02]],
...,
[[-4.4498e-03, 5.1611e-03, 3.7416e-03],
[ 3.2110e-04, 8.3762e-03, 3.6612e-03],
[ 9.3343e-03, 8.1829e-03, 1.1234e-03]],
[[-6.6849e-02, -5.9871e-02, -3.3931e-02],
[ 2.2337e-02, 3.1932e-02, 3.7244e-02],
[ 9.3296e-03, 3.7222e-02, 1.4052e-02]],
[[-2.0643e-03, 1.2408e-02, -3.1072e-03],
[-8.2882e-03, 1.3917e-02, -2.0680e-02],
[-1.9329e-02, 1.1953e-02, -2.3436e-02]]],
[[[-1.7788e-03, -3.5982e-03, -1.2592e-03],
[-1.5320e-02, -1.0690e-02, -2.0311e-02],
[-3.4649e-04, -2.2188e-03, -1.5021e-02]],
[[-2.8952e-02, -3.3958e-02, -2.5437e-02],
[-1.5919e-04, 1.5204e-02, 3.4554e-02],
[ 3.6892e-02, 7.0144e-02, 7.3610e-02]],
[[ 1.0721e-02, 2.1531e-03, -5.6155e-03],
[ 1.1754e-02, -4.8546e-03, -5.5013e-03],
[-3.7388e-04, -9.7639e-03, -1.5029e-02]],
...,
[[ 1.5622e-02, 9.8976e-03, 3.4725e-03],
[ 1.4711e-02, 7.0707e-03, -9.1826e-03],
[ 7.0986e-03, 6.3087e-03, -3.5893e-03]],
[[-6.4518e-03, -6.7673e-03, 1.1635e-02],
[ 1.4707e-02, 2.3831e-02, 4.9396e-02],
[ 1.8897e-02, 3.4981e-02, 4.5488e-02]],
[[ 1.5900e-02, 3.3369e-02, 2.6194e-02],
[ 1.0616e-02, 1.8515e-02, 3.0190e-03],
[ 1.1004e-02, 2.5503e-02, 1.3654e-02]]],
...,
[[[-2.1231e-02, -1.2804e-02, -1.5498e-02],
[ 7.6750e-03, 1.2120e-02, 1.5099e-02],
[ 1.8536e-02, 2.5110e-02, 2.5283e-02]],
[[ 7.4059e-03, -3.0540e-03, -1.5475e-03],
[-8.4415e-03, -2.2002e-02, -3.4099e-03],
[ 9.1918e-03, 2.2617e-03, -1.4260e-02]],
[[-5.2568e-03, -5.3507e-03, -3.2230e-03],
[-1.5805e-02, 6.0508e-03, -1.5917e-03],
[-8.9323e-03, 2.6483e-03, 5.0508e-03]],
...,
[[ 1.9826e-02, -2.1209e-03, 1.4889e-02],
[ 5.7275e-02, 3.5549e-02, 6.0175e-03],
[ 2.3347e-02, -2.2153e-02, -2.5497e-02]],
[[-1.3985e-02, -6.4766e-02, -1.7286e-02],
[ 1.1704e-02, 1.0714e-02, 4.6278e-02],
[-1.0038e-02, -3.5707e-03, 2.2691e-02]],
[[-8.3342e-03, -1.3070e-03, -1.0049e-02],
[ 3.2605e-02, 5.3259e-02, 2.2172e-02],
[ 3.7339e-02, 6.1155e-02, 4.4555e-02]]],
[[[-1.6584e-02, -1.3850e-02, -1.4604e-02],
[-1.7604e-02, -2.1268e-02, -1.6734e-02],
[-6.0039e-04, 3.8569e-03, 1.2837e-02]],
[[ 1.7623e-02, 2.3706e-02, 2.7633e-02],
[-2.2841e-02, -1.9576e-02, -1.6551e-02],
[-8.0822e-03, 4.3779e-03, -5.3622e-03]],
[[ 1.5582e-02, 3.7879e-02, 2.3555e-02],
[-6.4632e-03, 9.8620e-03, 1.2121e-02],
[-1.3743e-02, -6.1246e-03, -2.7332e-03]],
...,
[[-2.5037e-03, -1.2064e-02, -9.0989e-03],
[-4.7911e-04, -2.8339e-03, 2.1365e-03],
[-6.2077e-03, -2.6615e-03, 1.1215e-02]],
[[-8.1794e-03, -2.2417e-02, -3.4012e-02],
[-2.8553e-02, -2.9546e-02, -4.4372e-02],
[-5.0348e-02, -3.4973e-02, -5.2028e-02]],
[[ 1.4728e-02, 3.2834e-02, 2.6312e-02],
[ 1.3449e-02, 2.6407e-02, 2.6924e-02],
[ 2.5572e-02, 3.4316e-02, 2.6184e-02]]],
[[[ 7.2026e-03, -2.3931e-03, 2.2182e-03],
[ 4.2555e-03, -6.4084e-03, 7.8548e-03],
[ 2.0510e-02, 1.8644e-02, 2.3280e-02]],
[[-1.2471e-02, 1.3008e-02, 1.0010e-02],
[-1.7496e-03, 6.1331e-03, 4.3366e-03],
[ 5.2269e-03, 1.5111e-02, -8.1881e-03]],
[[-3.7337e-02, 1.9923e-02, -2.4149e-02],
[-4.9487e-02, -1.0510e-02, -4.2107e-02],
[-5.7684e-03, -4.8632e-03, -1.8332e-02]],
...,
[[ 7.2013e-03, -1.5208e-02, -1.6507e-02],
[-8.8276e-03, -1.8698e-02, -1.6637e-03],
[-1.2015e-02, 2.9667e-03, 6.2300e-03]],
[[-1.8341e-02, -9.0521e-03, 2.6030e-02],
[ 3.5930e-02, 5.3049e-02, 5.8487e-02],
[-1.3661e-02, -3.6888e-03, -7.1606e-03]],
[[-1.2594e-02, -4.0898e-02, 1.7162e-03],
[-1.7420e-02, -4.3435e-02, -1.3183e-02],
[-3.7506e-02, -5.5707e-02, -3.0051e-02]]]], device='cuda:0'),
Parameter containing:
tensor([0.1971, 0.1771, 0.1303, 0.1995, 0.1839, 0.0934, 0.2333, 0.2236, 0.1654,
0.1280, 0.0842, 0.1085, 0.3168, 0.2032, 0.3246, 0.2184, 0.3208, 0.2824,
0.3408, 0.3339, 0.3307, 0.5571, 0.2821, 0.3081, 0.2114, 0.2971, 0.2361,
0.5500, 0.1221, 0.3381, 0.1528, 0.1544, 0.1982, 0.0582, 0.1812, 0.2489,
0.1954, 0.0705, 0.0918, 0.1328, 0.2616, 0.2013, 0.0720, 0.1573, 0.1919,
0.0813, 0.1170, 0.2504, 0.2863, 0.3032, 0.1476, 0.3696, 0.1870, 0.2097,
0.1907, 0.2364, 0.1642, 0.1079, 0.2531, 0.1703, 0.1266, 0.0814, 0.2407,
0.2609, 0.2705, 0.2128, 0.5007, 0.2375, 0.0802, 0.2896, 0.1776, 0.0887,
0.1094, 0.1834, 0.2812, 0.1971, 0.2021, 0.3443, 0.1411, 0.1362, 0.2676,
0.1618, 0.2723, 0.2727, 0.2528, 0.0982, 0.4707, 0.2239, 0.3649, 0.1987,
0.0815, 0.2543, 0.3322, 0.1561, 0.2336, 0.1294, 0.2570, 0.1700, 0.1374,
0.2215, 0.5015, 0.3132, 0.1487, 0.1174, 0.0916, 0.2130, 0.1393, 0.3057,
0.5634, 0.1018, 0.0994, 0.0492, 0.4427, 0.3142, 0.4002, 0.1334, 0.2174,
0.5522, 0.2806, 0.2784, 0.4333, 0.2602, 0.3788, 0.1827, 0.2664, 0.1077,
0.3001, 0.2428, 0.5130, 0.0829, 0.1254, 0.1996, 0.1451, 0.2253, 0.1467,
0.3712, 0.0794, 0.5425, 0.2058, 0.2103, 0.1288, 0.4993, 0.1815, 0.1845,
0.4154, 0.3817, 0.2054, 0.2205, 0.1471, 0.4964, 0.4202, 0.0801, 0.0623,
0.3536, 0.2760, 0.3840, 0.1632, 0.1402, 0.2674, 0.0844, 0.2305, 0.2259,
0.2146, 0.4181, 0.2821, 0.2926, 0.3416, 0.4640, 0.3025, 0.3732, 0.5871,
0.0616, 0.2797, 0.3042, 0.2173, 0.3550, 0.2096, 0.2449, 0.3428, 0.2868,
0.3543, 0.4667, 0.3220, 0.3805, 0.2632, 0.2160, 0.1924, 0.4074, 0.4966,
0.3623, 0.1670, 0.1321, 0.2374, 0.2118, 0.1522, 0.1668, 0.3836, 0.0983,
0.3729, 0.3943, 0.4353, 0.2270, 0.1508, 0.3133, 0.3850, 0.5774, 0.1892,
0.2822, 0.0907, 0.2364, 0.0964, 0.2360, 0.0699, 0.2938, 0.5100, 0.3348,
0.2339, 0.1145, 0.2155, 0.2266, 0.2829, 0.2341, 0.1891, 0.2906, 0.2681,
0.3876, 0.3915, 0.1844, 0.1889, 0.4405, 0.1405, 0.3460, 0.2724, 0.2567,
0.2785, 0.1148, 0.1607, 0.1754, 0.0883, 0.1649, 0.1268, 0.2356, 0.2811,
0.0766, 0.1424, 0.1683, 0.3979, 0.2685, 0.6383, 0.1087, 0.3180, 0.1760,
0.3634, 0.2615, 0.1999, 0.2541], device='cuda:0'),
Parameter containing:
tensor([-0.0162, -0.2033, 0.0294, -0.1697, -0.1840, -0.0309, -0.2039, -0.1426,
-0.0443, -0.0886, -0.0647, -0.0968, -0.0380, -0.2073, -0.3061, 0.1443,
-0.3079, -0.1232, -0.1627, -0.0980, -0.2471, -0.2837, -0.1201, -0.2893,
-0.2303, -0.3562, -0.0825, -0.3483, 0.0707, -0.1321, -0.1074, -0.1451,
0.0235, 0.0225, -0.1885, -0.2507, -0.2461, 0.0631, -0.0023, -0.1209,
-0.2581, -0.1640, -0.0172, -0.1143, -0.2096, -0.0158, 0.0128, -0.1332,
-0.3139, -0.2294, -0.1527, -0.3503, 0.2086, 0.0785, -0.1597, -0.1990,
0.0346, 0.0388, -0.1269, 0.1019, 0.0981, -0.0390, -0.2537, -0.1356,
-0.1796, -0.2422, -0.4517, -0.3124, -0.0177, -0.2615, -0.1567, 0.0212,
-0.0753, -0.1426, -0.2788, 0.0062, -0.1895, -0.2327, -0.1298, -0.1200,
-0.1917, -0.0987, -0.1916, -0.1666, -0.2729, 0.1287, -0.4620, -0.2259,
-0.2270, 0.1939, 0.0230, -0.3303, -0.3202, -0.1292, -0.0716, 0.0048,
-0.2579, -0.0116, -0.0557, -0.1229, -0.4804, -0.2351, -0.1367, -0.0578,
-0.0537, -0.2743, -0.0827, -0.1922, -0.3481, -0.0358, -0.1094, 0.0138,
-0.1888, -0.2592, -0.3293, -0.0820, -0.1839, -0.1636, -0.3163, -0.0246,
-0.1667, -0.1653, -0.3076, -0.2229, -0.1834, -0.0536, -0.0621, -0.1752,
-0.5243, -0.1933, -0.1119, -0.2283, -0.0437, -0.1777, -0.1300, -0.2519,
-0.0456, -0.6305, -0.1364, -0.2138, 0.0406, -0.5287, -0.2014, -0.1442,
-0.1930, -0.3033, 0.1030, -0.1499, -0.2297, -0.5301, -0.2543, -0.0417,
0.0429, -0.3218, -0.1611, -0.2562, -0.1187, -0.1001, 0.0225, 0.0996,
-0.2138, -0.2019, 0.0808, -0.0121, -0.2364, -0.3247, -0.1482, -0.4846,
-0.3449, -0.1365, -0.6664, 0.0418, -0.2807, -0.0961, -0.2378, -0.1834,
-0.1890, -0.0377, -0.3056, -0.1843, -0.1357, -0.3038, -0.2680, -0.4143,
-0.2633, -0.1750, -0.1856, -0.2405, -0.1082, -0.2250, -0.1268, -0.1094,
0.0594, -0.1419, -0.1178, -0.1602, -0.0328, -0.0194, -0.1985, 0.0470,
-0.1887, -0.2776, -0.0930, -0.4092, -0.3378, -0.7252, 0.0260, -0.1829,
0.0561, -0.2227, -0.0026, -0.3218, -0.0093, -0.2843, -0.5121, -0.2337,
-0.0836, -0.0818, -0.1296, -0.2090, 0.0169, -0.1899, -0.1892, -0.3075,
-0.3108, -0.2986, -0.4712, -0.1823, -0.1893, -0.3131, -0.0876, -0.1166,
-0.2995, -0.0831, -0.3427, -0.0772, -0.1460, -0.1611, 0.0203, -0.0627,
-0.0610, -0.2574, -0.1383, 0.0470, -0.0302, -0.1638, -0.3323, -0.1741,
-0.6307, -0.0772, -0.2123, -0.1559, -0.0459, -0.2416, -0.0143, -0.2079],
device='cuda:0'),
Parameter containing:
tensor([[[[-1.1645e-02, -1.9010e-02, -2.1876e-02],
[ 2.0482e-02, 2.3962e-02, 2.9161e-02],
[ 4.3672e-02, 3.3278e-02, 4.9908e-02]],
[[-7.4040e-03, 2.8083e-03, -4.7339e-03],
[ 6.9030e-03, 1.4271e-02, -3.6954e-03],
[-3.1341e-03, 1.3736e-02, 1.6127e-03]],
[[ 1.8676e-02, -1.0553e-02, -1.4233e-02],
[ 8.9944e-03, -2.5068e-03, -1.2145e-02],
[-4.9455e-03, -2.9206e-02, -9.6385e-03]],
...,
[[-1.2655e-02, 1.7691e-02, 9.8264e-04],
[ 7.4271e-03, 7.6115e-03, 1.1135e-02],
[ 2.3242e-02, 1.1058e-02, 4.0498e-03]],
[[ 1.8557e-02, 1.2472e-02, 1.7220e-02],
[-4.8544e-03, 8.3627e-03, 2.2811e-02],
[-5.1675e-03, 2.3264e-02, 3.4068e-02]],
[[ 2.4934e-02, 2.2373e-02, 4.2614e-02],
[ 1.3486e-02, 1.6760e-03, 1.3019e-02],
[-6.2821e-03, -1.5112e-03, -8.9229e-03]]],
[[[-9.8089e-04, -6.3011e-03, 5.9932e-03],
[ 1.5936e-02, 1.3394e-02, 2.9934e-02],
[ 2.3149e-02, 2.0709e-02, 2.5485e-02]],
[[-2.0015e-02, -3.3349e-02, -8.0396e-03],
[-7.2800e-03, -1.2187e-02, -2.0389e-04],
[-1.3138e-02, -2.0427e-02, -1.6286e-02]],
[[-6.7681e-03, 5.0045e-03, -2.6683e-03],
[-2.1073e-02, 2.8275e-04, -1.8205e-02],
[-1.7382e-02, -5.0244e-03, -3.0386e-03]],
...,
[[-1.1035e-02, -2.2964e-02, -1.1028e-02],
[-6.3256e-03, -4.1667e-03, -1.7323e-02],
[-1.3611e-02, -2.3468e-02, -1.6436e-02]],
[[ 7.3663e-03, 6.6219e-03, 5.2776e-03],
[-3.5464e-03, 3.2750e-03, -9.1126e-03],
[ 3.5593e-04, -1.0151e-02, -1.9123e-02]],
[[ 1.8193e-03, 8.8087e-03, 5.1361e-03],
[ 3.1915e-03, 2.5287e-02, 2.4939e-02],
[ 1.3968e-02, 1.9613e-02, 2.2382e-02]]],
[[[-4.1548e-03, 8.8964e-03, 2.0143e-03],
[ 1.1327e-02, 1.3251e-02, 1.4014e-02],
[ 7.2196e-03, 1.3045e-02, 2.4827e-02]],
[[-1.5025e-02, 5.0530e-03, 7.4766e-03],
[-2.4685e-02, -1.6732e-02, -1.0888e-02],
[-2.8064e-02, -1.1875e-02, -3.4120e-03]],
[[ 2.8449e-02, 1.4594e-02, 6.9441e-03],
[ 2.4799e-02, 1.9453e-02, 1.1294e-02],
[-1.0787e-02, -2.1006e-02, -1.0372e-02]],
...,
[[ 1.4967e-02, 8.2449e-03, 2.0244e-03],
[ 1.4287e-02, -6.3867e-03, -8.0757e-03],
[ 2.7547e-02, 1.0791e-02, 1.6567e-02]],
[[ 3.6191e-02, 3.8918e-02, 3.9028e-02],
[-8.3489e-04, 1.3273e-02, 2.0172e-02],
[-2.0652e-02, -5.4010e-03, 1.7147e-03]],
[[ 2.0373e-04, 3.5919e-03, 8.5592e-03],
[ 6.2363e-03, -9.3086e-05, 1.2940e-02],
[ 1.3152e-02, 1.0732e-02, 1.9896e-02]]],
...,
[[[-1.7400e-02, -6.7019e-03, -9.1787e-03],
[-9.9672e-03, 2.6298e-04, 3.3439e-03],
[ 1.5721e-02, 1.4216e-02, 2.0509e-02]],
[[ 2.1410e-02, 3.6914e-02, 2.8239e-02],
[ 3.8158e-02, 4.8944e-02, 3.4652e-02],
[ 3.1723e-02, 4.4208e-02, 4.0035e-02]],
[[-3.3437e-03, -1.0482e-02, -5.3990e-03],
[-5.3186e-03, 1.1394e-02, 1.7593e-03],
[-5.6652e-03, -6.6373e-03, -1.3492e-02]],
...,
[[-1.7099e-02, -1.8145e-03, -1.3040e-02],
[-2.2750e-02, -3.6062e-03, -8.0294e-03],
[-1.6087e-02, -1.0175e-02, -1.3529e-02]],
[[ 4.1701e-04, -5.1785e-03, -2.1884e-02],
[ 2.6919e-03, 8.9139e-03, -1.4217e-04],
[-7.3746e-03, -6.6853e-03, -2.3725e-02]],
[[ 1.9425e-02, 1.3175e-02, 1.7511e-02],
[ 1.8235e-02, 4.4286e-02, 2.3767e-02],
[ 2.6504e-02, 3.3104e-02, 1.9696e-02]]],
[[[-1.0177e-02, -1.0701e-02, -2.0428e-02],
[-1.7986e-02, 5.9928e-03, -1.0584e-03],
[-1.8794e-02, -1.8773e-03, -6.9449e-03]],
[[-2.8498e-03, 1.6427e-03, 1.4575e-04],
[-5.4403e-03, 8.3667e-03, -9.4164e-03],
[-4.4999e-03, 5.4902e-03, 2.4863e-03]],
[[-1.3356e-02, -2.1525e-02, 5.3421e-04],
[-1.9160e-02, -2.4645e-02, -1.3791e-02],
[-6.1991e-03, -1.3174e-02, -3.6783e-03]],
...,
[[-3.3993e-03, -2.7823e-03, 7.6715e-03],
[-2.0649e-02, -1.2731e-02, -9.4138e-03],
[-1.3678e-03, -3.4410e-02, -2.6984e-02]],
[[-3.5651e-04, 2.0102e-03, 1.4130e-02],
[-1.3073e-02, -1.6616e-02, -1.2690e-02],
[-3.5934e-02, -4.1700e-02, -3.3968e-02]],
[[ 2.0470e-02, 8.0159e-04, -1.1607e-03],
[ 9.5101e-03, 3.0336e-02, 2.7362e-02],
[ 1.5588e-02, 3.2851e-02, 1.3015e-02]]],
[[[-1.5574e-02, -3.2971e-02, -3.1939e-02],
[-2.2502e-02, -5.7187e-03, -5.6729e-03],
[-2.7309e-02, -1.6981e-02, 1.2832e-04]],
[[-1.1925e-02, -2.9479e-02, -2.0437e-02],
[-2.4408e-02, -2.2069e-02, -1.9965e-03],
[-2.3279e-02, -5.5140e-03, 2.5630e-02]],
[[-1.6100e-02, -8.2417e-03, 1.5266e-04],
[-2.6195e-03, -8.2754e-03, -2.9435e-02],
[-2.7493e-03, -2.4889e-02, -2.3583e-02]],
...,
[[ 1.7985e-02, 1.8594e-02, 8.9198e-04],
[-1.7319e-02, 7.8735e-03, -2.8659e-03],
[ 3.8596e-03, 2.9061e-02, 2.4188e-02]],
[[-2.6735e-02, -1.4391e-02, -4.0148e-02],
[-2.6728e-02, -2.4455e-02, -6.9176e-03],
[-5.7244e-02, -2.1995e-04, 5.5438e-02]],
[[ 2.3487e-02, 2.7157e-03, -8.4719e-04],
[ 1.7886e-02, 5.4860e-03, 2.8059e-02],
[ 4.6468e-03, 1.8598e-02, 1.3761e-03]]]], device='cuda:0'),
Parameter containing:
