43
GitHub - johnolafenwa/TorchFusion: A modern deep learning framework built to acc...
source link: https://github.com/johnolafenwa/TorchFusion
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
README.md
TorchFusion
A modern deep learning framework built to accelerate research and development of AI systems.
Based on PyTorch and fully compatible with pure PyTorch and other pytorch packages, TorchFusion provides a comprehensive extensible training framework with trainers that you can easily use to train, evaluate and run inference with your PyTorch models, A GAN framework that greatly simplifies the process of experimenting with Generative Adversarial Networks Goodfellow et al. , with concrete implementations of a number of GAN algorithms, and a number of high level network layers and utilities to help you be more productive in your work.
The framework is highly extensible, so you can easily create your own custom trainers for specific purposes.
Unique features of our Trainers
-
Highly configurable
Highly detailed summary function that not only provides you details about number of parameters, layers, input and output sizes but also provides the number of Flops(Multiply-Adds) for every Linear and Convolution layer in your network. Now, you can know the exact computational cost of any CNN architecure with just a single function!!!
-
Live metrics and loss visualizations, with option to save them permanently
Support for persisting logs permanently
-
Easy to use callbacks
Note: This is a pre-release version of TorchFusion, the current set of features are just a sneak peek into what is coming! Future releases of TorchFusion will cut across multiple domains of Deep Learning.
An AI Commons project https://commons.specpal.science Developed and Maintained by John Olafenwa and Moses Olafenwa, brothers, creators of ImageAI and Authors of Introduction to Deep Computer Vision
Installation
Install TorchFusion
pip3 install https://github.com/johnolafenwa/TorchFusion/releases/download/0.1.0/torchfusion-0.1.0-py3-none-any.whl
CPU Only
With Python 3.6pip3 install http://download.pytorch.org/whl/cpu/torch-0.4.0-cp36-cp36m-win_amd64.whl torchvisionWith Python 3.5
pip3 install http://download.pytorch.org/whl/cpu/torch-0.4.0-cp35-cp35m-win_amd64.whl torchvision
With CUDA Support
With Python 3.6pip3 install http://download.pytorch.org/whl/cu80/torch-0.4.0-cp36-cp36m-win_amd64.whl torchvisionWith Python 3.5
pip3 install http://download.pytorch.org/whl/cu80/torch-0.4.0-cp35-cp35m-win_amd64.whl
CPU Only
With Python 3.6pip3 install http://download.pytorch.org/whl/cpu/torch-0.4.0-cp36-cp36m-linux_x86_64.whl torchvisionWith Python 3.5
pip3 install http://download.pytorch.org/whl/cpu/torch-0.4.0-cp35-cp35m-linux_x86_64.whl torchvision
With CUDA Support
pip3 install torch torchvision
CPU Only
pip3 install torch torchvision
With CUDA Support: Visit Pytorch.org for instructions on Installing on OSX with cuda support
MNIST in Five Minutes
import torchfusion as tf
from torchvision.datasets.mnist import MNIST
from torch.utils.data import DataLoader
from torchvision.transforms import transforms
import torch.nn as nn
from torch.optim import Adam
import torch.cuda as cuda
#Define a the classifier network
net = nn.Sequential(
tf.Flatten(),
nn.Linear(784, 100),
nn.ReLU(),
nn.Linear(100, 100),
nn.ReLU(),
nn.Linear(100, 100),
nn.ReLU(),
nn.Linear(100, 10)
)
batch_size = 64
#Transformations and data augmentation
transformations = transforms.Compose([
transforms.Resize(28),
transforms.ToTensor(),
transforms.Normalize((0.5,0.5,0.5),(0.5,0.5,0.5))
])
#Load the training and test sets
train_set = MNIST(root="./data",transform=transformations,download=True)
test_set = MNIST(root="./data",train=False,transform=transformations,download=True)
train_loader = DataLoader(train_set,shuffle=True,batch_size=batch_size,num_workers=4)
test_loader = DataLoader(test_set,shuffle=False,batch_size=batch_size,num_workers=4)
#Move to GPU if available
if cuda.is_available():
net.cuda()
#Setup the optimize and a loss function
optimizer = Adam(net.parameters(),lr=0.001)
loss_fn = nn.CrossEntropyLoss()
#Top 1 Train accuracy
train_metrics = tf.Accuracy(topK=1)
#Top 1 and Top 2 test accuracy
test_metrics_top1 = tf.Accuracy(name="Top 1 Acc ",topK=1)
test_metrics_top2 = tf.Accuracy(name="Top 2 Acc ",topK=2)
#Create an instance of the StandardModel
model = tf.StandardModel(net)
def train():
#print a summary of the network
print(model.summary((1,28,28)))
model.train(train_loader, loss_fn, optimizer, [train_metrics], test_loader,
[test_metrics_top1, test_metrics_top2], num_epochs=20,
model_dir="mnist_mlp_saved_models",save_logs="logs.txt")
if __name__ == "__main__":
train()
GAN in Five Minutes
import torchfusion.gan as tfgan
from torchvision.datasets import MNIST
from torchvision.transforms import transforms
from torch.optim import Adam
from torch.utils.data import DataLoader
import torch.nn as nn
import torch.cuda as cuda
#Transformations and data augmentation
train_transformations = transforms.Compose([
transforms.Resize(28),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])
batch_size = 64
# Load the training set
train_set = MNIST(root="./data", train=True, transform=train_transformations, download=True)
train_data = DataLoader(train_set,batch_size=batch_size,shuffle=True,num_workers=4)
#Create an instance of the NormalDistribution
source = tfgan.NormalDistribution(length=len(train_set),size=(100))
source_data = DataLoader(source,batch_size=batch_size,shuffle=True,num_workers=4)
#Create an instance of the Generator and Discriminator
G = tfgan.MLPGenerator(latent_size=100,output_size=(1,28,28))
D = tfgan.MLPDiscriminator(input_size=(1,28,28))
#Move the networks to GPU if available
if cuda.is_available():
G.cuda()
D.cuda()
#Setup the optimizers
g_optim = Adam(G.parameters(),lr=0.0002,betas=(0.5,0.