tensor([0.2427, 0.2232, 0.2511, 0.2288, 0.2074, 0.2905, 0.2482, 0.3102, 0.2749,
0.2892, 0.2448, 0.1759, 0.2426, 0.2780, 0.2315, 0.2631, 0.3383, 0.2785,
0.2536, 0.2989, 0.2335, 0.2812, 0.3486, 0.2778, 0.2280, 0.2547, 0.3032,
0.2468, 0.2512, 0.2973, 0.2577, 0.3200, 0.2385, 0.2714, 0.2532, 0.2625,
0.3344, 0.2626, 0.1838, 0.2839, 0.2187, 0.2666, 0.2858, 0.2471, 0.2915,
0.2332, 0.2637, 0.2691, 0.2432, 0.2384, 0.2356, 0.2525, 0.2564, 0.2451,
0.2529, 0.2522, 0.2800, 0.3165, 0.2340, 0.2634, 0.2569, 0.1942, 0.2621,
0.2205, 0.2301, 0.2323, 0.2811, 0.1897, 0.2280, 0.3472, 0.2717, 0.3191,
0.2440, 0.2719, 0.2781, 0.2262, 0.3444, 0.2648, 0.2725, 0.2851, 0.2039,
0.2935, 0.2742, 0.2774, 0.2654, 0.2430, 0.2721, 0.2708, 0.3085, 0.2895,
0.2596, 0.2147, 0.3119, 0.3449, 0.2262, 0.2814, 0.2326, 0.2712, 0.2637,
0.2323, 0.3333, 0.2714, 0.2991, 0.2747, 0.2515, 0.2394, 0.2709, 0.2836,
0.2866, 0.2408, 0.2560, 0.2048, 0.2394, 0.2813, 0.3267, 0.2761, 0.2123,
0.2715, 0.2540, 0.2771, 0.3209, 0.1905, 0.3989, 0.2676, 0.2357, 0.2169,
0.3216, 0.3596, 0.2838, 0.2648, 0.2702, 0.2469, 0.2442, 0.2553, 0.2599,
0.2693, 0.2399, 0.2700, 0.2063, 0.2711, 0.2834, 0.2781, 0.2529, 0.2013,
0.2343, 0.2082, 0.3063, 0.1635, 0.2673, 0.2197, 0.2787, 0.2724, 0.2744,
0.2287, 0.2969, 0.2662, 0.2982, 0.2396, 0.3039, 0.2319, 0.2773, 0.2661,
0.2898, 0.2489, 0.3060, 0.2612, 0.2937, 0.3045, 0.2999, 0.2580, 0.2093,
0.2714, 0.2993, 0.2679, 0.2963, 0.2754, 0.2580, 0.2566, 0.2634, 0.2325,
0.2442, 0.2934, 0.2398, 0.2631, 0.2851, 0.2870, 0.2239, 0.2410, 0.2676,
0.2681, 0.2638, 0.2732, 0.2812, 0.2203, 0.2670, 0.2764, 0.2550, 0.3160,
0.2888, 0.2615, 0.2178, 0.2485, 0.2414, 0.2798, 0.2872, 0.2767, 0.2551,
0.2429, 0.2459, 0.3288, 0.3024, 0.2912, 0.2625, 0.3019, 0.2643, 0.2721,
0.2108, 0.2368, 0.2269, 0.1988, 0.2830, 0.2569, 0.2349, 0.2755, 0.2442,
0.2717, 0.2747, 0.2785, 0.2516, 0.2227, 0.2783, 0.2465, 0.2652, 0.2641,
0.2960, 0.2671, 0.2679, 0.2537, 0.2847, 0.2507, 0.2525, 0.2024, 0.2311,
0.2618, 0.2764, 0.3031, 0.2452, 0.2716, 0.2273, 0.2295, 0.2611, 0.2329,
0.2690, 0.2753, 0.2737, 0.2590, 0.2421, 0.2685, 0.3392, 0.3073, 0.1371,
0.3650, 0.2980, 0.2460, 0.2487, 0.2912, 0.2704, 0.2560, 0.2213, 0.2569,
0.2661, 0.2367, 0.2742, 0.2847, 0.3055, 0.2671, 0.2819, 0.2791, 0.2401,
0.2549, 0.2210, 0.3507, 0.2852, 0.2162, 0.2821, 0.2369, 0.2905, 0.2826,
0.2300, 0.2745, 0.2437, 0.2522, 0.2489, 0.2395, 0.2851, 0.2887, 0.2621,
0.2500, 0.2689, 0.2427, 0.3010, 0.3067, 0.2861, 0.2387, 0.2462, 0.2859,
0.2550, 0.2630, 0.2442, 0.2145, 0.2898, 0.2282, 0.2327, 0.2242, 0.2738,
0.2485, 0.2379, 0.3058, 0.2798, 0.2761, 0.2252, 0.2866, 0.2660, 0.3250,
0.2612, 0.2767, 0.3205, 0.2932, 0.3183, 0.2939, 0.3103, 0.2553, 0.2981,
0.3667, 0.3086, 0.2254, 0.2352, 0.2348, 0.2555, 0.2597, 0.2369, 0.3017,
0.2776, 0.2728, 0.3174, 0.2785, 0.2721, 0.2637, 0.2702, 0.3633, 0.2869,
0.2675, 0.3405, 0.2587, 0.2732, 0.2747, 0.2821, 0.2750, 0.2630, 0.2018,
0.2358, 0.3034, 0.3155, 0.3013, 0.2775, 0.2511, 0.2945, 0.1605, 0.2825,
0.2964, 0.2194, 0.2061, 0.2332, 0.2348, 0.2663, 0.2543, 0.2927, 0.2215,
0.2521, 0.2827, 0.1993, 0.2453, 0.2597, 0.2654, 0.2757, 0.2650, 0.2444,
0.2949, 0.2308, 0.3071, 0.1904, 0.3024, 0.2786, 0.3659, 0.2966, 0.2746,
0.2449, 0.2201, 0.2564, 0.2853, 0.2392, 0.2457, 0.2467, 0.2374, 0.2664,
0.2460, 0.3182, 0.1793, 0.2379, 0.2596, 0.2847, 0.2452, 0.1974, 0.2388,
0.2949, 0.2879, 0.2786, 0.2765, 0.3296, 0.2530, 0.2690, 0.2547, 0.2333,
0.2348, 0.2690, 0.2718, 0.2679, 0.2516, 0.2710, 0.2366, 0.2601, 0.2764,
0.2880, 0.2008, 0.2637, 0.2263, 0.2511, 0.2604, 0.2805, 0.2989, 0.2965,
0.2597, 0.2767, 0.2553, 0.2959, 0.2512, 0.2925, 0.3008, 0.2423, 0.2394,
0.2708, 0.3704, 0.2879, 0.2532, 0.2248, 0.2023, 0.2279, 0.2366, 0.3082,
0.2980, 0.2909, 0.2777, 0.4293, 0.2658, 0.2940, 0.2418, 0.2816, 0.3247,
0.2647, 0.2216, 0.2758, 0.2421, 0.2078, 0.2332, 0.2271, 0.2611, 0.3650,
0.2017, 0.2598, 0.2160, 0.2641, 0.1408, 0.2664, 0.2502, 0.2553, 0.2227,
0.2417, 0.2696, 0.2388, 0.2833, 0.2333, 0.2667, 0.2224, 0.2691, 0.2710,
0.2459, 0.2674, 0.2430, 0.2593, 0.1851, 0.2950, 0.3664, 0.2212, 0.3026,
0.1840, 0.3443, 0.2140, 0.3717, 0.2360, 0.3081, 0.2638, 0.2233],
device='cuda:0'),
Parameter containing:
tensor([-0.1986, -0.1593, -0.2054, -0.1598, -0.1268, -0.3226, -0.1597, -0.3477,
-0.2497, -0.2730, -0.2319, -0.0286, -0.1899, -0.2813, -0.1733, -0.2412,
-0.3712, -0.2747, -0.2053, -0.2585, -0.1535, -0.2748, -0.3241, -0.2525,
-0.1906, -0.2252, -0.3436, -0.2202, -0.1664, -0.2716, -0.1920, -0.3399,
-0.2026, -0.2972, -0.2616, -0.2238, -0.2486, -0.2606, -0.0893, -0.3572,
-0.1283, -0.2583, -0.2450, -0.1523, -0.3165, -0.1445, -0.2522, -0.1963,
-0.1794, -0.1071, -0.1662, -0.2053, -0.2530, -0.1447, -0.2517, -0.2062,
-0.2817, -0.3376, -0.1382, -0.2389, -0.2557, -0.0156, -0.2169, -0.1763,
-0.1486, -0.2122, -0.2002, -0.0716, -0.2089, -0.3580, -0.2588, -0.3599,
-0.1528, -0.2107, -0.2925, -0.1855, -0.3970, -0.1257, -0.2574, -0.2412,
-0.0863, -0.3065, -0.2701, -0.3380, -0.2485, -0.1935, -0.2987, -0.2279,
-0.3600, -0.2764, -0.2480, -0.1208, -0.3378, -0.2661, -0.1677, -0.2470,
-0.2152, -0.2591, -0.1936, -0.1543, -0.4117, -0.1570, -0.2372, -0.2997,
-0.2124, -0.2034, -0.1848, -0.3070, -0.3438, -0.1839, -0.1937, -0.0916,
-0.2338, -0.3558, -0.1967, -0.3303, -0.1398, -0.2177, -0.1665, -0.1857,
-0.3115, -0.1049, -0.4229, -0.2408, -0.1320, -0.1631, -0.3378, -0.3300,
-0.3183, -0.2268, -0.2787, -0.1950, -0.1950, -0.1463, -0.2437, -0.2297,
-0.1282, -0.2164, -0.1179, -0.2437, -0.2611, -0.2656, -0.1948, -0.1208,
-0.1668, -0.1351, -0.2713, -0.0560, -0.2243, -0.1318, -0.2356, -0.2720,
-0.2051, -0.1736, -0.2891, -0.2627, -0.3358, -0.1779, -0.2309, -0.1477,
-0.2685, -0.1882, -0.2629, -0.1983, -0.3522, -0.1905, -0.2778, -0.3395,
-0.2895, -0.2240, -0.1150, -0.2462, -0.2426, -0.2581, -0.3133, -0.2315,
-0.2271, -0.2077, -0.2109, -0.1371, -0.1323, -0.2529, -0.1716, -0.2532,
-0.2277, -0.2084, -0.1803, -0.1868, -0.2404, -0.2166, -0.2197, -0.2870,
-0.3062, -0.1507, -0.1054, -0.2199, -0.2415, -0.3310, -0.2700, -0.1568,
-0.1449, -0.2610, -0.1828, -0.2648, -0.3134, -0.2937, -0.2687, -0.2115,
-0.2164, -0.4522, -0.2999, -0.3032, -0.2292, -0.3099, -0.2642, -0.2695,
-0.1441, -0.1671, -0.1570, -0.1415, -0.2222, -0.1736, -0.1481, -0.2573,
-0.2060, -0.1703, -0.2360, -0.1770, -0.2132, -0.2016, -0.3001, -0.1518,
-0.2086, -0.2805, -0.2698, -0.2292, -0.1293, -0.2514, -0.2600, -0.2454,
-0.1744, -0.1029, -0.1679, -0.2353, -0.2007, -0.3363, -0.1640, -0.2430,
-0.1699, -0.1697, -0.1837, -0.1625, -0.2415, -0.2687, -0.2305, -0.2029,
-0.2209, -0.2240, -0.2675, -0.3233, 0.1462, -0.4777, -0.2376, -0.1489,
-0.1462, -0.3055, -0.2234, -0.1697, -0.1952, -0.2131, -0.2340, -0.2039,
-0.3054, -0.2596, -0.3470, -0.2176, -0.2706, -0.2897, -0.1729, -0.2300,
-0.1066, -0.3556, -0.2912, -0.1777, -0.2007, -0.1699, -0.3009, -0.3046,
-0.1693, -0.2602, -0.2053, -0.1810, -0.1808, -0.1730, -0.3757, -0.1808,
-0.1805, -0.1895, -0.2643, -0.2075, -0.2365, -0.1975, -0.3064, -0.1984,
-0.1811, -0.3676, -0.1198, -0.1485, -0.1770, -0.0781, -0.2052, -0.1360,
-0.1417, -0.1691, -0.2395, -0.1785, -0.1747, -0.2484, -0.2717, -0.3096,
-0.1465, -0.2239, -0.2584, -0.3572, -0.2311, -0.2878, -0.3841, -0.3475,
-0.3896, -0.1891, -0.2861, -0.2431, -0.2837, -0.4365, -0.3353, -0.1802,
-0.1976, -0.1529, -0.1978, -0.2535, -0.1954, -0.2667, -0.2813, -0.2487,
-0.3070, -0.2339, -0.2212, -0.1925, -0.2224, -0.4178, -0.3151, -0.2663,
-0.3581, -0.1935, -0.2385, -0.2424, -0.1850, -0.2265, -0.1803, -0.0777,
-0.1492, -0.3361, -0.4133, -0.3123, -0.2745, -0.1247, -0.3102, 0.0041,
-0.1981, -0.3301, -0.2047, -0.1053, -0.1653, -0.1634, -0.1116, -0.2314,
-0.3191, -0.1818, -0.2657, -0.2220, -0.1029, -0.1999, -0.2702, -0.2139,
-0.2256, -0.2653, -0.1630, -0.3322, -0.1617, -0.3446, 0.0288, -0.2456,
-0.3171, -0.3580, -0.2857, -0.2520, -0.2031, -0.1522, -0.2203, -0.3490,
-0.1685, -0.1424, -0.1602, -0.1553, -0.3057, -0.2420, -0.3536, -0.0551,
-0.0987, -0.2272, -0.2619, -0.2035, -0.0906, -0.1976, -0.3040, -0.2732,
-0.3161, -0.2102, -0.3384, -0.1740, -0.1475, -0.1842, -0.1823, -0.1151,
-0.2183, -0.2010, -0.2659, -0.2205, -0.2567, -0.1633, -0.2213, -0.2658,
-0.2938, -0.1069, -0.2522, -0.1103, -0.2216, -0.2244, -0.2908, -0.2176,
-0.3605, -0.2374, -0.2391, -0.2251, -0.2256, -0.1339, -0.1970, -0.2970,
-0.2206, -0.2051, -0.2229, -0.3602, -0.2923, -0.2498, -0.1466, -0.0979,
-0.1686, -0.2158, -0.2881, -0.3002, -0.2760, -0.2496, -0.3536, -0.2868,
-0.3251, -0.1847, -0.3062, -0.3861, -0.2650, -0.1339, -0.1846, -0.1630,
-0.0630, -0.1717, -0.1415, -0.1906, -0.4611, -0.1391, -0.1920, -0.1369,
-0.1647, -0.0055, -0.2598, -0.2653, -0.2319, -0.1780, -0.1913, -0.2055,
-0.1891, -0.2625, -0.1633, -0.2497, -0.1696, -0.1907, -0.2431, -0.1825,
-0.2607, -0.1943, -0.2361, -0.0581, -0.2758, -0.2593, -0.1466, -0.3589,
-0.0439, -0.3440, -0.1089, -0.4219, -0.1503, -0.2792, -0.3035, -0.1156],
device='cuda:0'),
Parameter containing:
tensor([[[[ 1.6218e-04, -1.4720e-02, -1.7000e-02],
[-1.2850e-02, -3.3085e-02, -3.6656e-02],
[ 2.7812e-02, 1.7691e-02, -1.8369e-02]],
[[ 1.0528e-02, 3.1379e-02, 2.4801e-02],
[-1.2698e-02, -2.9453e-02, -1.1834e-02],
[-9.4094e-03, -8.9462e-03, -3.1349e-02]],
[[-7.8447e-03, -2.9256e-02, 5.3590e-03],
[-1.3791e-02, -1.1116e-02, 5.0388e-03],
[-2.4919e-03, 7.3514e-03, 5.4013e-03]],
...,
[[-1.0276e-03, -1.0275e-02, -2.9986e-02],
[-3.8465e-03, 1.9549e-03, -1.6291e-02],
[-1.8100e-03, 8.3778e-03, -8.5481e-03]],
[[-1.8196e-02, -1.3533e-02, -1.7457e-02],
[ 2.2457e-02, 5.7402e-02, 1.9325e-02],
[-2.4977e-02, -3.2113e-02, -8.1780e-03]],
[[ 3.6550e-03, 4.9358e-03, -5.7597e-03],
[-1.6875e-02, 1.3999e-04, 3.7629e-04],
[-2.6272e-03, 1.0947e-03, 1.1145e-03]]],
[[[ 1.4018e-02, 3.9198e-03, -1.7189e-03],
[-1.3175e-03, 4.3503e-04, -1.1798e-02],
[-9.8003e-03, -1.7693e-02, -1.9910e-02]],
[[-1.4957e-02, -1.9796e-02, -2.8724e-02],
[ 5.8908e-03, -1.5228e-02, -5.6715e-03],
[ 2.9284e-03, -1.8028e-02, -7.1433e-03]],
[[-1.1625e-02, -3.3804e-02, -1.0025e-02],
[-1.6606e-02, -5.5716e-02, -2.3204e-02],
[-2.5758e-02, -4.3135e-02, -2.5901e-02]],
...,
[[-1.5007e-02, -1.4333e-02, -2.5937e-03],
[-2.3078e-02, -1.5820e-02, -2.2818e-03],
[-4.1318e-03, -8.0353e-03, -2.3236e-03]],
[[-1.8531e-02, -1.8004e-02, -2.8084e-02],
[-3.6680e-02, -6.8641e-02, -5.2469e-02],
[-1.1712e-02, -2.4334e-02, -1.6733e-02]],
[[-2.2078e-02, -2.9163e-02, -3.8717e-03],
[-7.0301e-03, 1.6718e-02, 5.4339e-03],
[-1.3131e-02, 1.1999e-02, -1.7480e-02]]],
[[[-5.2378e-03, -3.4890e-03, -2.0851e-03],
[ 1.5306e-02, -2.1752e-02, -8.7682e-03],
[ 2.2460e-02, 9.9175e-03, -3.3635e-03]],
[[ 7.4677e-03, -9.1762e-03, -9.2569e-05],
[ 1.9441e-04, 1.2344e-03, -8.9978e-03],
[-5.1243e-04, 2.1850e-04, -4.8828e-03]],
[[ 1.7078e-02, 3.3955e-03, 9.3503e-03],
[ 2.0334e-02, -1.0621e-04, -8.2017e-05],
[ 1.0706e-02, -1.8414e-03, 1.0828e-02]],
...,
[[ 3.2008e-02, 2.3494e-02, 2.5386e-02],
[ 1.9307e-02, 2.3924e-02, 2.8972e-02],
[ 9.9003e-03, 2.0158e-02, 2.2655e-02]],
[[-9.8395e-03, -1.1114e-02, -3.7696e-03],
[-2.9508e-02, -3.6956e-02, -1.8228e-02],
[-1.3663e-03, -2.5845e-03, 1.0352e-02]],
[[-7.3867e-03, -2.5413e-02, -2.1942e-02],
[-1.6699e-02, -1.5133e-02, -1.3030e-02],
[-2.0090e-02, 3.7970e-03, -1.0341e-02]]],
...,
[[[-1.6157e-02, -1.6883e-02, -2.8328e-04],
[-7.7759e-03, -2.4465e-03, -1.4641e-02],
[ 2.4639e-02, 3.9862e-02, 2.1048e-02]],
[[ 2.4491e-03, -9.3885e-03, -1.1786e-02],
[ 2.5301e-02, 2.5625e-04, 7.1335e-03],
[ 2.2342e-02, 1.9042e-02, 7.2526e-03]],
[[-1.4652e-02, -2.7802e-02, -4.3564e-03],
[-1.7961e-02, -4.3846e-02, 2.7409e-03],
[-4.7968e-03, -8.4231e-03, 1.2070e-02]],
...,
[[-2.0171e-02, -3.3546e-02, -1.6728e-02],
[-1.7847e-02, -5.1713e-02, -2.6780e-02],
[-1.3145e-03, -4.3181e-03, -9.6373e-03]],
[[-5.3917e-03, -2.0410e-04, 2.7798e-03],
[-9.6882e-04, -2.5141e-02, 1.4804e-02],
[ 2.8748e-02, 9.0832e-03, 4.2548e-02]],
[[-1.5698e-02, -1.9303e-02, -9.1469e-03],
[-2.0025e-02, -1.1131e-02, -3.3902e-02],
[-5.7436e-03, -7.3640e-03, -1.0044e-02]]],
[[[-8.8612e-03, -4.5370e-03, -1.2354e-02],
[-5.9245e-03, -1.7058e-02, -2.8041e-02],
[-1.0435e-02, 7.6695e-04, -1.0578e-02]],
[[ 9.5200e-03, -5.1975e-03, 1.2947e-02],
[ 4.4305e-03, -2.3992e-02, -8.4569e-04],
[ 4.6608e-03, 9.6787e-03, 8.2174e-03]],
[[ 5.1559e-03, 4.4635e-04, -7.9934e-03],
[ 3.3069e-03, 1.4450e-02, 8.9234e-03],
[ 6.3402e-03, 1.9043e-02, 1.9021e-02]],
...,
[[ 7.6964e-03, -1.3777e-02, 6.0539e-03],
[-1.5745e-03, -2.3391e-02, -1.0052e-02],
[ 9.5183e-03, -1.2251e-02, 2.2436e-03]],
[[ 1.0375e-02, 3.5875e-03, -5.7940e-04],
[ 7.0412e-03, -1.0673e-02, -4.9120e-03],
[-2.6034e-03, 1.1306e-02, 7.0696e-03]],
[[-1.7509e-02, -2.3182e-02, -1.7897e-02],
[-1.7769e-03, 1.9672e-03, -7.3220e-03],
[-6.6833e-03, 9.8286e-03, 2.0653e-03]]],
[[[ 2.8375e-02, -8.1936e-03, 1.8009e-02],
[ 1.5829e-02, -1.3571e-02, -1.9335e-02],
[ 4.0766e-03, -1.5722e-02, -5.0620e-02]],
[[-5.5310e-03, -1.8996e-02, -7.9436e-03],
[ 1.3825e-03, -4.9608e-02, 1.7256e-03],
[ 7.6629e-03, -7.6101e-03, 1.2541e-02]],
[[ 1.8052e-02, 3.1718e-02, 4.2556e-03],
[-3.6760e-03, 3.0490e-03, -1.2264e-02],
[-8.9404e-03, -1.6604e-02, 1.6348e-03]],
...,
[[ 5.3192e-03, 1.8204e-02, 1.8114e-02],
[-6.1202e-03, 1.5905e-03, 2.0264e-02],
[-1.1471e-02, -1.5697e-02, 9.0871e-03]],
[[ 3.7707e-03, 8.0599e-03, 1.8290e-02],
[ 1.7257e-02, 6.9638e-03, 1.8746e-02],
[ 1.0751e-02, 1.3663e-02, -1.0081e-03]],
[[ 1.9711e-02, -1.4569e-02, -2.4663e-02],
[ 2.5966e-03, -2.4807e-02, 9.3861e-03],
[-1.2876e-03, 1.3974e-03, 1.3434e-02]]]], device='cuda:0'),
Parameter containing:
tensor([0.4474, 0.5138, 0.4335, 0.3421, 0.3855, 0.3495, 0.3741, 0.5836, 0.4327,
0.5043, 0.4618, 0.3866, 0.3498, 0.4798, 0.3310, 0.3913, 0.3880, 0.5225,
0.3975, 0.3292, 0.4151, 0.4458, 0.3970, 0.3614, 0.3914, 0.4633, 0.3463,
0.3644, 0.3272, 0.4584, 0.4280, 0.4538, 0.4030, 0.4673, 0.4209, 0.3987,
0.4233, 0.3876, 0.4212, 0.3460, 0.3522, 0.3744, 0.4550, 0.2888, 0.4590,
0.4817, 0.4450, 0.5110, 0.4052, 0.4247, 0.3558, 0.3075, 0.4462, 0.4724,
0.4253, 0.3884, 0.4492, 0.3727, 0.4630, 0.3985, 0.3512, 0.3665, 0.3860,
0.5082, 0.4022, 0.3458, 0.4805, 0.5390, 0.4223, 0.4275, 0.4590, 0.4736,
0.3673, 0.5405, 0.3243, 0.5178, 0.4743, 0.3506, 0.3759, 0.4328, 0.3867,
0.4591, 0.3843, 0.4982, 0.5288, 0.3946, 0.4589, 0.3197, 0.4676, 0.4806,
0.4308, 0.4235, 0.3284, 0.3877, 0.4140, 0.4469, 0.4041, 0.4407, 0.4356,
0.5120, 0.5059, 0.4628, 0.4585, 0.3311, 0.3424, 0.4150, 0.5170, 0.4593,
0.5228, 0.4252, 0.4214, 0.4995, 0.4098, 0.5380, 0.4874, 0.3719, 0.4649,
0.4320, 0.3277, 0.3743, 0.4360, 0.4838, 0.4399, 0.3763, 0.4150, 0.5147,
0.5012, 0.4382, 0.3655, 0.4037, 0.4498, 0.4720, 0.3914, 0.3237, 0.3208,
0.3224, 0.4291, 0.4009, 0.3947, 0.3779, 0.4349, 0.4120, 0.3274, 0.4334,
0.3740, 0.4189, 0.4288, 0.3071, 0.4260, 0.3410, 0.4375, 0.4407, 0.3750,
0.5853, 0.4518, 0.5045, 0.3005, 0.4968, 0.4155, 0.3755, 0.5514, 0.4146,
0.4677, 0.1404, 0.5001, 0.4193, 0.4246, 0.4452, 0.5109, 0.4488, 0.4574,
0.3896, 0.4145, 0.4497, 0.4245, 0.3971, 0.3957, 0.4072, 0.5305, 0.4986,
0.3733, 0.4280, 0.3469, 0.4178, 0.3766, 0.4029, 0.3814, 0.4493, 0.5132,
0.4080, 0.4155, 0.3635, 0.4391, 0.3489, 0.4228, 0.4833, 0.3494, 0.4406,
0.3795, 0.4298, 0.4910, 0.3878, 0.6299, 0.4322, 0.5436, 0.4140, 0.4312,
0.3161, 0.3612, 0.3597, 0.4281, 0.4506, 0.4294, 0.3646, 0.4110, 0.4038,
0.4098, 0.3901, 0.3928, 0.5421, 0.3629, 0.4078, 0.4586, 0.4217, 0.3953,
0.3997, 0.3838, 0.4374, 0.3576, 0.4217, 0.4128, 0.3904, 0.4137, 0.5145,
0.4039, 0.3577, 0.4429, 0.5639, 0.3848, 0.6104, 0.4482, 0.6203, 0.5336,
0.3480, 0.5401, 0.6044, 0.4077, 0.3469, 0.4281, 0.4631, 0.5948, 0.3479,
0.3689, 0.3658, 0.3191, 0.5492, 0.3410, 0.5386, 0.4041, 0.3373, 0.4186,
0.5187, 0.3933, 0.3188, 0.3502, 0.3736, 0.4238, 0.4752, 0.3322, 0.5078,
0.4317, 0.5318, 0.4413, 0.5510, 0.5648, 0.4130, 0.4017, 0.4304, 0.4077,
0.4285, 0.4360, 0.3749, 0.4261, 0.3905, 0.3030, 0.3412, 0.3768, 0.4507,
0.3127, 0.4592, 0.4298, 0.3936, 0.3106, 0.3869, 0.3594, 0.4046, 0.4722,
0.4373, 0.3902, 0.3515, 0.4448, 0.4299, 0.4347, 0.4693, 0.4807, 0.2549,
0.4171, 0.4387, 0.4156, 0.3976, 0.4092, 0.4953, 0.4824, 0.3468, 0.4382,
0.4179, 0.4668, 0.3299, 0.5986, 0.4949, 0.4167, 0.4996, 0.4528, 0.4550,
0.4945, 0.3415, 0.4658, 0.4356, 0.3976, 0.5439, 0.4643, 0.5122, 0.4669,
0.4463, 0.4810, 0.3492, 0.3961, 0.3593, 0.4053, 0.3878, 0.3959, 0.5001,
0.2808, 0.5470, 0.4448, 0.4894, 0.4621, 0.3417, 0.3485, 0.5060, 0.3637,
0.3774, 0.3248, 0.4520, 0.3936, 0.3403, 0.4660, 0.4114, 0.3643, 0.4196,
0.3903, 0.5128, 0.4221, 0.4115, 0.4240, 0.3610, 0.4999, 0.3672, 0.4721,
0.4252, 0.5590, 0.4694, 0.7322, 0.5849, 0.4749, 0.4426, 0.3934, 0.3909,
0.4576, 0.3636, 0.4146, 0.4129, 0.5081, 0.3681, 0.3652, 0.4254, 0.2945,
0.4142, 0.3145, 0.4304, 0.4252, 0.3493, 0.4257, 0.5133, 0.3261, 0.4367,
0.3637, 0.3712, 0.4183, 0.3772, 0.4418, 0.4231, 0.4133, 0.4731, 0.4955,
0.4046, 0.4079, 0.4719, 0.3875, 0.4673, 0.4129, 0.4569, 0.3530, 0.4793,
0.3844, 0.3785, 0.3343, 0.4351, 0.6512, 0.4295, 0.4122, 0.3788, 0.3692,
0.4343, 0.4214, 0.3873, 0.4566, 0.4456, 0.4107, 0.4596, 0.7082, 0.4452,
0.3515, 0.4785, 0.4217, 0.5756, 0.4312, 0.4047, 0.4043, 0.4764, 0.5489,
0.4430, 0.5559, 0.3744, 0.3951, 0.4376, 0.4752, 0.4340, 0.4399, 0.3586,
0.4161, 0.3930, 0.4599, 0.4354, 0.3448, 0.4649, 0.4442, 0.4275, 0.3881,
0.3247, 0.4909, 0.3426, 0.3989, 0.4320, 0.3363, 0.3991, 0.4732, 0.3514,
0.4736, 0.4244, 0.4603, 0.3298, 0.4357, 0.4353, 0.3742, 0.4191, 0.3880,
0.4212, 0.4527, 0.7213, 0.3969, 0.5217, 0.3786, 0.3512, 0.5318, 0.4138,
0.3243, 0.3244, 0.3652, 0.4774, 0.3997, 0.2800, 0.4562, 0.4463, 0.4816,
0.4290, 0.4399, 0.4633, 0.3575, 0.4774, 0.3105, 0.4356, 0.3797, 0.4304,
0.4261, 0.3740, 0.3370, 0.3917, 0.3637, 0.4347, 0.5235, 0.3845],
device='cuda:0'),
Parameter containing:
tensor([-0.1759, -0.2156, -0.2047, -0.1695, -0.1628, -0.1473, -0.2158, -0.2905,
-0.1112, -0.2196, -0.1020, -0.1549, -0.1989, -0.0445, -0.1508, -0.1920,
-0.2114, -0.1655, -0.1854, -0.1733, -0.1289, -0.2376, -0.1965, -0.1965,
-0.1776, -0.1774, -0.1760, -0.1546, -0.1648, -0.2599, -0.1752, -0.2498,
-0.1741, -0.2410, -0.2498, -0.2938, -0.1496, -0.1578, -0.1800, -0.1851,
-0.1516, -0.1345, -0.2746, -0.1248, -0.2246, -0.2531, -0.2398, -0.1859,
-0.1739, -0.2393, -0.1214, -0.1803, -0.2729, -0.2617, -0.1855, -0.2316,
-0.2333, -0.1860, -0.2097, -0.0692, -0.1912, -0.2078, -0.1084, -0.2810,
-0.1303, -0.1654, -0.2119, -0.3641, -0.2951, -0.2384, -0.1632, -0.1892,
-0.1792, -0.2031, -0.1770, -0.2738, -0.3324, -0.1725, -0.1793, -0.2638,
-0.2207, -0.1609, -0.1534, -0.1414, -0.2992, -0.1450, -0.1838, -0.1779,
-0.1422, -0.2198, -0.1900, -0.1580, -0.1666, -0.2490, -0.1569, -0.1718,
-0.1660, -0.1972, -0.2287, -0.2366, -0.2230, -0.1543, -0.2030, -0.1431,
-0.1363, -0.2015, -0.1804, -0.2093, -0.2964, -0.1984, -0.2683, -0.2216,
-0.2147, -0.3404, -0.2668, -0.1890, -0.1733, -0.2226, -0.1772, -0.1698,
-0.1095, -0.2180, -0.1154, -0.1654, -0.1910, -0.3535, -0.3112, -0.2161,
-0.1496, -0.1667, -0.2849, -0.2207, -0.1529, -0.1807, -0.2118, -0.1869,
-0.1376, -0.1770, -0.1861, -0.1969, -0.1741, -0.3011, -0.0787, -0.2017,
-0.1947, -0.2247, -0.2459, -0.1058, -0.1401, -0.1213, -0.1199, -0.1760,
-0.2156, -0.3307, -0.3515, -0.2366, -0.1185, -0.2155, -0.1751, -0.1892,
-0.3365, -0.1598, -0.2554, 0.0644, -0.2856, -0.1198, -0.1583, -0.2297,
-0.3352, -0.1987, -0.2686, -0.1632, -0.2461, -0.2900, -0.2428, -0.1449,
-0.1900, -0.2149, -0.1541, -0.2917, -0.2504, -0.2213, -0.0463, -0.1547,
-0.1511, -0.1527, -0.1735, -0.1931, -0.1987, -0.2239, -0.2086, -0.2688,
-0.1845, -0.1797, -0.1833, -0.3880, -0.1539, -0.1553, -0.1567, -0.2238,
-0.1511, -0.2540, -0.2849, -0.1826, -0.2687, -0.2328, -0.2108, -0.2410,
-0.1022, -0.1507, -0.1978, -0.1734, -0.2282, -0.0985, -0.1847, -0.1770,
-0.1576, -0.1937, -0.1643, -0.2822, -0.1866, -0.2754, -0.2266, -0.2169,
-0.1352, -0.2194, -0.1060, -0.2139, -0.1322, -0.1889, -0.2130, -0.1913,
-0.2364, -0.1402, -0.2228, -0.2354, -0.1632, -0.1905, -0.1428, -0.1177,
-0.2419, -0.2733, -0.2963, -0.1600, -0.3558, -0.3673, -0.2201, -0.1505,
-0.2084, -0.0870, -0.2052, -0.2070, -0.1986, -0.2299, -0.0745, -0.1765,
-0.1412, -0.2180, -0.1450, -0.1426, -0.1452, -0.2916, -0.0871, -0.1359,
-0.2003, -0.1125, -0.2588, -0.1988, -0.2028, -0.2443, -0.0864, -0.3415,
-0.2579, -0.2343, -0.3552, -0.1859, -0.1153, -0.1732, -0.1780, -0.1909,
-0.2018, -0.1886, -0.2751, -0.1501, 0.1165, -0.1891, -0.1845, -0.2037,
-0.0339, -0.3464, -0.1956, -0.1962, -0.1537, -0.1902, -0.1431, -0.3022,
-0.1780, -0.1971, -0.2118, -0.0952, -0.1711, -0.2409, -0.2184, -0.2114,
-0.2042, -0.0566, -0.0700, -0.2081, -0.1872, -0.2079, -0.1540, -0.2266,
-0.1981, -0.1679, -0.2022, -0.2010, -0.1051, -0.1705, -0.2139, 0.0396,
-0.1077, -0.2745, -0.2690, -0.2603, -0.2819, -0.1917, -0.1940, -0.2944,
-0.1822, -0.2903, -0.1064, -0.2076, -0.2648, -0.3032, -0.2878, -0.1579,
-0.0071, -0.2142, -0.2022, -0.1516, -0.1123, 0.0246, -0.0978, -0.1382,
-0.1800, -0.3214, -0.2179, -0.1369, -0.0800, 0.0117, -0.1839, -0.1926,
-0.1614, -0.2769, -0.1909, -0.2101, -0.2305, -0.2055, -0.2017, -0.2741,
-0.1005, -0.3152, -0.1121, -0.1700, -0.1364, -0.2157, -0.2673, -0.1584,
-0.1997, -0.1745, -0.1886, -0.2307, -0.2024, -0.3376, -0.2266, -0.2355,
-0.2133, -0.2346, -0.2412, -0.2358, -0.1265, -0.2341, -0.1887, -0.1646,
-0.1417, -0.1882, -0.1076, -0.3048, -0.1162, -0.1651, -0.2046, -0.1833,
-0.3102, -0.1778, -0.1575, -0.2676, -0.1777, -0.1569, -0.1741, -0.1892,
-0.3028, -0.1457, -0.2179, -0.2226, -0.1609, -0.1423, -0.2683, -0.2920,
-0.1740, -0.2079, -0.1940, -0.2679, -0.1973, -0.1951, -0.1665, -0.2286,
-0.1903, -0.2667, -0.4010, -0.2550, -0.1817, -0.2025, -0.1589, -0.2476,
-0.0573, -0.2203, -0.2084, -0.1587, -0.1212, -0.1795, -0.3449, -0.1662,
-0.2523, -0.2435, -0.2878, -0.2797, -0.1897, -0.2113, -0.1943, -0.2050,
-0.1694, -0.2243, -0.2987, -0.1328, -0.1428, -0.2399, -0.1593, -0.1999,
-0.3225, -0.1860, -0.1763, -0.2691, -0.2097, -0.2396, -0.1140, -0.1897,
-0.1870, -0.1829, -0.2615, -0.2073, -0.1858, -0.0598, -0.1915, -0.2183,
-0.2088, -0.1742, -0.2715, -0.1999, -0.2117, -0.2492, -0.1717, -0.1566,
-0.1669, -0.3015, -0.1685, -0.2434, -0.2297, -0.1947, -0.2860, -0.3288,
-0.2197, -0.1862, -0.1755, -0.0987, -0.1756, -0.1304, -0.1555, -0.1679,
-0.2222, -0.2819, -0.2652, -0.0947, -0.2412, -0.2731, -0.2572, -0.2604,
-0.2934, -0.2470, -0.1820, -0.2740, -0.1336, -0.1698, -0.1919, -0.1796,
-0.2325, -0.1352, -0.1077, -0.2184, -0.1539, -0.2015, -0.3243, -0.1713],
device='cuda:0'),
Parameter containing:
tensor([[[[ 0.0057]],
[[ 0.0020]],
[[ 0.0167]],
...,
[[ 0.0085]],
[[-0.0274]],
[[ 0.0097]]],
[[[-0.0271]],
[[-0.0157]],
[[ 0.0543]],
...,
[[-0.0206]],
[[-0.0308]],
[[ 0.0013]]],
[[[-0.0523]],
[[-0.0353]],
[[ 0.0394]],
...,
[[-0.0382]],
[[-0.0264]],
[[-0.0443]]],
...,
[[[ 0.0494]],
[[ 0.0436]],
[[ 0.0103]],
...,
[[ 0.0072]],
[[ 0.0014]],
[[-0.0669]]],
[[[ 0.0533]],
[[-0.0148]],
[[-0.0480]],
...,
[[ 0.0055]],
[[ 0.0429]],
[[ 0.0129]]],
[[[ 0.0300]],
[[-0.0092]],
[[ 0.0090]],
...,
[[ 0.0101]],
[[-0.0111]],
[[-0.0080]]]], device='cuda:0'),
Parameter containing:
tensor([ 0.1694, 0.3368, 0.2993, 0.3745, 0.1513, 0.1781, 0.3167, 0.3947,
0.1858, 0.2068, 0.1090, 0.2042, 0.2955, 0.0765, 0.2023, 0.2487,
0.3295, 0.3349, 0.2532, 0.2739, 0.1661, 0.3432, 0.3424, 0.2969,
0.2226, 0.0993, 0.3328, 0.2349, 0.2894, 0.2296, 0.2719, 0.3945,
0.1990, 0.2564, 0.2557, 0.3541, 0.1848, 0.2513, 0.3101, 0.2782,
0.2109, 0.2441, 0.3282, 0.3248, 0.2499, 0.1873, 0.2643, 0.3949,
0.1962, 0.2587, 0.1708, 0.3381, 0.2238, 0.2498, 0.2787, 0.3783,
0.3445, 0.2681, 0.2956, 0.1146, 0.2688, 0.3479, 0.1295, 0.2843,
0.1552, 0.3026, 0.2738, 0.1891, 0.3568, 0.2302, 0.2199, 0.2070,
0.2119, 0.0971, 0.2482, 0.2264, 0.3555, 0.3113, 0.2386, 0.2654,
0.2975, 0.2666, 0.2180, 0.1451, 0.2460, 0.1734, 0.2358, 0.2891,
0.2091, 0.1971, 0.2185, 0.2008, 0.2461, 0.3726, 0.2028, 0.1993,
0.3652, 0.2258, 0.2606, 0.1900, 0.2764, 0.2011, 0.1973, 0.2958,
0.3222, 0.4117, 0.1475, 0.2674, 0.1928, 0.3615, 0.2774, 0.2143,
0.2688, 0.4286, 0.2560, 0.2777, 0.1339, 0.5103, 0.3238, 0.2417,
0.1529, 0.1843, 0.0579, 0.2288, 0.1797, 0.2803, 0.2279, 0.1579,
0.3196, 0.1842, 0.3378, 0.1688, 0.1654, 0.3049, 0.3533, 0.2948,
0.1140, 0.2503, 0.1892, 0.2647, 0.2405, 0.3880, 0.1933, 0.1918,
0.2511, 0.2901, 0.3151, 0.3252, 0.1296, 0.2491, 0.1417, 0.1295,
0.3062, 0.2836, 0.3483, 0.2306, 0.2741, 0.2700, 0.1873, 0.2431,
0.3526, 0.3546, 0.2721, 0.2708, 0.3065, 0.0832, 0.2968, 0.2286,
0.3276, 0.2695, 0.2452, 0.2444, 0.2857, 0.3365, 0.2784, 0.2933,
0.3397, 0.2231, 0.2330, 0.1486, 0.3846, 0.3104, 0.1724, 0.1724,
0.3466, 0.2978, 0.2582, 0.1879, 0.2419, 0.2249, 0.2720, 0.3735,
0.4259, 0.3754, 0.1731, 0.3698, 0.2349, 0.2694, 0.3148, 0.1658,
0.1181, 0.2994, 0.4018, 0.2126, 0.3864, 0.2955, 0.1848, 0.3686,
0.1972, 0.3265, 0.2319, 0.1676, 0.1756, 0.2367, 0.2139, 0.1974,
0.2561, 0.2619, 0.2170, 0.2284, 0.3486, 0.4500, 0.2563, 0.2559,
0.2814, 0.1797, 0.1736, 0.2013, 0.3411, 0.2245, 0.1385, 0.2284,
0.2230, 0.2566, 0.2301, 0.3639, 0.1380, 0.2381, 0.2590, 0.0830,
0.1863, 0.1267, 0.4501, 0.2741, 0.2590, 0.2782, 0.2248, 0.2718,
0.1949, 0.1815, 0.2969, 0.3168, 0.3389, 0.2790, 0.1594, 0.2752,
0.2947, 0.2909, 0.1418, 0.3336, 0.1953, 0.2646, 0.0879, 0.2553,
0.3335, 0.1943, 0.2777, 0.2386, 0.3676, 0.3042, 0.1234, 0.2615,
0.2548, 0.3224, 0.3462, 0.2090, 0.2142, 0.2054, 0.2115, 0.2153,
0.2163, 0.2509, 0.2429, 0.3326, -0.0527, 0.2244, 0.2319, 0.2674,
0.1103, 0.2320, 0.2822, 0.3234, 0.2818, 0.2093, 0.2261, 0.2900,
0.3127, 0.3456, 0.2592, 0.1677, 0.3924, 0.2694, 0.1997, 0.2973,
0.3324, 0.2270, 0.0656, 0.2964, 0.1948, 0.2383, 0.3021, 0.2510,
0.3117, 0.3185, 0.1721, 0.1867, 0.1665, 0.2851, 0.3512, -0.0486,
0.1558, 0.2213, 0.3281, 0.3861, 0.2375, 0.3057, 0.1178, 0.2681,
0.1921, 0.2211, 0.1679, 0.2877, 0.2495, 0.2451, 0.2678, 0.2393,
0.0988, 0.2778, 0.2465, 0.1747, 0.1005, 0.0502, 0.2809, 0.2810,
0.1716, 0.2114, 0.2213, 0.2817, 0.1506, 0.0769, 0.2381, 0.2411,
0.2942, 0.2543, 0.2556, 0.3451, 0.2948, 0.3040, 0.3204, 0.2757,
0.1657, 0.2941, 0.1301, 0.1854, 0.2866, 0.3198, 0.2127, 0.3608,
0.3440, 0.0954, 0.2586, 0.1709, 0.2007, 0.1967, 0.1972, 0.1942,
0.3201, 0.3484, 0.3437, 0.3153, 0.2020, 0.3251, 0.3227, 0.3038,
0.2634, 0.2364, 0.2492, 0.3080, 0.2591, 0.2391, 0.2720, 0.2601,
0.3210, 0.1818, 0.3526, 0.3579, 0.2861, 0.2526, 0.1642, 0.2897,
0.3996, 0.2651, 0.2031, 0.2502, 0.3694, 0.2085, 0.2804, 0.2233,
0.2309, 0.1609, 0.2369, 0.2116, 0.3549, 0.1635, 0.1642, 0.3072,
0.3077, 0.2152, 0.2821, 0.2857, 0.1701, 0.2305, 0.2134, 0.3189,
0.1061, 0.2628, 0.2608, 0.1749, 0.0820, 0.1815, 0.3566, 0.1204,
0.3159, 0.1595, 0.3790, 0.3272, 0.2086, 0.3096, 0.2253, 0.1456,
0.1346, 0.2304, 0.2913, 0.2727, 0.2027, 0.2688, 0.1958, 0.2277,
0.3036, 0.3250, 0.3000, 0.3328, 0.2417, 0.2665, 0.2473, 0.0913,
0.2503, 0.2543, 0.3710, 0.3321, 0.3693, 0.1099, 0.1701, 0.1758,
0.3888, 0.2206, 0.2766, 0.2813, 0.1755, 0.2616, 0.1544, 0.2519,
0.1945, 0.2452, 0.3405, 0.2446, 0.2426, 0.1822, 0.3002, 0.3037,
0.3118, 0.2414, 0.2326, 0.1303, 0.3081, 0.0979, 0.2776, 0.2918,
0.3848, 0.1789, 0.3622, 0.3005, 0.1923, 0.2672, 0.1663, 0.2998,
0.2710, 0.2040, 0.2565, 0.2289, 0.2552, 0.2121, 0.3532, 0.2293,
0.2510, 0.3085, 0.2368, 0.3000, 0.2111, 0.3456, 0.3422, 0.1576],
device='cuda:0'),
Parameter containing:
tensor([-0.1759, -0.2156, -0.2047, -0.1695, -0.1628, -0.1473, -0.2158, -0.2905,
-0.1112, -0.2196, -0.1020, -0.1549, -0.1989, -0.0445, -0.1508, -0.1920,
-0.2114, -0.1655, -0.1854, -0.1733, -0.1289, -0.2376, -0.1965, -0.1965,
-0.1776, -0.1774, -0.1760, -0.1546, -0.1648, -0.2599, -0.1752, -0.2498,
-0.1741, -0.2410, -0.2498, -0.2938, -0.1496, -0.1578, -0.1800, -0.1851,
-0.1516, -0.1345, -0.2746, -0.1248, -0.2246, -0.2531, -0.2398, -0.1859,
-0.1739, -0.2393, -0.1214, -0.1803, -0.2729, -0.2617, -0.1855, -0.2316,
-0.2333, -0.1860, -0.2097, -0.0692, -0.1912, -0.2078, -0.1084, -0.2810,
-0.1303, -0.1654, -0.2119, -0.3641, -0.2951, -0.2384, -0.1632, -0.1892,
-0.1792, -0.2031, -0.1770, -0.2738, -0.3324, -0.1725, -0.1793, -0.2638,
-0.2207, -0.1609, -0.1534, -0.1414, -0.2992, -0.1450, -0.1838, -0.1779,
-0.1422, -0.2198, -0.1900, -0.1580, -0.1666, -0.2490, -0.1569, -0.1718,
-0.1660, -0.1972, -0.2287, -0.2366, -0.2230, -0.1543, -0.2030, -0.1431,
-0.1363, -0.2015, -0.1804, -0.2093, -0.2964, -0.1984, -0.2683, -0.2216,
-0.2147, -0.3404, -0.2668, -0.1890, -0.1733, -0.2226, -0.1772, -0.1698,
-0.1095, -0.2180, -0.1154, -0.1654, -0.1910, -0.3535, -0.3112, -0.2161,
-0.1496, -0.1667, -0.2849, -0.2207, -0.1529, -0.1807, -0.2118, -0.1869,
-0.1376, -0.1770, -0.1861, -0.1969, -0.1741, -0.3011, -0.0787, -0.2017,
-0.1947, -0.2247, -0.2459, -0.1058, -0.1401, -0.1213, -0.1199, -0.1760,
-0.2156, -0.3307, -0.3515, -0.2366, -0.1185, -0.2155, -0.1751, -0.1892,
-0.3365, -0.1598, -0.2554, 0.0644, -0.2856, -0.1198, -0.1583, -0.2297,
-0.3352, -0.1987, -0.2686, -0.1632, -0.2461, -0.2900, -0.2428, -0.1449,
-0.1900, -0.2149, -0.1541, -0.2917, -0.2504, -0.2213, -0.0463, -0.1547,
-0.1511, -0.1527, -0.1735, -0.1931, -0.1987, -0.2239, -0.2086, -0.2688,
-0.1845, -0.1797, -0.1833, -0.3880, -0.1539, -0.1553, -0.1567, -0.2238,
-0.1511, -0.2540, -0.2849, -0.1826, -0.2687, -0.2328, -0.2108, -0.2410,
-0.1022, -0.1507, -0.1978, -0.1734, -0.2282, -0.0985, -0.1847, -0.1770,
-0.1576, -0.1937, -0.1643, -0.2822, -0.1866, -0.2754, -0.2266, -0.2169,
-0.1352, -0.2194, -0.1060, -0.2139, -0.1322, -0.1889, -0.2130, -0.1913,
-0.2364, -0.1402, -0.2228, -0.2354, -0.1632, -0.1905, -0.1428, -0.1177,
-0.2419, -0.2733, -0.2963, -0.1600, -0.3558, -0.3673, -0.2201, -0.1505,
-0.2084, -0.0870, -0.2052, -0.2070, -0.1986, -0.2299, -0.0745, -0.1765,
-0.1412, -0.2180, -0.1450, -0.1426, -0.1452, -0.2916, -0.0871, -0.1359,
-0.2003, -0.1125, -0.2588, -0.1988, -0.2028, -0.2443, -0.0864, -0.3415,
-0.2579, -0.2343, -0.3552, -0.1859, -0.1153, -0.1732, -0.1780, -0.1909,
-0.2018, -0.1886, -0.2751, -0.1501, 0.1165, -0.1891, -0.1845, -0.2037,
-0.0339, -0.3464, -0.1956, -0.1962, -0.1537, -0.1902, -0.1431, -0.3022,
-0.1780, -0.1971, -0.2118, -0.0952, -0.1711, -0.2409, -0.2184, -0.2114,
-0.2042, -0.0566, -0.0700, -0.2081, -0.1872, -0.2079, -0.1540, -0.2266,
-0.1981, -0.1679, -0.2022, -0.2010, -0.1051, -0.1705, -0.2139, 0.0396,
-0.1077, -0.2745, -0.2690, -0.2603, -0.2819, -0.1917, -0.1940, -0.2944,
-0.1822, -0.2903, -0.1064, -0.2076, -0.2648, -0.3032, -0.2878, -0.1579,
-0.0071, -0.2142, -0.2022, -0.1516, -0.1123, 0.0246, -0.0978, -0.1382,
-0.1800, -0.3214, -0.2179, -0.1369, -0.0800, 0.0117, -0.1839, -0.1926,
-0.1614, -0.2769, -0.1909, -0.2101, -0.2305, -0.2055, -0.2017, -0.2741,
-0.1005, -0.3152, -0.1121, -0.1700, -0.1364, -0.2157, -0.2673, -0.1584,
-0.1997, -0.1745, -0.1886, -0.2307, -0.2024, -0.3376, -0.2266, -0.2355,
-0.2133, -0.2346, -0.2412, -0.2358, -0.1265, -0.2341, -0.1887, -0.1646,
-0.1417, -0.1882, -0.1076, -0.3048, -0.1162, -0.1651, -0.2046, -0.1833,
-0.3102, -0.1778, -0.1575, -0.2676, -0.1777, -0.1569, -0.1741, -0.1892,
-0.3028, -0.1457, -0.2179, -0.2226, -0.1609, -0.1423, -0.2683, -0.2920,
-0.1740, -0.2079, -0.1940, -0.2679, -0.1973, -0.1951, -0.1665, -0.2286,
-0.1903, -0.2667, -0.4010, -0.2550, -0.1817, -0.2025, -0.1589, -0.2476,
-0.0573, -0.2203, -0.2084, -0.1587, -0.1212, -0.1795, -0.3449, -0.1662,
-0.2523, -0.2435, -0.2878, -0.2797, -0.1897, -0.2113, -0.1943, -0.2050,
-0.1694, -0.2243, -0.2987, -0.1328, -0.1428, -0.2399, -0.1593, -0.1999,
-0.3225, -0.1860, -0.1763, -0.2691, -0.2097, -0.2396, -0.1140, -0.1897,
-0.1870, -0.1829, -0.2615, -0.2073, -0.1858, -0.0598, -0.1915, -0.2183,
-0.2088, -0.1742, -0.2715, -0.1999, -0.2117, -0.2492, -0.1717, -0.1566,
-0.1669, -0.3015, -0.1685, -0.2434, -0.2297, -0.1947, -0.2860, -0.3288,
-0.2197, -0.1862, -0.1755, -0.0987, -0.1756, -0.1304, -0.1555, -0.1679,
-0.2222, -0.2819, -0.2652, -0.0947, -0.2412, -0.2731, -0.2572, -0.2604,
-0.2934, -0.2470, -0.1820, -0.2740, -0.1336, -0.1698, -0.1919, -0.1796,
-0.2325, -0.1352, -0.1077, -0.2184, -0.1539, -0.2015, -0.3243, -0.1713],
device='cuda:0'),
Parameter containing:
tensor([[[[-8.0284e-03, -5.7776e-03, 6.4154e-03],
[ 5.0498e-03, -6.7796e-03, 1.2691e-02],
[ 1.3331e-02, 1.4523e-02, 2.4522e-02]],
[[-1.9876e-03, 1.2466e-02, 1.0494e-02],
[-1.9364e-02, -1.6696e-02, -1.1857e-02],
[-1.1569e-02, -3.7674e-03, -3.4679e-03]],
[[-1.1440e-02, -1.3884e-02, 1.1559e-03],
[-1.7906e-02, -2.9349e-02, -1.3876e-02],
[-1.4057e-02, -2.6989e-02, -2.3963e-02]],
...,
[[-6.3040e-03, -3.1167e-03, -1.3304e-02],
[ 7.1623e-03, 6.4669e-03, 1.6063e-02],
[-1.0750e-02, -1.0480e-02, -6.1070e-03]],
[[ 7.4484e-03, 6.3878e-03, -1.2579e-02],
[-7.7356e-03, 1.8112e-03, -1.7890e-02],
[-2.9142e-03, 7.7705e-03, -9.7314e-03]],
[[ 2.1760e-02, 2.2364e-02, 2.2731e-02],
[ 2.6681e-02, 2.9127e-02, 3.3356e-02],
[ 1.2892e-02, -3.5818e-03, 5.3022e-03]]],
[[[-1.0597e-02, -9.1551e-03, -2.3418e-02],
[-1.0768e-02, -3.3171e-03, -1.8559e-02],
[-1.8607e-02, -4.2634e-03, -1.5591e-02]],
[[-2.6090e-02, -2.2517e-02, -3.0593e-02],
[-3.9406e-02, -2.6639e-02, -2.8202e-02],
[-2.6143e-02, -1.9647e-02, -2.1466e-02]],
[[-3.5259e-03, 1.6623e-03, -6.5624e-03],
[-5.0597e-03, -8.7162e-04, -5.3742e-03],
[-7.9651e-03, -9.7778e-03, -1.0736e-02]],
...,
[[ 1.8492e-02, -3.6799e-03, 1.0043e-02],
[-5.2974e-03, -2.0757e-02, -1.5120e-02],
[ 2.1435e-02, 6.4916e-03, 4.7660e-03]],
[[-1.8810e-02, -6.0469e-04, -7.6999e-03],
[-1.7697e-02, -7.8692e-03, -1.6543e-02],
[-1.7206e-02, -2.4746e-02, -3.0270e-02]],
[[-3.1191e-02, -1.4363e-02, 2.2032e-03],
[-1.2033e-02, -2.3699e-03, -1.6630e-02],
[-1.2905e-02, -1.5363e-02, -3.6297e-03]]],
[[[-3.2648e-02, -4.8158e-03, -2.0476e-02],
[-2.5846e-02, -1.4660e-03, -2.8170e-02],
[-2.6640e-02, 4.3022e-03, -2.7636e-02]],
[[-6.3289e-03, -1.5401e-02, -1.3096e-03],
[-1.7499e-02, -2.6212e-02, -2.3646e-02],
[-7.3207e-03, -1.5592e-02, -8.9578e-03]],
[[ 8.9701e-04, -6.6914e-03, -5.3129e-03],
[-1.1727e-03, -1.0726e-02, -9.0103e-03],
[ 3.2311e-03, -4.5854e-03, 4.3512e-03]],
...,
[[-2.1822e-02, -3.6889e-02, -2.2588e-02],
[-1.3054e-02, -3.4191e-02, -2.7238e-02],
[-1.2383e-02, -2.3452e-02, -2.2486e-02]],
[[ 6.8177e-03, 2.1561e-02, 1.3674e-02],
[ 3.1192e-03, 1.0660e-02, 1.0409e-02],
[ 8.0477e-03, -4.6817e-03, -4.3912e-03]],
[[-1.1983e-02, -1.6201e-02, -2.2626e-02],
[-1.3461e-02, -7.0928e-03, -1.4384e-02],
[-2.4456e-02, 1.4885e-02, 1.2247e-02]]],
...,
[[[-2.6347e-02, -2.9923e-02, -3.7810e-02],
[-1.5663e-02, -4.1126e-03, -1.1482e-02],
[-1.3415e-02, -1.5432e-02, -1.8204e-02]],
[[-3.8392e-03, -1.1093e-02, -8.0841e-04],
[-5.9634e-03, -5.9165e-03, -9.3332e-03],
[-2.2761e-03, 5.4781e-03, -5.6050e-03]],
[[-1.8406e-03, -2.8134e-03, 8.3246e-03],
[-1.2453e-03, 2.1453e-04, 7.4868e-03],
[ 1.3450e-02, 3.0599e-02, 2.6405e-02]],
...,
[[ 3.5268e-04, 2.3897e-03, 6.2558e-03],
[-1.4338e-02, -2.3146e-02, -1.9024e-02],
[-2.7306e-02, -3.0079e-02, -3.1762e-02]],
[[ 1.4584e-02, 4.3430e-03, 1.2053e-02],
[-6.1130e-03, -2.8539e-02, -1.8268e-02],
[-1.6844e-02, -4.7816e-02, -2.6274e-02]],
[[-1.8850e-02, -9.3396e-03, 7.8905e-03],
[-1.5322e-03, 8.3153e-03, 1.7783e-02],
[-8.3318e-03, -1.5759e-02, -1.2061e-02]]],
[[[ 9.9578e-03, 7.4573e-03, -1.8738e-03],
[-1.7752e-03, -6.8015e-04, -7.4443e-03],
[-1.8319e-02, -1.4264e-02, -7.1446e-03]],
[[ 7.8524e-03, -2.6520e-03, -1.7556e-02],
[ 4.5240e-03, -4.8661e-03, -1.5215e-02],
[-5.0211e-03, -1.1864e-02, -1.4846e-02]],
[[ 2.9163e-02, 1.0344e-02, 2.4736e-02],
[ 1.2012e-02, -1.0346e-02, 3.5472e-03],
[ 8.2238e-03, -1.8237e-02, -5.4892e-03]],
...,
[[-8.8434e-03, -4.3184e-03, -5.7536e-03],
[ 7.7230e-03, -4.1936e-04, 7.7260e-03],
[ 1.3536e-02, 1.5705e-02, 2.0893e-02]],
[[ 1.6743e-03, 1.9720e-03, 2.1567e-02],
[-8.0074e-03, -4.6606e-03, 4.0560e-03],
[-1.6688e-02, -1.3754e-02, -1.1708e-02]],
[[-9.7959e-03, -9.4502e-03, -9.3443e-03],
[ 6.9547e-03, -3.9134e-05, 6.2691e-03],
[-1.3193e-02, 9.3272e-04, 1.4579e-02]]],
[[[-1.4963e-03, 5.5133e-04, 1.1571e-02],
[ 1.0174e-02, 1.7889e-03, 1.1035e-02],
[ 7.0212e-03, 1.4651e-03, 1.2769e-03]],
[[-1.3021e-02, 6.4109e-03, -1.5199e-02],
[ 2.4775e-02, 2.1926e-02, 3.3679e-02],
[ 2.6471e-04, -3.0235e-03, 1.1690e-02]],
[[-2.9665e-02, -1.5314e-02, -1.7500e-02],
[-1.8339e-02, -2.0845e-02, -1.5494e-02],
[-1.6086e-03, 1.0831e-02, -1.4309e-02]],
...,
[[-7.7044e-03, -2.1100e-02, -2.2816e-02],
[ 5.7688e-03, 1.9362e-04, 7.7105e-04],
[-6.1357e-03, 9.7275e-03, -2.5464e-03]],
[[ 1.1043e-02, 2.4205e-02, 3.4213e-02],
[ 2.9181e-02, 2.6904e-02, 4.5372e-02],
[-2.1594e-02, -1.1072e-03, -7.8312e-03]],
[[-8.3287e-03, -7.9521e-03, -5.3358e-03],
[-6.2527e-04, -5.3243e-03, -8.6296e-03],
[ 3.6094e-03, -1.2544e-03, -4.3801e-03]]]], device='cuda:0'),
Parameter containing:
tensor([0.2587, 0.3073, 0.2595, 0.3223, 0.2662, 0.2652, 0.2575, 0.2660, 0.2766,
0.2414, 0.3045, 0.2853, 0.2821, 0.2880, 0.3094, 0.3444, 0.3155, 0.4129,
0.2110, 0.2903, 0.2496, 0.2601, 0.2967, 0.3033, 0.4152, 0.2719, 0.3661,
0.3251, 0.3898, 0.3346, 0.2753, 0.2712, 0.2414, 0.3351, 0.3394, 0.3167,
0.3360, 0.2666, 0.2109, 0.2705, 0.2587, 0.3070, 0.2720, 0.2316, 0.2885,
0.2884, 0.2955, 0.3057, 0.3043, 0.2596, 0.2673, 0.1929, 0.3136, 0.3593,
0.2622, 0.2931, 0.3295, 0.2514, 0.3208, 0.2798, 0.3259, 0.2939, 0.2390,
0.3105, 0.3471, 0.2812, 0.2148, 0.2997, 0.3061, 0.2740, 0.2791, 0.3790,
0.3592, 0.3247, 0.2995, 0.2735, 0.3356, 0.2703, 0.3255, 0.3127, 0.2783,
0.2702, 0.3900, 0.2942, 0.2899, 0.3461, 0.3432, 0.4685, 0.2634, 0.2553,
0.3019, 0.3961, 0.2742, 0.2995, 0.3858, 0.2785, 0.3212, 0.3109, 0.3642,
0.2193, 0.2643, 0.2333, 0.3151, 0.3102, 0.2936, 0.2374, 0.2419, 0.2976,
0.3335, 0.2619, 0.3984, 0.2721, 0.2718, 0.2678, 0.2757, 0.2445, 0.3508,
0.2174, 0.3309, 0.2653, 0.2564, 0.1748, 0.3177, 0.2751, 0.2067, 0.2905,
0.2762, 0.3329, 0.2738, 0.3224, 0.2199, 0.2997, 0.2206, 0.3213, 0.2760,
0.3927, 0.3174, 0.2698, 0.2988, 0.2610, 0.2550, 0.2788, 0.4445, 0.2862,
0.3606, 0.3279, 0.2869, 0.3294, 0.2244, 0.2338, 0.1754, 0.2318, 0.3186,
0.3322, 0.2255, 0.3041, 0.2837, 0.3276, 0.2392, 0.3668, 0.1971, 0.2946,
0.3613, 0.2736, 0.2554, 0.2860, 0.2511, 0.3490, 0.3253, 0.2934, 0.2027,
0.2580, 0.2200, 0.3089, 0.3074, 0.3332, 0.2943, 0.3375, 0.2330, 0.2611,
0.3383, 0.2837, 0.3546, 0.3093, 0.3791, 0.2197, 0.2648, 0.2830, 0.2587,
0.3588, 0.2830, 0.3971, 0.3194, 0.3066, 0.2754, 0.2647, 0.0970, 0.2182,
0.2334, 0.2624, 0.1829, 0.2933, 0.2747, 0.3001, 0.2996, 0.3107, 0.3256,
0.2940, 0.3901, 0.2790, 0.3030, 0.2838, 0.3010, 0.3044, 0.3479, 0.3087,
0.2611, 0.1958, 0.2941, 0.2558, 0.2889, 0.3148, 0.2516, 0.2664, 0.2862,
0.3940, 0.2933, 0.2781, 0.3796, 0.3022, 0.2583, 0.3021, 0.2784, 0.2967,
0.2994, 0.3856, 0.3277, 0.2587, 0.2539, 0.2824, 0.2634, 0.1489, 0.2205,
0.3929, 0.3401, 0.2717, 0.2789, 0.2917, 0.3177, 0.1992, 0.3684, 0.3120,
0.3201, 0.2810, 0.2302, 0.2779, 0.2865, 0.2858, 0.2713, 0.1601, 0.2496,
0.2895, 0.3154, 0.3443, 0.3285, 0.3444, 0.3251, 0.3235, 0.3375, 0.2282,
0.2128, 0.1795, 0.3077, 0.3005, 0.2775, 0.3054, 0.2914, 0.3535, 0.2871,
0.2669, 0.3961, 0.2674, 0.3898, 0.3183, 0.3242, 0.2789, 0.1911, 0.2569,
0.3427, 0.2464, 0.2778, 0.2098, 0.3019, 0.3145, 0.3271, 0.2914, 0.2619,
0.2643, 0.3039, 0.2520, 0.2099, 0.3643, 0.2915, 0.1957, 0.3286, 0.2355,
0.3210, 0.2982, 0.3388, 0.3450, 0.3716, 0.2898, 0.2846, 0.2805, 0.2219,
0.2910, 0.2681, 0.3163, 0.1964, 0.3176, 0.3092, 0.2706, 0.2505, 0.2508,
0.3166, 0.3583, 0.1563, 0.2608, 0.2892, 0.3401, 0.2891, 0.3126, 0.2172,
0.2459, 0.2651, 0.4052, 0.2986, 0.3026, 0.3773, 0.2262, 0.2675, 0.2900,
0.3759, 0.3201, 0.2567, 0.3443, 0.2348, 0.3057, 0.2347, 0.3277, 0.2938,
0.2746, 0.2805, 0.2421, 0.3590, 0.2622, 0.2773, 0.2396, 0.2134, 0.2727,
0.2984, 0.2744, 0.2591, 0.2628, 0.3568, 0.2009, 0.3220, 0.2868, 0.2561,
0.3113, 0.2138, 0.3136, 0.2745, 0.3046, 0.3042, 0.1972, 0.2815, 0.2542,
0.2983, 0.2613, 0.2668, 0.3142, 0.2930, 0.3800, 0.1966, 0.2948, 0.3363,
0.2713, 0.3625, 0.2909, 0.2695, 0.3111, 0.3242, 0.3009, 0.3231, 0.3051,
0.2012, 0.2716, 0.3692, 0.2694, 0.1481, 0.2858, 0.2819, 0.2391, 0.2867,
0.3466, 0.3431, 0.2365, 0.3357, 0.1685, 0.2925, 0.3092, 0.3127, 0.1883,
0.2561, 0.3086, 0.1732, 0.2989, 0.3235, 0.2693, 0.2630, 0.2913, 0.2786,
0.3124, 0.3098, 0.2695, 0.2403, 0.2906, 0.2784, 0.2654, 0.3485, 0.3939,
0.3033, 0.3145, 0.2622, 0.1540, 0.2790, 0.2967, 0.1954, 0.2632, 0.2957,
0.2581, 0.3231, 0.2795, 0.2859, 0.3139, 0.2488, 0.2404, 0.3714, 0.2649,
0.2267, 0.2878, 0.3462, 0.3063, 0.3180, 0.1726, 0.3153, 0.2625, 0.3020,
0.2996, 0.3632, 0.1541, 0.3192, 0.2200, 0.2894, 0.2622, 0.2534, 0.2935,
0.3208, 0.2231, 0.2743, 0.3023, 0.2829, 0.2394, 0.2506, 0.3512, 0.3366,
0.2666, 0.2930, 0.3049, 0.2321, 0.3397, 0.2727, 0.2900, 0.3146, 0.2682,
0.3094, 0.3718, 0.3387, 0.3202, 0.2423, 0.2745, 0.2966, 0.2500, 0.2329,
0.3419, 0.2928, 0.3536, 0.3739, 0.1935, 0.2670, 0.2846, 0.2583, 0.3783,
0.2826, 0.2929, 0.2728, 0.3645, 0.2770, 0.2756, 0.2523, 0.2500],
device='cuda:0'),
Parameter containing:
tensor([-0.1668, -0.3019, -0.2187, -0.2917, -0.1971, -0.2325, -0.1869, -0.1857,
-0.2474, -0.1629, -0.2448, -0.2508, -0.1895, -0.2651, -0.3250, -0.3811,
-0.2953, -0.4963, -0.0294, -0.2724, -0.2007, -0.2220, -0.2945, -0.2579,
-0.5152, -0.1994, -0.5016, -0.2736, -0.4528, -0.3968, -0.2281, -0.1772,
-0.1293, -0.2655, -0.3252, -0.3232, -0.3337, -0.1901, -0.0692, -0.2196,
-0.2132, -0.2565, -0.1646, -0.1567, -0.2087, -0.2178, -0.2480, -0.2767,
-0.3071, -0.1988, -0.1985, -0.0235, -0.2458, -0.4156, -0.1660, -0.1923,
-0.3328, -0.1481, -0.3047, -0.2277, -0.3182, -0.2744, -0.1643, -0.3365,
-0.4050, -0.2082, -0.0621, -0.2671, -0.2809, -0.2185, -0.2148, -0.4465,
-0.3376, -0.3213, -0.2921, -0.1998, -0.3369, -0.2092, -0.2831, -0.2893,
-0.1719, -0.2189, -0.4016, -0.2484, -0.2070, -0.3849, -0.3753, -0.5874,
-0.1637, -0.1748, -0.2217, -0.5067, -0.2496, -0.2117, -0.4291, -0.1944,
-0.3089, -0.2621, -0.4096, -0.0602, -0.2009, -0.1316, -0.3336, -0.2627,
-0.2320, -0.0910, -0.1560, -0.2889, -0.3286, -0.1628, -0.5128, -0.2036,
-0.1726, -0.1844, -0.2285, -0.1925, -0.3432, -0.0929, -0.3138, -0.1912,
-0.1926, -0.0342, -0.3268, -0.1699, -0.0828, -0.2417, -0.2069, -0.3870,
-0.2210, -0.2867, -0.0526, -0.3092, -0.0655, -0.2594, -0.2160, -0.5062,
-0.2905, -0.2125, -0.3124, -0.2128, -0.1946, -0.2520, -0.5475, -0.2321,
-0.3350, -0.3473, -0.2158, -0.3603, -0.0759, -0.1472, -0.0327, -0.1404,
-0.3128, -0.3063, -0.1120, -0.2664, -0.2700, -0.3112, -0.1519, -0.3843,
-0.0645, -0.2373, -0.4227, -0.2546, -0.1611, -0.2350, -0.1524, -0.3494,
-0.3453, -0.2081, -0.0918, -0.2025, -0.1246, -0.2533, -0.2768, -0.3156,
-0.2530, -0.3957, -0.0981, -0.1257, -0.3697, -0.2333, -0.3664, -0.2829,
-0.4320, -0.0836, -0.1583, -0.2395, -0.1818, -0.4408, -0.2376, -0.4450,
-0.3232, -0.2787, -0.1858, -0.2137, 0.0481, -0.1058, -0.1093, -0.2035,
-0.0496, -0.2117, -0.1598, -0.2389, -0.2830, -0.2878, -0.3406, -0.2560,
-0.4468, -0.2444, -0.2492, -0.2222, -0.2792, -0.3005, -0.4180, -0.2568,
-0.1872, -0.0270, -0.2645, -0.1873, -0.3022, -0.3400, -0.1803, -0.1810,
-0.2079, -0.4775, -0.2047, -0.1878, -0.4504, -0.2516, -0.1657, -0.2765,
-0.2329, -0.2446, -0.2956, -0.4163, -0.2816, -0.1571, -0.2199, -0.2125,
-0.1684, 0.0356, -0.0914, -0.4484, -0.3535, -0.2212, -0.2550, -0.2509,
-0.2702, -0.0599, -0.3505, -0.2924, -0.2360, -0.2339, -0.1259, -0.2597,
-0.2267, -0.1978, -0.1371, -0.0129, -0.1175, -0.2527, -0.3099, -0.3231,
-0.3468, -0.3553, -0.3537, -0.3315, -0.3713, -0.1091, -0.0959, -0.0258,
-0.2756, -0.2808, -0.2012, -0.2812, -0.1991, -0.3948, -0.2257, -0.2469,
-0.4211, -0.2110, -0.4670, -0.3069, -0.3549, -0.2337, -0.0612, -0.1321,
-0.2968, -0.1870, -0.2316, -0.0686, -0.3113, -0.2895, -0.3149, -0.2686,
-0.2081, -0.2096, -0.3011, -0.1810, -0.0227, -0.3873, -0.2665, -0.0225,
-0.2973, -0.0973, -0.2980, -0.3219, -0.2926, -0.3196, -0.4332, -0.1980,
-0.2117, -0.2302, -0.0980, -0.2344, -0.2154, -0.2921, -0.0350, -0.3361,
-0.2620, -0.2188, -0.1566, -0.1795, -0.2726, -0.4103, 0.0413, -0.1507,
-0.2552, -0.3137, -0.2466, -0.2961, -0.0938, -0.1481, -0.2129, -0.5480,
-0.2915, -0.2802, -0.5077, -0.1306, -0.1862, -0.2400, -0.4362, -0.3017,
-0.1633, -0.3447, -0.1047, -0.2846, -0.1244, -0.3036, -0.2404, -0.2333,
-0.2494, -0.1866, -0.3294, -0.1677, -0.2540, -0.1295, -0.0512, -0.1966,
-0.2801, -0.1702, -0.1879, -0.1850, -0.3274, -0.0369, -0.2979, -0.2612,
-0.1889, -0.3270, -0.1377, -0.2787, -0.2201, -0.2417, -0.2834, -0.0555,
-0.2538, -0.1040, -0.2660, -0.1644, -0.1723, -0.2672, -0.2797, -0.4214,
-0.0378, -0.2386, -0.3498, -0.2435, -0.4348, -0.2554, -0.1719, -0.2836,
-0.3316, -0.2787, -0.2879, -0.2640, -0.0560, -0.1789, -0.4195, -0.2152,
0.0567, -0.2359, -0.2249, -0.0911, -0.2644, -0.3875, -0.3317, -0.1415,
-0.3425, -0.0020, -0.1941, -0.2821, -0.2809, -0.0965, -0.1841, -0.2971,
-0.0173, -0.3043, -0.3013, -0.1729, -0.1872, -0.2683, -0.2033, -0.3059,
-0.2939, -0.2163, -0.1889, -0.2581, -0.2296, -0.2066, -0.3462, -0.4298,
-0.2600, -0.3095, -0.1800, -0.0116, -0.2124, -0.2552, -0.0523, -0.2216,
-0.2605, -0.2134, -0.2867, -0.2556, -0.2275, -0.3437, -0.1698, -0.1560,
-0.4120, -0.2067, -0.1159, -0.2408, -0.3093, -0.2621, -0.2593, -0.0135,
-0.3099, -0.2179, -0.2766, -0.2400, -0.3934, 0.0072, -0.2982, -0.0930,
-0.2166, -0.1635, -0.1827, -0.2308, -0.2525, -0.0991, -0.2325, -0.2938,
-0.2480, -0.0934, -0.1911, -0.3772, -0.3369, -0.1606, -0.2752, -0.3005,
-0.1372, -0.2990, -0.2156, -0.2622, -0.3160, -0.1342, -0.2903, -0.3865,
-0.2916, -0.3243, -0.2051, -0.2656, -0.2359, -0.1508, -0.1063, -0.3595,
-0.2312, -0.3046, -0.4178, -0.0276, -0.2204, -0.2426, -0.1616, -0.4789,
-0.1713, -0.2802, -0.2305, -0.4327, -0.2413, -0.1862, -0.1486, -0.1507],
device='cuda:0'),
Parameter containing:
tensor([[[[ 2.8729e-04, 4.2632e-03, -2.0266e-03],
[ 1.9513e-04, 2.4381e-03, -5.8632e-03],
[ 4.4803e-03, 8.6577e-03, 8.5538e-04]],
[[-1.1335e-02, -1.3195e-02, -1.0305e-02],
[-4.9507e-03, -4.5898e-03, -3.1041e-03],
[-7.5883e-03, -8.3795e-03, -8.9239e-03]],
[[-1.1914e-02, -1.2104e-02, -1.0167e-02],
[-1.2093e-02, -1.1557e-02, -8.9600e-03],
[-1.2515e-02, -9.3296e-03, -6.4079e-03]],
...,
[[-9.3573e-03, -1.0662e-02, -1.2672e-02],
[-8.0600e-03, -8.5423e-03, -1.2121e-02],
[-8.1498e-03, -8.8037e-03, -1.0611e-02]],
[[ 4.2632e-03, 5.6461e-03, 2.8460e-03],
[ 4.7070e-03, 6.2550e-03, 7.5862e-03],
[ 1.1504e-02, 1.1518e-02, 1.0728e-02]],
[[-6.2455e-03, -9.1693e-03, -9.6664e-03],
[-4.2935e-03, -6.5311e-03, -5.0513e-03],
[-3.1141e-03, -5.0124e-03, -5.8122e-03]]],
[[[ 2.7483e-03, 3.7146e-04, 3.3262e-05],
[-4.5675e-03, -6.6689e-03, -6.4447e-03],
[-6.7610e-03, -7.3204e-03, -9.5855e-03]],
[[-1.4630e-02, -1.2320e-02, -1.4457e-02],
[-8.6197e-03, -5.8059e-03, -1.1075e-02],
[-6.2154e-03, -6.8218e-03, -9.3805e-03]],
[[ 1.0879e-03, 4.3850e-04, -1.9456e-03],
[-1.2517e-03, 3.2917e-04, -2.1435e-03],
[ 4.8136e-03, 2.5333e-03, 5.1504e-03]],
...,
[[ 2.4644e-02, 1.7434e-02, 2.0734e-02],
[ 2.3101e-02, 1.3487e-02, 2.0728e-02],
[ 1.9381e-02, 1.5243e-02, 1.7340e-02]],
[[ 1.2212e-02, 1.2448e-02, 1.5048e-02],
[ 5.2993e-03, 4.0090e-03, 9.3927e-03],
[ 6.6766e-03, 2.4941e-03, 8.3288e-03]],
[[ 3.1040e-02, 2.8243e-02, 3.2319e-02],
[ 3.8608e-02, 3.3099e-02, 3.8652e-02],
[ 2.5839e-02, 2.6524e-02, 2.4995e-02]]],
[[[-2.1761e-03, 4.5553e-03, 2.0612e-03],
[ 4.9747e-03, 1.1420e-02, 8.5734e-03],
[ 4.8583e-03, 1.1469e-02, 1.0039e-02]],
[[-6.2547e-05, 6.5336e-04, 9.4747e-04],
[ 5.0603e-03, 7.7136e-03, 6.5484e-03],
[-4.8432e-04, 2.3057e-03, 2.9219e-03]],
[[-3.2788e-02, -2.7615e-02, -3.2608e-02],
[-3.6296e-02, -2.8170e-02, -3.0277e-02],
[-3.6814e-02, -3.1547e-02, -3.0231e-02]],
...,
[[-5.2998e-03, -2.8590e-04, -4.9266e-03],
[-7.0530e-03, -2.3684e-04, -1.5838e-03],
[-6.9291e-03, 4.8084e-04, -3.1548e-03]],
[[ 1.1854e-02, 8.4836e-03, 1.3839e-02],
[ 2.8741e-03, -9.7358e-05, 4.4888e-03],
[-2.5515e-03, -2.7788e-03, -3.2464e-03]],
[[-1.2408e-02, -1.5001e-02, -1.3377e-02],
[-1.4540e-02, -1.8537e-02, -1.7392e-02],
[-6.7315e-03, -9.5205e-03, -9.0692e-03]]],
...,
[[[ 3.0369e-03, 1.9542e-03, 1.7140e-03],
[-7.6240e-03, -2.8765e-03, -5.1760e-03],
[-9.3019e-03, -4.8800e-03, -4.2932e-03]],
[[ 4.4836e-03, 2.4909e-03, 1.5746e-03],
[ 1.2065e-02, 1.2936e-02, 1.0344e-02],
[ 1.9010e-02, 1.7459e-02, 1.5988e-02]],
[[-1.4914e-03, -8.1727e-03, -8.0671e-03],
[-6.6247e-03, -6.2421e-03, -9.2717e-03],
[-8.7991e-03, -7.7528e-03, -8.6336e-03]],
...,
[[-1.8040e-02, -1.5366e-02, -1.5334e-02],
[-1.3148e-02, -1.2180e-02, -1.0915e-02],
[-1.4545e-02, -1.4756e-02, -1.1787e-02]],
[[ 3.5762e-03, 6.6073e-03, -1.4055e-03],
[ 4.3975e-03, 7.8375e-03, 8.8085e-05],
[-5.0697e-03, -5.6633e-04, -5.9284e-03]],
[[-1.9234e-03, -8.8012e-03, -5.8821e-03],
[ 3.6685e-03, -1.3784e-03, -3.2117e-03],
[-4.7037e-04, 1.5340e-04, -3.4046e-03]]],
[[[-1.8305e-02, -1.7735e-02, -2.1683e-02],
[-1.6598e-02, -1.2508e-02, -2.0530e-02],
[-1.0800e-02, -9.8670e-03, -1.7195e-02]],
[[ 2.0721e-02, 2.2466e-02, 2.5049e-02],
[ 1.8682e-02, 1.3160e-02, 2.3696e-02],
[ 2.2104e-02, 1.7261e-02, 2.4877e-02]],
[[-5.7091e-03, -2.6876e-03, -9.2260e-04],
[-9.4530e-03, -7.0543e-03, -6.2770e-03],
[-4.5806e-03, -2.7182e-03, -2.5823e-03]],
...,
[[ 2.4150e-02, 1.4002e-02, 1.6559e-02],
[ 2.1363e-02, 1.4359e-02, 1.5854e-02],
[ 2.5786e-02, 2.7233e-02, 2.5104e-02]],
[[-4.6450e-03, 1.2419e-03, -1.8768e-03],
[ 1.3005e-03, 4.0888e-03, -6.5483e-04],
[-7.9783e-03, -6.6539e-03, -8.9957e-03]],
[[ 1.1494e-02, 2.6621e-02, 1.5649e-02],
[ 6.5960e-03, 1.7290e-02, 7.5466e-03],
[-8.0256e-03, 4.6246e-03, -5.7808e-03]]],
[[[ 1.4232e-02, 1.1769e-02, 9.4342e-03],
[ 6.2592e-03, 5.1087e-03, 2.3311e-03],
[-1.9694e-03, 2.7110e-03, -2.8945e-03]],
[[-7.0772e-03, 1.0365e-03, -5.8451e-03],
[-9.1879e-03, -3.1388e-03, -8.1517e-03],
[-8.0300e-03, -5.1313e-03, -9.5734e-03]],
[[ 2.4314e-02, 1.8942e-02, 2.4256e-02],
[ 2.0090e-02, 1.1472e-02, 1.5993e-02],
[ 2.2910e-02, 2.0622e-02, 2.3820e-02]],
...,
[[-1.6375e-02, -1.6928e-02, -1.9019e-02],
[-9.7367e-03, -1.1274e-02, -1.0261e-02],
[-1.2310e-02, -1.5931e-02, -1.4151e-02]],
[[ 4.7098e-03, -4.5205e-04, 2.8042e-03],
[ 2.1428e-03, -4.6175e-03, -1.6818e-03],
[-1.3336e-03, -5.5009e-03, -2.6237e-03]],
[[-1.4367e-02, -1.3520e-02, -1.1387e-02],
[-4.7420e-03, -1.7309e-03, -2.6426e-03],
[ 5.1448e-03, 7.0428e-03, 5.0202e-03]]]], device='cuda:0'),
Parameter containing:
tensor([1.8419, 1.8307, 1.7650, 1.8288, 1.9505, 1.8026, 1.9536, 2.2790, 1.7662,
1.8902, 1.7768, 1.7749, 1.9055, 1.7328, 1.8762, 1.8211, 1.7967, 2.3428,
1.7985, 1.7271, 1.7915, 1.9512, 1.8928, 1.9017, 1.8784, 1.9809, 1.8569,
1.7830, 1.8911, 1.8859, 1.7764, 1.9832, 1.8389, 1.7616, 1.8728, 1.8753,
1.9008, 1.8209, 1.7039, 1.7377, 1.7786, 1.6944, 1.7829, 1.7815, 1.7594,
1.8428, 1.9238, 2.0871, 1.8980, 1.8413, 1.8471, 1.8584, 1.7640, 1.8453,
1.7606, 1.9504, 1.9620, 1.8755, 1.9424, 1.8731, 1.8674, 1.9422, 1.8750,
1.9208, 1.7464, 1.8558, 1.6539, 2.0660, 2.0298, 1.9174, 1.8972, 1.7589,
1.7551, 1.9560, 1.7909, 1.7971, 1.7851, 1.7733, 1.8061, 1.7949, 1.8169,
1.8089, 1.8641, 2.1542, 1.7739, 1.7913, 1.8022, 1.7155, 1.7679, 1.7704,
1.6266, 1.8645, 1.9076, 1.8576, 1.6924, 1.8020, 1.7100, 1.7713, 1.8572,
1.7103, 2.0664, 1.9054, 1.9422, 1.8078, 1.7412, 1.6061, 1.9105, 1.8947,
1.7954, 1.8989, 1.8239, 1.7619, 1.7951, 1.8149, 1.8539, 1.8502, 1.7095,
2.1831, 1.8599, 1.8252, 1.8193, 1.8460, 1.7968, 1.6229, 1.8450, 1.8290,
1.8706, 1.9293, 1.6881, 1.9725, 1.8981, 1.8925, 1.8851, 1.8445, 1.9764,
2.0674, 1.8384, 1.8414, 1.8762, 1.7931, 1.7131, 1.9644, 1.7854, 1.9369,
1.8972, 1.8940, 1.8700, 1.7967, 1.8775, 1.9409, 1.7391, 1.7944, 1.9678,
1.7678, 1.6851, 1.9414, 1.9663, 1.9882, 1.7915, 1.8141, 1.8325, 2.1200,
1.9256, 2.3592, 2.0304, 1.9594, 1.7334, 1.9048, 1.8221, 1.7811, 1.9084,
1.8053, 1.9171, 1.9644, 1.8256, 1.6432, 1.9173, 1.9094, 1.9923, 1.7963,
1.9077, 1.7619, 2.1724, 1.7931, 1.7564, 1.8889, 1.9832, 1.9136, 1.8035,
1.8419, 1.8278, 1.8057, 1.9063, 1.8646, 1.7848, 1.8230, 1.7986, 1.7091,
1.7724, 1.7939, 1.7611, 1.9325, 2.0162, 1.7295, 2.0196, 1.8876, 1.8325,
1.8225, 1.7870, 1.9160, 1.7197, 1.7170, 1.9133, 1.7770, 1.9943, 1.8389,
1.8070, 1.8516, 1.7857, 1.9648, 1.9553, 1.9232, 1.8086, 1.8114, 1.7141,
1.8058, 1.8532, 1.9255, 1.7682, 1.8314, 1.8495, 1.8296, 1.8278, 1.8819,
1.7698, 1.7838, 1.7807, 1.9974, 1.6994, 1.9483, 1.7793, 1.8029, 2.2210,
1.6455, 1.8357, 2.1706, 1.9204, 1.7414, 1.7809, 1.8648, 1.9145, 1.8849,
1.8346, 1.9368, 1.8169, 2.2302, 1.8262, 2.0651, 1.9888, 1.8169, 1.8462,
1.9681, 1.8083, 1.8595, 1.8539, 1.7699, 1.9001, 1.7285, 1.7553, 1.8924,
1.7829, 1.9428, 1.8724, 1.7228, 2.0548, 1.7732, 1.8561, 1.7699, 1.9269,
1.8171, 2.4075, 1.7257, 1.7819, 1.7244, 1.8521, 1.8302, 1.8797, 1.7617,
1.9650, 1.9807, 1.7102, 1.7486, 1.8350, 1.9919, 1.8505, 1.9000, 1.8269,
1.9787, 1.7635, 1.6071, 1.7998, 1.9545, 1.7348, 1.7140, 1.8851, 1.7981,
1.9100, 1.8315, 1.7864, 1.9165, 1.8839, 1.9017, 1.9334, 1.7405, 1.7661,
1.8015, 1.9987, 1.7622, 1.9107, 1.8444, 1.7128, 1.8726, 1.8529, 1.9270,
1.8769, 1.7261, 1.8393, 1.9075, 1.7953, 1.8246, 1.7605, 2.0470, 1.9221,
1.9205, 1.8910, 1.7666, 1.6801, 1.8308, 1.8845, 1.8339, 1.8238, 1.7616,
1.6114, 1.8411, 1.7437, 1.8423, 1.9540, 1.7465, 1.7741, 1.8746, 1.8856,
1.7740, 1.7603, 1.7682, 1.8396, 1.6869, 1.8080, 1.8836, 1.8283, 1.8341,
1.8522, 1.9749, 1.8707, 1.7719, 1.8993, 1.8108, 1.8480, 1.8267, 1.8731,
1.9576, 1.8347, 1.9509, 1.9641, 1.7997, 1.7652, 1.9253, 1.7126, 1.7551,
1.9427, 1.8559, 1.9163, 1.7681, 1.7803, 1.8500, 1.8535, 1.8865, 1.7599,
2.0692, 1.8021, 1.7077, 1.8890, 1.9457, 1.8516, 1.7882, 1.8356, 1.8472,
1.6708, 1.7435, 1.9080, 1.9653, 2.0401, 1.8935, 1.8450, 1.7536, 1.7733,
1.8135, 1.8534, 1.9368, 1.7348, 1.8738, 1.9632, 1.9033, 1.7422, 1.7842,
1.8516, 2.0218, 1.7044, 1.8793, 1.8655, 1.8516, 1.8002, 1.8687, 1.8460,
1.7589, 1.8174, 1.9830, 1.9034, 2.1222, 1.8460, 1.9209, 1.8893, 1.9422,
1.8489, 1.8396, 1.9953, 2.0865, 1.8253, 1.7700, 1.8035, 1.7535, 1.8923,
1.8620, 1.8627, 1.7264, 1.8140, 1.9613, 1.8812, 1.8729, 2.0050, 1.7092,
1.7726, 1.9410, 1.8381, 1.8366, 1.7276, 1.8796, 1.7548, 1.9536, 1.8062,
1.8883, 2.0278, 1.8775, 1.9446, 1.8676, 1.8423, 1.7798, 1.9403, 1.8375,
2.0473, 1.9507, 1.8337, 1.8184, 1.7791, 1.8993, 1.8781, 1.8691, 1.8493,
1.7623, 1.9458, 1.7564, 1.7448, 1.8633, 1.6863, 1.8062, 1.8702, 2.0048,
1.8504, 1.8964, 1.9489, 1.8264, 1.9019, 1.8196, 1.9712, 1.8969, 1.8652,
1.8709, 1.6984, 1.8677, 1.8846, 1.9256, 1.8620, 1.6366, 1.8434, 1.7506,
1.8438, 1.5788, 1.9316, 1.9535, 1.7878, 1.7354, 2.0920, 1.9456],
device='cuda:0'),
Parameter containing:
tensor([0.2371, 0.3433, 0.3279, 0.4642, 0.2233, 0.2370, 0.2176, 0.3793, 0.3140,
0.2803, 0.2434, 0.2116, 0.2478, 0.2435, 0.2298, 0.3172, 0.2725, 0.6511,
0.2925, 0.2281, 0.2279, 0.4254, 0.2342, 0.3328, 0.2632, 0.2176, 0.3180,
0.3893, 0.1387, 0.2274, 0.3379, 0.0767, 0.2253, 0.2504, 0.1990, 0.1951,
0.2566, 0.3253, 0.2797, 0.3149, 0.2373, 0.2533, 0.1956, 0.3236, 0.2093,
0.2333, 0.2300, 0.5019, 0.2830, 0.1885, 0.3264, 0.2722, 0.2369, 0.2430,
0.3625, 0.2165, 0.4700, 0.3047, 0.3675, 0.2641, 0.1979, 0.2664, 0.3448,
0.2005, 0.2450, 0.4351, 0.2689, 0.1632, 0.3087, 0.1209, 0.2153, 0.1592,
0.2960, 0.1423, 0.2951, 0.2706, 0.2007, 0.2939, 0.2210, 0.2243, 0.2465,
0.3910, 0.4599, 0.5417, 0.2147, 0.3469, 0.2703, 0.2229, 0.3645, 0.2647,
0.2421, 0.2492, 0.1666, 0.2763, 0.2560, 0.2151, 0.3363, 0.2767, 0.2516,
0.2988, 0.2622, 0.3499, 0.3001, 0.3907, 0.3184, 0.2233, 0.2649, 0.2110,
0.2034, 0.2752, 0.2314, 0.3480, 0.2238, 0.2892, 0.1991, 0.2923, 0.3259,
0.0722, 0.3039, 0.3041, 0.3803, 0.2568, 0.2382, 0.3057, 0.2652, 0.1532,
0.2110, 0.2567, 0.3148, 0.2746, 0.1833, 0.1950, 0.1116, 0.2279, 0.3705,
0.2477, 0.2000, 0.3060, 0.2548, 0.2468, 0.3028, 0.1921, 0.2952, 0.1980,
0.2135, 0.1583, 0.1586, 0.3944, 0.2352, 0.3947, 0.2740, 0.2861, 0.1856,
0.2702, 0.2986, 0.1728, 0.2658, 0.2696, 0.2028, 0.1838, 0.3176, 0.6246,
0.2631, 0.3855, 0.2074, 0.2317, 0.4171, 0.2044, 0.2926, 0.3506, 0.2305,
0.2400, 0.1420, 0.1093, 0.2757, 0.3253, 0.2334, 0.1650, 0.4026, 0.2066,
0.1790, 0.3032, 0.5658, 0.3246, 0.3834, 0.3254, 0.1772, 0.2909, 0.2350,
0.2519, 0.1968, 0.2003, 0.3213, 0.4802, 0.2543, 0.2578, 0.3280, 0.2270,
0.3044, 0.2273, 0.2447, 0.2527, 0.4136, 0.2588, 0.3589, 0.2688, 0.2115,
0.2022, 0.3186, 0.3740, 0.1785, 0.2074, 0.2346, 0.3566, 0.2623, 0.2620,
0.2880, 0.1462, 0.1896, 0.2777, 0.1852, 0.3240, 0.2748, 0.2164, 0.3066,
0.1845, 0.3992, 0.1695, 0.4411, 0.2812, 0.2730, 0.2784, 0.1861, 0.3589,
0.1934, 0.3320, 0.3350, 0.2655, 0.2740, 0.3185, 0.2633, 0.2458, 0.2003,
0.2809, 0.3049, 0.2050, 0.2904, 0.2381, 0.3278, 0.3484, 0.4293, 0.2422,
0.2859, 0.1864, 0.2954, 0.5634, 0.2081, 0.3743, 0.2902, 0.3820, 0.3069,
0.2101, 0.2750, 0.2878, 0.1870, 0.3015, 0.1661, 0.2998, 0.3101, 0.2522,
0.2419, 0.1758, 0.2681, 0.2812, 0.1495, 0.2868, 0.3157, 0.2587, 0.2437,
0.1467, 0.5416, 0.2490, 0.2831, 0.2783, 0.1614, 0.1963, 0.2034, 0.2364,
0.2527, 0.1573, 0.3184, 0.2841, 0.1613, 0.1489, 0.2850, 0.1625, 0.3277,
0.4936, 0.2780, 0.3178, 0.1743, 0.2158, 0.2222, 0.2821, 0.4267, 0.2713,
0.1778, 0.3067, 0.2270, 0.1772, 0.3897, 0.2923, 0.4843, 0.2345, 0.2327,
0.2740, 0.2700, 0.2804, 0.4035, 0.1501, 0.3329, 0.3286, 0.2803, 0.2309,
0.1738, 0.3270, 0.3097, 0.1808, 0.2384, 0.2107, 0.3240, 0.3346, 0.2236,
0.2061, 0.2687, 0.2360, 0.3338, 0.2694, 0.3203, 0.2895, 0.1884, 0.1491,
0.3957, 0.5167, 0.3407, 0.1854, 0.1816, 0.2626, 0.1855, 0.2219, 0.1482,
0.2584, 0.2458, 0.2616, 0.2396, 0.2402, 0.2423, 0.3463, 0.2731, 0.1524,
0.2514, 0.2760, 0.1734, 0.2715, 0.4052, 0.2252, 0.3676, 0.3070, 0.3127,
0.1836, 0.4330, 0.2203, 0.2073, 0.2803, 0.2984, 0.2191, 0.3272, 0.2267,
0.2749, 0.3056, 0.4566, 0.2962, 0.3528, 0.3236, 0.4220, 0.2715, 0.2256,
0.2903, 0.1829, 0.3994, 0.2820, 0.2471, 0.1647, 0.3654, 0.4504, 0.2685,
0.2992, 0.2825, 0.2435, 0.2212, 0.4300, 0.4342, 0.1988, 0.2863, 0.3398,
0.2444, 0.2905, 0.2559, 0.2586, 0.1702, 0.1906, 0.2536, 0.2978, 0.2498,
0.3777, 0.2252, 0.2472, 0.2243, 0.1732, 0.2194, 0.2091, 0.2820, 0.2898,
0.2887, 0.3292, 0.1644, 0.2962, 0.3279, 0.2535, 0.2795, 0.2238, 0.2607,
0.1937, 0.2680, 0.2418, 0.5193, 0.2502, 0.3147, 0.2166, 0.2313, 0.2027,
0.1880, 0.2180, 0.3826, 0.3871, 0.2358, 0.3556, 0.2272, 0.3272, 0.3442,
0.3154, 0.1993, 0.3135, 0.2254, 0.3048, 0.2658, 0.3337, 0.2679, 0.2670,
0.2363, 0.4347, 0.1931, 0.1995, 0.2072, 0.3202, 0.2667, 0.2305, 0.2383,
0.2246, 0.2562, 0.2837, 0.4046, 0.2786, 0.2243, 0.1591, 0.1923, 0.1894,
0.2496, 0.1140, 0.3128, 0.3197, 0.3530, 0.2999, 0.2115, 0.4718, 0.2979,
0.3472, 0.2890, 0.4740, 0.2230, 0.3630, 0.4015, 0.2446, 0.1897, 0.1460,
0.1874, 0.2734, 0.2366, 0.3001, 0.2359, 0.2688, 0.3256, 0.2749, 0.2848,
0.2299, 0.3001, 0.4818, 0.3074, 0.3164, 0.3114, 0.3549, 0.2859],
device='cuda:0'),
Parameter containing:
tensor([[ 0.0345, -0.3183, -0.4702, ..., 0.0426, -0.2295, 0.1899],
[-0.0395, -0.0307, 0.0222, ..., -0.0358, 0.0175, -0.0546],
[ 0.0213, 0.0333, 0.0223, ..., -0.0761, 0.0598, -0.0121],
...,
[ 0.0981, -0.0773, 0.0473, ..., 0.0361, -0.0102, -0.0076],
[-0.0779, -0.0764, -0.1302, ..., 0.0722, -0.0760, -0.0214],
[-0.0006, -0.0591, -0.1613, ..., -0.0418, 0.0960, -0.0646]],
device='cuda:0', requires_grad=True),
Parameter containing:
tensor([ 2.3882e-02, 2.3160e-02, 1.0865e-02, -9.5149e-02, 2.0054e-02,
-2.8129e-02, -8.1836e-02, -4.5059e-02, -3.3202e-02, -5.0288e-02,
1.5722e-02, -8.1809e-02, 3.8963e-02, 2.1192e-02, 1.5252e-02,
1.8656e-02, 2.2183e-02, -8.6538e-02, -1.8721e-02, -2.0963e-02,
2.8428e-02, -9.9596e-03, -1.4719e-02, -3.3502e-02, -3.6034e-02,
2.7984e-02, -4.0156e-02, 1.6873e-03, -7.8723e-03, 1.3919e-01,
-4.9428e-02, 1.0176e-02, -3.0426e-02, 6.4723e-02, -8.4473e-02,
-7.7931e-02, -6.8479e-02, 6.8622e-02, 2.2563e-02, -1.2761e-02,
-6.6919e-02, -5.6423e-02, -1.2707e-01, 1.6811e-02, -1.9661e-01,
-2.9402e-02, -2.7494e-02, -7.1889e-02, 5.0625e-02, -1.1795e-01,
-1.7137e-01, -3.9132e-02, -6.9623e-02, 7.1832e-03, -8.5855e-04,
-8.9473e-03, -3.7105e-02, 2.7011e-02, 1.4709e-01, -8.0529e-02,
1.3977e-02, -2.4924e-02, 1.9562e-02, 1.9913e-02, -1.8919e-01,
-1.5520e-02, -8.8260e-02, -1.1059e-01, -1.5079e-02, -5.3378e-02,
-8.9203e-02, -1.0469e-01, -3.4431e-04, 1.8321e-02, 4.4223e-02,
-2.3680e-02, 2.5024e-01, -7.5267e-02, 6.9595e-03, 5.8727e-02,
-7.6249e-02, 1.2731e-01, 4.1504e-02, -1.2871e-01, -4.4928e-02,
-1.4547e-01, -5.6118e-02, 2.2972e-03, -3.9553e-02, -2.9428e-02,
-6.9022e-03, -1.9316e-02, -2.1254e-04, 2.4385e-03, -8.6958e-02,
3.2247e-02, -1.1623e-01, -2.3061e-02, -4.3962e-02, 1.4804e-01,
-1.7960e-02, -2.6240e-02, -5.1388e-02, -6.4724e-02, 2.3701e-02,
-2.6375e-02, -6.1721e-03, -3.4196e-02, -5.2473e-02, -6.8840e-02,
-1.2092e-01, -1.3121e-02, -6.2931e-02, 9.5268e-03, -9.0820e-02,
-1.1162e-01, -6.0939e-02, -6.0011e-02, -9.7252e-02, -1.1699e-01,
1.2594e-02, -4.5866e-02, -6.7777e-02, 6.3572e-02, -1.0414e-02,
1.5994e-03, -2.6629e-02, 5.6616e-03], device='cuda:0',
requires_grad=True),
Parameter containing:
tensor([[-4.8745e-01, -1.2102e-02, 6.5422e-02, 4.0561e-02, -5.8295e-02,
5.0964e-02, 6.9002e-02, -9.1042e-02, -9.9724e-02, -3.6839e-02,
-6.9494e-02, 2.8650e-02, 5.6055e-02, -1.6195e-01, -1.1176e-01,
-6.5000e-02, -6.6726e-02, -1.0028e-01, -6.0856e-02, 7.6315e-02,
8.8964e-02, 1.6451e-03, 2.2384e-02, -9.3930e-02, -5.4525e-02,
5.2312e-02, 2.3416e-01, 4.6622e-02, -1.3078e-02, -2.0146e-01,
-5.9992e-02, 5.0101e-02, -5.5169e-02, 7.1130e-02, -8.4937e-02,
-1.5820e-01, -6.8509e-03, 2.9893e-02, -6.8028e-02, -7.0754e-02,
2.8120e-02, 1.1718e-01, 7.4299e-02, 1.8101e-02, -2.7412e-01,
-8.6546e-02, 1.2974e-01, -7.0481e-02, -3.1670e-02, 3.4876e-02,
5.5180e-02, 9.1050e-03, -1.4184e-02, -5.8973e-02, 3.0867e-02,
-1.2077e-01, 7.1129e-02, -2.6028e-02, 1.9229e-01, 7.9713e-02,
6.0510e-02, 4.6453e-03, -2.9189e-02, 2.6882e-02, -4.6318e-03,
2.0667e-02, -4.7193e-02, 1.3151e-01, -2.0508e-01, -4.9152e-02,
5.9015e-02, -7.2685e-02, 7.5160e-02, -5.9431e-02, 4.2996e-02,
-8.8975e-02, -3.3310e-01, -1.0241e-01, -5.9923e-02, -1.8942e-01,
-6.2474e-03, -5.2615e-02, -7.6916e-02, 1.1194e-02, 7.9096e-02,
-3.5705e-02, -8.1465e-02, -8.6878e-02, 1.2709e-03, -5.3036e-02,
-1.1674e-02, -9.6784e-02, -6.1610e-02, 1.5168e-02, -1.6326e-01,
-3.6787e-02, -1.1708e-01, -9.7155e-02, 1.0132e-01, 1.9494e-01,
-1.7423e-02, 8.5618e-02, -1.2843e-02, 6.1894e-02, 8.5567e-03,
-6.4105e-02, -1.4142e-01, -2.5577e-01, -8.5833e-03, -5.3060e-03,
4.4549e-02, 1.1195e-01, -5.7068e-02, -1.5534e-02, 7.4066e-02,
9.0180e-02, 2.7298e-02, -1.0081e-01, 7.2939e-02, -4.0357e-02,
4.9829e-02, -8.6681e-03, -2.9728e-02, -1.4713e-01, -2.5577e-01,
7.1003e-02, -6.5052e-02, 1.4503e-02],
[-1.8943e-01, -9.5311e-02, -7.1541e-02, -1.4473e-01, -6.6290e-02,
-6.2950e-02, -1.0339e-01, -6.1152e-02, -2.8316e-02, -5.3656e-03,
-6.2524e-02, -2.6620e-02, 1.3207e-01, 1.2545e-01, -3.2633e-02,
-2.8837e-02, 1.1900e-01, -4.3299e-02, 7.1765e-02, 9.2635e-03,
-2.4723e-01, 9.9162e-02, -1.3985e-02, -1.7620e-02, -6.8765e-02,
9.7512e-02, 7.5705e-02, 7.3257e-03, 7.8437e-02, 3.0069e-01,
-4.5243e-03, -1.8803e-02, -1.0776e-01, 4.4883e-02, -1.1846e-01,
7.9874e-02, -2.5009e-01, 1.2238e-01, 2.4933e-02, 7.4091e-02,
1.2272e-01, -2.2968e-02, -2.2241e-01, 6.0401e-02, -1.9668e-01,
1.5688e-02, -1.3491e-01, -6.7967e-02, 5.5645e-02, -7.3564e-02,
-1.9097e-01, -1.6189e-02, -7.1664e-02, -4.2120e-02, -6.8107e-02,
9.6100e-02, 4.8505e-02, -3.6589e-02, -2.5408e-01, 2.8137e-02,
4.3260e-03, 4.1233e-02, 1.2848e-01, -1.2362e-01, -8.1960e-02,
-1.0072e-02, 3.7041e-02, -7.4255e-02, -1.8413e-01, 1.9235e-01,
-8.6005e-02, -2.5089e-02, 1.4303e-02, 6.4616e-02, -7.7839e-02,
1.1373e-01, 1.7529e-01, -9.5082e-02, -7.1428e-02, 1.4543e-01,
5.3760e-02, -6.3516e-02, -1.1800e-02, -1.8895e-01, -5.9803e-02,
-1.0762e-01, -8.4542e-02, 1.1888e-01, -1.6030e-03, 2.5119e-02,
-4.7351e-02, 5.0538e-02, -1.5219e-01, 2.0257e-02, 2.1084e-02,
7.5876e-03, -1.3822e-01, -1.8634e-01, -1.8087e-01, -3.9941e-02,
-1.1258e-02, 6.8517e-04, 1.3429e-02, -1.8672e-01, -5.2764e-02,
-3.4422e-02, 4.8491e-02, -7.2013e-02, 2.1781e-02, -1.1914e-01,
-1.8172e-01, 2.7533e-02, -1.2778e-01, -1.8883e-02, -1.2676e-01,
-1.4823e-01, -8.9103e-02, 1.9874e-01, 1.6138e-01, -7.0796e-02,
-9.4076e-02, 5.6416e-02, -7.7269e-02, -4.4464e-02, 4.0271e-02,
1.4718e-01, 9.6151e-02, 9.1682e-02],
[ 1.2582e-01, -2.7707e-02, -6.8097e-02, 1.6570e-01, -3.8364e-02,
3.3157e-02, 6.0634e-02, 1.6884e-01, 7.4974e-02, 8.1004e-02,
-3.4523e-02, -1.9934e-02, 1.2322e-01, -4.1651e-02, -7.0397e-02,
-3.2501e-03, 9.0664e-02, 1.0403e-02, 3.6285e-02, 6.8427e-02,
-8.7453e-02, -2.6708e-02, -4.2627e-02, 1.1437e-02, -3.0080e-02,
-3.3618e-02, -3.6782e-02, 1.7268e-02, -5.1166e-02, 1.9659e-02,
1.3958e-01, -9.4712e-02, -2.8425e-02, -2.9876e-02, 1.0994e-01,
6.7381e-02, -1.0591e-02, -2.5223e-03, 6.6493e-02, -4.4545e-02,
4.2149e-02, 3.7589e-02, 9.0897e-02, -4.2674e-02, 2.0207e-01,
5.5852e-02, -5.7460e-02, 4.1154e-02, 1.2223e-01, 8.5875e-02,
-6.4510e-02, 1.4996e-02, 2.8891e-02, 3.5435e-02, 5.1713e-02,
6.8203e-02, 4.7446e-02, -5.4786e-02, 3.2103e-02, -1.4055e-02,
-1.9741e-03, -4.5006e-02, -9.7270e-03, -8.0248e-02, 1.6742e-01,
-3.3621e-02, -4.2772e-02, -2.5777e-02, -3.6817e-02, -9.6395e-02,
-2.9825e-04, -3.9866e-02, -1.5693e-02, -1.1500e-02, -4.3258e-02,
9.4713e-03, -1.6954e-02, 7.0691e-02, 3.4435e-03, 1.7479e-02,
4.2611e-02, -1.8800e-02, 3.6177e-02, -2.6217e-01, -7.5517e-02,
-3.7183e-02, 2.2031e-02, -1.1382e-02, 3.5616e-02, 6.5777e-02,
-8.4540e-02, -2.3781e-02, -1.7181e-03, 4.6190e-03, 5.5758e-02,
5.6313e-02, 7.4771e-02, -1.8167e-02, -6.2846e-02, -6.5132e-02,
-8.8609e-02, -6.5764e-02, -7.1349e-02, 6.9642e-02, -3.1135e-02,
4.2801e-02, -1.3643e-02, 2.1819e-02, 9.1523e-03, 1.6578e-01,
8.5112e-02, 6.0084e-02, -3.9364e-03, 1.5565e-02, -4.0231e-02,
5.3259e-02, 6.5105e-04, -1.9842e-02, -4.7308e-03, -1.4229e-03,
2.2462e-02, -2.7025e-02, 1.0798e-01, -5.5201e-02, 5.9115e-02,
4.2280e-02, -7.6240e-02, -4.8316e-02],
[ 1.0445e-01, 2.0288e-02, 3.9860e-02, -1.5768e-01, -1.0205e-01,
-3.5041e-02, -1.5510e-01, 3.4439e-02, -6.7905e-03, 7.6577e-02,
-3.3769e-02, 6.3343e-03, 7.6656e-02, -3.6402e-02, -4.3498e-02,
1.9293e-02, 9.3600e-03, 7.4897e-02, 5.4095e-02, -7.1164e-02,
-3.7666e-01, 3.5166e-02, -4.0090e-02, 1.4658e-01, -9.7445e-02,
5.9873e-02, -1.5538e-01, -8.8138e-02, -5.3674e-02, -2.3402e-01,
-2.6658e-02, 4.9032e-02, -6.8837e-02, -2.4822e-02, -7.0911e-02,
-2.2661e-02, -5.8355e-02, 1.1377e-01, 4.9464e-02, 5.2420e-02,
-7.1146e-03, 4.5222e-02, -3.0543e-02, -2.4075e-02, -2.9880e-02,
2.1736e-02, 1.4291e-02, 1.1837e-02, 4.3869e-02, -6.8389e-02,
-3.0216e-01, -8.9730e-03, -5.7487e-02, -2.9430e-02, 6.6473e-02,
9.5135e-02, 7.2007e-02, -6.2084e-03, -4.4718e-03, -4.4319e-02,
1.8050e-02, -5.4630e-02, -1.0518e-02, -1.0007e-01, 6.4115e-03,
5.4974e-02, 5.8158e-02, -6.4369e-02, -9.4605e-02, 5.6295e-02,
-1.2679e-01, -2.3375e-02, 1.5921e-02, 2.9418e-02, 8.5743e-02,
-3.0346e-02, 8.4903e-02, 9.0598e-02, 5.6882e-02, 6.8326e-02,
-2.7533e-02, 1.4607e-01, 1.2707e-01, -2.3441e-01, -7.8791e-02,
-2.5067e-02, -1.9230e-02, -1.2828e-01, 9.0558e-02, 6.1793e-02,
4.9037e-02, -3.4608e-02, 6.3156e-03, 5.6468e-02, 4.2965e-02,
-9.2287e-02, 2.9720e-02, 6.7632e-02, -1.3350e-01, -9.6728e-02,
-3.8071e-02, -3.4323e-02, 8.4998e-02, -4.3782e-02, 5.0118e-02,
-6.7789e-02, -6.2655e-02, -2.2966e-02, 5.7994e-02, 5.4030e-02,
-8.9043e-03, -7.6267e-02, 2.1659e-02, 5.9630e-02, -7.3024e-02,
-9.2469e-02, -9.9133e-02, 1.9687e-03, -2.5081e-01, -6.0421e-02,
3.1804e-02, 3.6854e-02, 6.9453e-02, -1.6007e-03, -5.1810e-02,
-2.0790e-02, 6.6285e-02, 4.5710e-02],
[-9.3480e-03, 1.0044e-02, 6.2689e-02, -1.1751e-01, 5.7494e-02,
-5.3356e-02, -5.1719e-02, -8.0611e-02, -3.2994e-02, 4.7971e-02,
1.8605e-02, 2.1457e-02, -4.9067e-02, -5.8949e-02, -2.0263e-02,
5.1273e-02, -5.6748e-02, -3.3365e-02, 6.3614e-02, -2.4196e-02,
-4.5514e-02, 4.7095e-02, -2.0277e-02, 6.6178e-02, 6.0306e-02,
8.2595e-02, 5.0167e-02, -1.4945e-02, -2.3142e-03, -1.7626e-01,
3.1366e-02, -2.5953e-02, -6.1366e-02, -7.6500e-02, -6.7769e-02,
9.2949e-02, 2.0062e-01, -1.0792e-01, -1.6528e-02, -1.6701e-02,
-1.4432e-02, -1.5019e-02, 5.8451e-02, -2.3426e-03, -1.4286e-01,
4.3311e-02, 2.9433e-02, 4.3419e-02, -1.9438e-01, 5.9776e-02,
-2.8685e-01, -1.5671e-03, 2.4169e-01, 3.4834e-02, 5.8942e-02,
6.5155e-02, 5.1638e-02, -8.2930e-02, 1.1019e-01, -1.2815e-02,
-1.9315e-02, 4.5446e-02, -1.0844e-01, 6.4890e-02, -1.7658e-02,
2.2320e-02, -8.7311e-02, -1.0393e-01, 2.4817e-01, -1.8710e-02,
-1.3833e-01, -1.4084e-01, -1.2664e-02, 7.0760e-02, -6.6813e-02,
1.3063e-01, 3.4110e-02, 2.4923e-02, -1.0050e-01, 5.3968e-02,
-1.1325e-01, -1.6362e-01, -7.3702e-02, -1.9392e-01, -8.9765e-02,
2.2131e-02, 5.5211e-02, 1.4346e-01, -9.6013e-02, -6.6208e-02,
-3.5739e-02, -3.2729e-02, -1.0804e-02, 1.7304e-03, 4.0798e-02,
-2.6155e-02, 1.0030e-01, 8.9367e-02, -3.2173e-02, -2.3184e-01,
-3.8889e-02, -4.7861e-02, 7.5138e-02, -6.1155e-02, 4.4688e-02,
6.7546e-02, 8.1820e-02, 7.9347e-03, 2.8454e-03, 2.2361e-02,
1.2709e-01, 8.0562e-03, -9.7130e-02, -7.2896e-02, -1.1342e-01,
1.2087e-01, -5.0355e-02, 6.3797e-02, -8.1286e-02, 5.6618e-02,
2.6636e-02, 6.7817e-02, -1.4628e-01, 9.3704e-02, 1.1827e-01,
-3.7516e-02, -5.5686e-02, 4.4176e-02],
[-4.1551e-01, -1.1826e-02, 1.1219e-01, 1.2629e-01, -2.0804e-02,
4.4033e-02, 9.6014e-02, 1.1022e-01, -1.9680e-01, -1.2967e-01,
1.2251e-02, 1.1734e-01, -6.6136e-02, -8.3951e-02, -6.4435e-02,
-1.0817e-01, -3.0742e-02, 1.5271e-02, 4.3513e-02, -8.4594e-02,
6.8212e-02, 2.1760e-02, 3.8373e-02, -1.8607e-01, 1.1946e-01,
8.2204e-02, -5.8330e-02, 4.6245e-03, 1.2751e-02, -2.1609e-01,
-1.1078e-01, -1.2965e-02, -7.1851e-02, 2.1416e-02, -1.6482e-01,
-2.2266e-02, 5.4793e-02, -3.6825e-02, -9.3510e-02, -1.0132e-02,
8.4685e-02, 8.3058e-02, 1.2457e-02, -3.4615e-02, -7.3403e-02,
-6.9102e-02, -1.1287e-01, -8.0894e-02, 3.7375e-02, 8.5972e-02,
1.3127e-01, 6.0794e-02, 5.6340e-02, -6.9808e-02, -6.3643e-02,
2.0344e-02, -4.4906e-02, -3.2832e-02, -1.1888e-01, -3.1430e-02,
5.2615e-03, 4.5307e-03, -5.4874e-02, 2.6163e-02, -3.4820e-03,
9.6005e-02, 8.1418e-02, 1.0566e-01, -4.6627e-02, 2.5222e-02,
6.5191e-02, 1.3285e-01, -5.2298e-02, 1.3938e-02, -8.9627e-02,
-9.3309e-02, -2.4065e-01, 2.8894e-02, -5.8771e-03, -2.6192e-01,
1.8775e-02, 6.4480e-02, -1.5761e-01, 1.7129e-01, 4.4078e-02,
1.6031e-01, -3.2656e-01, -2.3240e-02, 3.1537e-02, -1.2648e-01,
1.1993e-02, 2.3773e-02, -1.4316e-01, 7.1942e-02, -1.5182e-01,
1.5646e-02, -3.4267e-03, -9.6387e-02, 1.0521e-01, -2.8769e-02,
-1.7415e-02, 9.3605e-02, 2.2417e-02, -5.7616e-02, 6.1135e-02,
3.8660e-02, -6.4307e-02, -1.3600e-01, 7.7186e-03, -8.5118e-02,
-1.5163e-02, 5.6989e-02, 4.3245e-03, -2.1169e-02, 4.1125e-02,
-1.3783e-02, 8.6877e-02, 2.0958e-02, 6.1320e-02, 9.8687e-02,
-4.3692e-02, -7.5546e-02, -1.4047e-01, -2.5837e-01, -3.4284e-01,
2.7730e-02, -6.0936e-02, 2.2814e-02]], device='cuda:0',
requires_grad=True),
Parameter containing:
tensor([ 0.0272, -0.2218, -0.0513, 0.1923, 0.0594, -0.1952], device='cuda:0',
requires_grad=True)]
len(list(oModelPreTrn.parameters()))
64
(list(oModelPreTrn.fc.parameters()))
[Parameter containing:
tensor([[ 0.0345, -0.3183, -0.4702, ..., 0.0426, -0.2295, 0.1899],
[-0.0395, -0.0307, 0.0222, ..., -0.0358, 0.0175, -0.0546],
[ 0.0213, 0.0333, 0.0223, ..., -0.0761, 0.0598, -0.0121],
...,
[ 0.0981, -0.0773, 0.0473, ..., 0.0361, -0.0102, -0.0076],
[-0.0779, -0.0764, -0.1302, ..., 0.0722, -0.0760, -0.0214],
[-0.0006, -0.0591, -0.1613, ..., -0.0418, 0.0960, -0.0646]],
device='cuda:0', requires_grad=True),
Parameter containing:
tensor([ 2.3882e-02, 2.3160e-02, 1.0865e-02, -9.5149e-02, 2.0054e-02,
-2.8129e-02, -8.1836e-02, -4.5059e-02, -3.3202e-02, -5.0288e-02,
1.5722e-02, -8.1809e-02, 3.8963e-02, 2.1192e-02, 1.5252e-02,
1.8656e-02, 2.2183e-02, -8.6538e-02, -1.8721e-02, -2.0963e-02,
2.8428e-02, -9.9596e-03, -1.4719e-02, -3.3502e-02, -3.6034e-02,
2.7984e-02, -4.0156e-02, 1.6873e-03, -7.8723e-03, 1.3919e-01,
-4.9428e-02, 1.0176e-02, -3.0426e-02, 6.4723e-02, -8.4473e-02,
-7.7931e-02, -6.8479e-02, 6.8622e-02, 2.2563e-02, -1.2761e-02,
-6.6919e-02, -5.6423e-02, -1.2707e-01, 1.6811e-02, -1.9661e-01,
-2.9402e-02, -2.7494e-02, -7.1889e-02, 5.0625e-02, -1.1795e-01,
-1.7137e-01, -3.9132e-02, -6.9623e-02, 7.1832e-03, -8.5855e-04,
-8.9473e-03, -3.7105e-02, 2.7011e-02, 1.4709e-01, -8.0529e-02,
1.3977e-02, -2.4924e-02, 1.9562e-02, 1.9913e-02, -1.8919e-01,
-1.5520e-02, -8.8260e-02, -1.1059e-01, -1.5079e-02, -5.3378e-02,
-8.9203e-02, -1.0469e-01, -3.4431e-04, 1.8321e-02, 4.4223e-02,
-2.3680e-02, 2.5024e-01, -7.5267e-02, 6.9595e-03, 5.8727e-02,
-7.6249e-02, 1.2731e-01, 4.1504e-02, -1.2871e-01, -4.4928e-02,
-1.4547e-01, -5.6118e-02, 2.2972e-03, -3.9553e-02, -2.9428e-02,
-6.9022e-03, -1.9316e-02, -2.1254e-04, 2.4385e-03, -8.6958e-02,
3.2247e-02, -1.1623e-01, -2.3061e-02, -4.3962e-02, 1.4804e-01,
-1.7960e-02, -2.6240e-02, -5.1388e-02, -6.4724e-02, 2.3701e-02,
-2.6375e-02, -6.1721e-03, -3.4196e-02, -5.2473e-02, -6.8840e-02,
-1.2092e-01, -1.3121e-02, -6.2931e-02, 9.5268e-03, -9.0820e-02,
-1.1162e-01, -6.0939e-02, -6.0011e-02, -9.7252e-02, -1.1699e-01,
1.2594e-02, -4.5866e-02, -6.7777e-02, 6.3572e-02, -1.0414e-02,
1.5994e-03, -2.6629e-02, 5.6616e-03], device='cuda:0',
requires_grad=True),
Parameter containing:
tensor([[-4.8745e-01, -1.2102e-02, 6.5422e-02, 4.0561e-02, -5.8295e-02,
5.0964e-02, 6.9002e-02, -9.1042e-02, -9.9724e-02, -3.6839e-02,
-6.9494e-02, 2.8650e-02, 5.6055e-02, -1.6195e-01, -1.1176e-01,
-6.5000e-02, -6.6726e-02, -1.0028e-01, -6.0856e-02, 7.6315e-02,
8.8964e-02, 1.6451e-03, 2.2384e-02, -9.3930e-02, -5.4525e-02,
5.2312e-02, 2.3416e-01, 4.6622e-02, -1.3078e-02, -2.0146e-01,
-5.9992e-02, 5.0101e-02, -5.5169e-02, 7.1130e-02, -8.4937e-02,
-1.5820e-01, -6.8509e-03, 2.9893e-02, -6.8028e-02, -7.0754e-02,
2.8120e-02, 1.1718e-01, 7.4299e-02, 1.8101e-02, -2.7412e-01,
-8.6546e-02, 1.2974e-01, -7.0481e-02, -3.1670e-02, 3.4876e-02,
5.5180e-02, 9.1050e-03, -1.4184e-02, -5.8973e-02, 3.0867e-02,
-1.2077e-01, 7.1129e-02, -2.6028e-02, 1.9229e-01, 7.9713e-02,
6.0510e-02, 4.6453e-03, -2.9189e-02, 2.6882e-02, -4.6318e-03,
2.0667e-02, -4.7193e-02, 1.3151e-01, -2.0508e-01, -4.9152e-02,
5.9015e-02, -7.2685e-02, 7.5160e-02, -5.9431e-02, 4.2996e-02,
-8.8975e-02, -3.3310e-01, -1.0241e-01, -5.9923e-02, -1.8942e-01,
-6.2474e-03, -5.2615e-02, -7.6916e-02, 1.1194e-02, 7.9096e-02,
-3.5705e-02, -8.1465e-02, -8.6878e-02, 1.2709e-03, -5.3036e-02,
-1.1674e-02, -9.6784e-02, -6.1610e-02, 1.5168e-02, -1.6326e-01,
-3.6787e-02, -1.1708e-01, -9.7155e-02, 1.0132e-01, 1.9494e-01,
-1.7423e-02, 8.5618e-02, -1.2843e-02, 6.1894e-02, 8.5567e-03,
-6.4105e-02, -1.4142e-01, -2.5577e-01, -8.5833e-03, -5.3060e-03,
4.4549e-02, 1.1195e-01, -5.7068e-02, -1.5534e-02, 7.4066e-02,
9.0180e-02, 2.7298e-02, -1.0081e-01, 7.2939e-02, -4.0357e-02,
4.9829e-02, -8.6681e-03, -2.9728e-02, -1.4713e-01, -2.5577e-01,
7.1003e-02, -6.5052e-02, 1.4503e-02],
[-1.8943e-01, -9.5311e-02, -7.1541e-02, -1.4473e-01, -6.6290e-02,
-6.2950e-02, -1.0339e-01, -6.1152e-02, -2.8316e-02, -5.3656e-03,
-6.2524e-02, -2.6620e-02, 1.3207e-01, 1.2545e-01, -3.2633e-02,
-2.8837e-02, 1.1900e-01, -4.3299e-02, 7.1765e-02, 9.2635e-03,
-2.4723e-01, 9.9162e-02, -1.3985e-02, -1.7620e-02, -6.8765e-02,
9.7512e-02, 7.5705e-02, 7.3257e-03, 7.8437e-02, 3.0069e-01,
-4.5243e-03, -1.8803e-02, -1.0776e-01, 4.4883e-02, -1.1846e-01,
7.9874e-02, -2.5009e-01, 1.2238e-01, 2.4933e-02, 7.4091e-02,
1.2272e-01, -2.2968e-02, -2.2241e-01, 6.0401e-02, -1.9668e-01,
1.5688e-02, -1.3491e-01, -6.7967e-02, 5.5645e-02, -7.3564e-02,
-1.9097e-01, -1.6189e-02, -7.1664e-02, -4.2120e-02, -6.8107e-02,
9.6100e-02, 4.8505e-02, -3.6589e-02, -2.5408e-01, 2.8137e-02,
4.3260e-03, 4.1233e-02, 1.2848e-01, -1.2362e-01, -8.1960e-02,
-1.0072e-02, 3.7041e-02, -7.4255e-02, -1.8413e-01, 1.9235e-01,
-8.6005e-02, -2.5089e-02, 1.4303e-02, 6.4616e-02, -7.7839e-02,
1.1373e-01, 1.7529e-01, -9.5082e-02, -7.1428e-02, 1.4543e-01,
5.3760e-02, -6.3516e-02, -1.1800e-02, -1.8895e-01, -5.9803e-02,
-1.0762e-01, -8.4542e-02, 1.1888e-01, -1.6030e-03, 2.5119e-02,
-4.7351e-02, 5.0538e-02, -1.5219e-01, 2.0257e-02, 2.1084e-02,
7.5876e-03, -1.3822e-01, -1.8634e-01, -1.8087e-01, -3.9941e-02,
-1.1258e-02, 6.8517e-04, 1.3429e-02, -1.8672e-01, -5.2764e-02,
-3.4422e-02, 4.8491e-02, -7.2013e-02, 2.1781e-02, -1.1914e-01,
-1.8172e-01, 2.7533e-02, -1.2778e-01, -1.8883e-02, -1.2676e-01,
-1.4823e-01, -8.9103e-02, 1.9874e-01, 1.6138e-01, -7.0796e-02,
-9.4076e-02, 5.6416e-02, -7.7269e-02, -4.4464e-02, 4.0271e-02,
1.4718e-01, 9.6151e-02, 9.1682e-02],
[ 1.2582e-01, -2.7707e-02, -6.8097e-02, 1.6570e-01, -3.8364e-02,
3.3157e-02, 6.0634e-02, 1.6884e-01, 7.4974e-02, 8.1004e-02,
-3.4523e-02, -1.9934e-02, 1.2322e-01, -4.1651e-02, -7.0397e-02,
-3.2501e-03, 9.0664e-02, 1.0403e-02, 3.6285e-02, 6.8427e-02,
-8.7453e-02, -2.6708e-02, -4.2627e-02, 1.1437e-02, -3.0080e-02,
-3.3618e-02, -3.6782e-02, 1.7268e-02, -5.1166e-02, 1.9659e-02,
1.3958e-01, -9.4712e-02, -2.8425e-02, -2.9876e-02, 1.0994e-01,
6.7381e-02, -1.0591e-02, -2.5223e-03, 6.6493e-02, -4.4545e-02,
4.2149e-02, 3.7589e-02, 9.0897e-02, -4.2674e-02, 2.0207e-01,
5.5852e-02, -5.7460e-02, 4.1154e-02, 1.2223e-01, 8.5875e-02,
-6.4510e-02, 1.4996e-02, 2.8891e-02, 3.5435e-02, 5.1713e-02,
6.8203e-02, 4.7446e-02, -5.4786e-02, 3.2103e-02, -1.4055e-02,
-1.9741e-03, -4.5006e-02, -9.7270e-03, -8.0248e-02, 1.6742e-01,
-3.3621e-02, -4.2772e-02, -2.5777e-02, -3.6817e-02, -9.6395e-02,
-2.9825e-04, -3.9866e-02, -1.5693e-02, -1.1500e-02, -4.3258e-02,
9.4713e-03, -1.6954e-02, 7.0691e-02, 3.4435e-03, 1.7479e-02,
4.2611e-02, -1.8800e-02, 3.6177e-02, -2.6217e-01, -7.5517e-02,
-3.7183e-02, 2.2031e-02, -1.1382e-02, 3.5616e-02, 6.5777e-02,
-8.4540e-02, -2.3781e-02, -1.7181e-03, 4.6190e-03, 5.5758e-02,
5.6313e-02, 7.4771e-02, -1.8167e-02, -6.2846e-02, -6.5132e-02,
-8.8609e-02, -6.5764e-02, -7.1349e-02, 6.9642e-02, -3.1135e-02,
4.2801e-02, -1.3643e-02, 2.1819e-02, 9.1523e-03, 1.6578e-01,
8.5112e-02, 6.0084e-02, -3.9364e-03, 1.5565e-02, -4.0231e-02,
5.3259e-02, 6.5105e-04, -1.9842e-02, -4.7308e-03, -1.4229e-03,
2.2462e-02, -2.7025e-02, 1.0798e-01, -5.5201e-02, 5.9115e-02,
4.2280e-02, -7.6240e-02, -4.8316e-02],
[ 1.0445e-01, 2.0288e-02, 3.9860e-02, -1.5768e-01, -1.0205e-01,
-3.5041e-02, -1.5510e-01, 3.4439e-02, -6.7905e-03, 7.6577e-02,
-3.3769e-02, 6.3343e-03, 7.6656e-02, -3.6402e-02, -4.3498e-02,
1.9293e-02, 9.3600e-03, 7.4897e-02, 5.4095e-02, -7.1164e-02,
-3.7666e-01, 3.5166e-02, -4.0090e-02, 1.4658e-01, -9.7445e-02,
5.9873e-02, -1.5538e-01, -8.8138e-02, -5.3674e-02, -2.3402e-01,
-2.6658e-02, 4.9032e-02, -6.8837e-02, -2.4822e-02, -7.0911e-02,
-2.2661e-02, -5.8355e-02, 1.1377e-01, 4.9464e-02, 5.2420e-02,
-7.1146e-03, 4.5222e-02, -3.0543e-02, -2.4075e-02, -2.9880e-02,
2.1736e-02, 1.4291e-02, 1.1837e-02, 4.3869e-02, -6.8389e-02,
-3.0216e-01, -8.9730e-03, -5.7487e-02, -2.9430e-02, 6.6473e-02,
9.5135e-02, 7.2007e-02, -6.2084e-03, -4.4718e-03, -4.4319e-02,
1.8050e-02, -5.4630e-02, -1.0518e-02, -1.0007e-01, 6.4115e-03,
5.4974e-02, 5.8158e-02, -6.4369e-02, -9.4605e-02, 5.6295e-02,
-1.2679e-01, -2.3375e-02, 1.5921e-02, 2.9418e-02, 8.5743e-02,
-3.0346e-02, 8.4903e-02, 9.0598e-02, 5.6882e-02, 6.8326e-02,
-2.7533e-02, 1.4607e-01, 1.2707e-01, -2.3441e-01, -7.8791e-02,
-2.5067e-02, -1.9230e-02, -1.2828e-01, 9.0558e-02, 6.1793e-02,
4.9037e-02, -3.4608e-02, 6.3156e-03, 5.6468e-02, 4.2965e-02,
-9.2287e-02, 2.9720e-02, 6.7632e-02, -1.3350e-01, -9.6728e-02,
-3.8071e-02, -3.4323e-02, 8.4998e-02, -4.3782e-02, 5.0118e-02,
-6.7789e-02, -6.2655e-02, -2.2966e-02, 5.7994e-02, 5.4030e-02,
-8.9043e-03, -7.6267e-02, 2.1659e-02, 5.9630e-02, -7.3024e-02,
-9.2469e-02, -9.9133e-02, 1.9687e-03, -2.5081e-01, -6.0421e-02,
3.1804e-02, 3.6854e-02, 6.9453e-02, -1.6007e-03, -5.1810e-02,
-2.0790e-02, 6.6285e-02, 4.5710e-02],
[-9.3480e-03, 1.0044e-02, 6.2689e-02, -1.1751e-01, 5.7494e-02,
-5.3356e-02, -5.1719e-02, -8.0611e-02, -3.2994e-02, 4.7971e-02,
1.8605e-02, 2.1457e-02, -4.9067e-02, -5.8949e-02, -2.0263e-02,
5.1273e-02, -5.6748e-02, -3.3365e-02, 6.3614e-02, -2.4196e-02,
-4.5514e-02, 4.7095e-02, -2.0277e-02, 6.6178e-02, 6.0306e-02,
8.2595e-02, 5.0167e-02, -1.4945e-02, -2.3142e-03, -1.7626e-01,
3.1366e-02, -2.5953e-02, -6.1366e-02, -7.6500e-02, -6.7769e-02,
9.2949e-02, 2.0062e-01, -1.0792e-01, -1.6528e-02, -1.6701e-02,
-1.4432e-02, -1.5019e-02, 5.8451e-02, -2.3426e-03, -1.4286e-01,
4.3311e-02, 2.9433e-02, 4.3419e-02, -1.9438e-01, 5.9776e-02,
-2.8685e-01, -1.5671e-03, 2.4169e-01, 3.4834e-02, 5.8942e-02,
6.5155e-02, 5.1638e-02, -8.2930e-02, 1.1019e-01, -1.2815e-02,
-1.9315e-02, 4.5446e-02, -1.0844e-01, 6.4890e-02, -1.7658e-02,
2.2320e-02, -8.7311e-02, -1.0393e-01, 2.4817e-01, -1.8710e-02,
-1.3833e-01, -1.4084e-01, -1.2664e-02, 7.0760e-02, -6.6813e-02,
1.3063e-01, 3.4110e-02, 2.4923e-02, -1.0050e-01, 5.3968e-02,
-1.1325e-01, -1.6362e-01, -7.3702e-02, -1.9392e-01, -8.9765e-02,
2.2131e-02, 5.5211e-02, 1.4346e-01, -9.6013e-02, -6.6208e-02,
-3.5739e-02, -3.2729e-02, -1.0804e-02, 1.7304e-03, 4.0798e-02,
-2.6155e-02, 1.0030e-01, 8.9367e-02, -3.2173e-02, -2.3184e-01,
-3.8889e-02, -4.7861e-02, 7.5138e-02, -6.1155e-02, 4.4688e-02,
6.7546e-02, 8.1820e-02, 7.9347e-03, 2.8454e-03, 2.2361e-02,
1.2709e-01, 8.0562e-03, -9.7130e-02, -7.2896e-02, -1.1342e-01,
1.2087e-01, -5.0355e-02, 6.3797e-02, -8.1286e-02, 5.6618e-02,
2.6636e-02, 6.7817e-02, -1.4628e-01, 9.3704e-02, 1.1827e-01,
-3.7516e-02, -5.5686e-02, 4.4176e-02],
[-4.1551e-01, -1.1826e-02, 1.1219e-01, 1.2629e-01, -2.0804e-02,
4.4033e-02, 9.6014e-02, 1.1022e-01, -1.9680e-01, -1.2967e-01,
1.2251e-02, 1.1734e-01, -6.6136e-02, -8.3951e-02, -6.4435e-02,
-1.0817e-01, -3.0742e-02, 1.5271e-02, 4.3513e-02, -8.4594e-02,
6.8212e-02, 2.1760e-02, 3.8373e-02, -1.8607e-01, 1.1946e-01,
8.2204e-02, -5.8330e-02, 4.6245e-03, 1.2751e-02, -2.1609e-01,
-1.1078e-01, -1.2965e-02, -7.1851e-02, 2.1416e-02, -1.6482e-01,
-2.2266e-02, 5.4793e-02, -3.6825e-02, -9.3510e-02, -1.0132e-02,
8.4685e-02, 8.3058e-02, 1.2457e-02, -3.4615e-02, -7.3403e-02,
-6.9102e-02, -1.1287e-01, -8.0894e-02, 3.7375e-02, 8.5972e-02,
1.3127e-01, 6.0794e-02, 5.6340e-02, -6.9808e-02, -6.3643e-02,
2.0344e-02, -4.4906e-02, -3.2832e-02, -1.1888e-01, -3.1430e-02,
5.2615e-03, 4.5307e-03, -5.4874e-02, 2.6163e-02, -3.4820e-03,
9.6005e-02, 8.1418e-02, 1.0566e-01, -4.6627e-02, 2.5222e-02,
6.5191e-02, 1.3285e-01, -5.2298e-02, 1.3938e-02, -8.9627e-02,
-9.3309e-02, -2.4065e-01, 2.8894e-02, -5.8771e-03, -2.6192e-01,
1.8775e-02, 6.4480e-02, -1.5761e-01, 1.7129e-01, 4.4078e-02,
1.6031e-01, -3.2656e-01, -2.3240e-02, 3.1537e-02, -1.2648e-01,
1.1993e-02, 2.3773e-02, -1.4316e-01, 7.1942e-02, -1.5182e-01,
1.5646e-02, -3.4267e-03, -9.6387e-02, 1.0521e-01, -2.8769e-02,
-1.7415e-02, 9.3605e-02, 2.2417e-02, -5.7616e-02, 6.1135e-02,
3.8660e-02, -6.4307e-02, -1.3600e-01, 7.7186e-03, -8.5118e-02,
-1.5163e-02, 5.6989e-02, 4.3245e-03, -2.1169e-02, 4.1125e-02,
-1.3783e-02, 8.6877e-02, 2.0958e-02, 6.1320e-02, 9.8687e-02,
-4.3692e-02, -7.5546e-02, -1.4047e-01, -2.5837e-01, -3.4284e-01,
2.7730e-02, -6.0936e-02, 2.2814e-02]], device='cuda:0',
requires_grad=True),
Parameter containing:
tensor([ 0.0272, -0.2218, -0.0513, 0.1923, 0.0594, -0.1952], device='cuda:0',
requires_grad=True)]
# Training Loop
dModelHist = {}
for ii, (modelName, oModel) in enumerate(lModel):
print(f'Training with the {modelName} model')
oModel = oModel.to(runDevice) #<! Transfer model to device
oOpt = torch.optim.AdamW(oModel.parameters(), lr = 1e-4, betas = (0.9, 0.99), weight_decay = 2e-4) #<! Define optimizer
oSch = torch.optim.lr_scheduler.OneCycleLR(oOpt, max_lr = 2e-2, total_steps = numEpochs * len(dlTrain))
_, lTrainLoss, lTrainScore, lValLoss, lValScore, lLearnRate = TrainModelSch(oModel, dlTrain, dlVal, oOpt, oSch, numEpochs, hL, hS)
dModelHist[modelName] = lTrainLoss, lTrainScore, lValLoss, lValScore, lLearnRate
Training with the Pre Defined Model model
Epoch 1 / 10 | Train Loss: 0.880 | Val Loss: 1.544 | Train Score: 0.662 | Val Score: 0.541 | Epoch Time: 87.48 | <-- Checkpoint! |
Epoch 2 / 10 | Train Loss: 0.798 | Val Loss: 1.403 | Train Score: 0.713 | Val Score: 0.547 | Epoch Time: 62.12 | <-- Checkpoint! |
Epoch 3 / 10 | Train Loss: 0.677 | Val Loss: 2.505 | Train Score: 0.755 | Val Score: 0.453 | Epoch Time: 61.74 |
Epoch 4 / 10 | Train Loss: 0.609 | Val Loss: 0.877 | Train Score: 0.782 | Val Score: 0.671 | Epoch Time: 61.84 | <-- Checkpoint! |
Epoch 5 / 10 | Train Loss: 0.516 | Val Loss: 0.684 | Train Score: 0.817 | Val Score: 0.742 | Epoch Time: 61.69 | <-- Checkpoint! |
Epoch 6 / 10 | Train Loss: 0.442 | Val Loss: 0.510 | Train Score: 0.841 | Val Score: 0.816 | Epoch Time: 61.84 | <-- Checkpoint! |
Epoch 7 / 10 | Train Loss: 0.382 | Val Loss: 0.607 | Train Score: 0.866 | Val Score: 0.793 | Epoch Time: 61.62 |
Epoch 8 / 10 | Train Loss: 0.323 | Val Loss: 0.364 | Train Score: 0.887 | Val Score: 0.872 | Epoch Time: 61.68 | <-- Checkpoint! |
Epoch 9 / 10 | Train Loss: 0.267 | Val Loss: 0.317 | Train Score: 0.904 | Val Score: 0.886 | Epoch Time: 61.69 | <-- Checkpoint! |
Epoch 10 / 10 | Train Loss: 0.221 | Val Loss: 0.300 | Train Score: 0.922 | Val Score: 0.897 | Epoch Time: 61.59 | <-- Checkpoint! |
Training with the Pre Trained Model model
Epoch 1 / 10 | Train Loss: 0.473 | Val Loss: 0.350 | Train Score: 0.851 | Val Score: 0.871 | Epoch Time: 30.29 | <-- Checkpoint! |
Epoch 2 / 10 | Train Loss: 0.344 | Val Loss: 0.497 | Train Score: 0.875 | Val Score: 0.822 | Epoch Time: 30.98 |
Epoch 3 / 10 | Train Loss: 0.332 | Val Loss: 0.322 | Train Score: 0.879 | Val Score: 0.887 | Epoch Time: 30.00 | <-- Checkpoint! |
Epoch 4 / 10 | Train Loss: 0.318 | Val Loss: 0.262 | Train Score: 0.890 | Val Score: 0.909 | Epoch Time: 30.46 | <-- Checkpoint! |
Epoch 5 / 10 | Train Loss: 0.263 | Val Loss: 0.265 | Train Score: 0.905 | Val Score: 0.906 | Epoch Time: 30.13 |
Epoch 6 / 10 | Train Loss: 0.265 | Val Loss: 0.247 | Train Score: 0.907 | Val Score: 0.912 | Epoch Time: 30.37 | <-- Checkpoint! |
Epoch 7 / 10 | Train Loss: 0.238 | Val Loss: 0.255 | Train Score: 0.913 | Val Score: 0.911 | Epoch Time: 29.83 |
Epoch 8 / 10 | Train Loss: 0.207 | Val Loss: 0.237 | Train Score: 0.925 | Val Score: 0.914 | Epoch Time: 30.32 | <-- Checkpoint! |
Epoch 9 / 10 | Train Loss: 0.189 | Val Loss: 0.234 | Train Score: 0.931 | Val Score: 0.914 | Epoch Time: 30.00 |
Epoch 10 / 10 | Train Loss: 0.179 | Val Loss: 0.235 | Train Score: 0.934 | Val Score: 0.918 | Epoch Time: 30.16 | <-- Checkpoint! |
(@) Add TensorBoard based monitoring. You should use the
TBLoggerclass.(?) Compare run time and memory consumption during the training of the models. How can it be utilized?
# Plot Training Phase
hF, vHa = plt.subplots(nrows = 1, ncols = 3, figsize = (18, 5))
vHa = np.ravel(vHa)
for modelKey in dModelHist:
hA = vHa[0]
hA.plot(dModelHist[modelKey][0], lw = 2, label = f'Train {modelKey}')
hA.plot(dModelHist[modelKey][2], lw = 2, label = f'Validation {modelKey}')
hA.set_title('Cross Entropy Loss')
hA.set_xlabel('Epoch')
hA.set_ylabel('Loss')
hA.legend()
hA = vHa[1]
hA.plot(dModelHist[modelKey][1], lw = 2, label = f'Train {modelKey}')
hA.plot(dModelHist[modelKey][3], lw = 2, label = f'Validation {modelKey}')
hA.set_title('Accuracy Score')
hA.set_xlabel('Epoch')
hA.set_ylabel('Score')
hA.legend()
hA = vHa[2]
hA.plot(lLearnRate, lw = 2, label = f'{modelKey}')
hA.set_title('Learn Rate Scheduler')
hA.set_xlabel('Iteration')
hA.set_ylabel('Learn Rate')
hA.legend()
(@) Build the
Testdata loader (You may usedsTest) and exam the models on few samples.(@) Redo the training with a different model.
(?) Look at the
Places365(Places365 v2) data set.
If the base model for transfer learning is trained onPlaces365, what effect will it have on the results?
Think of the type of the task. You may try it withRelease of Places365-CNNs.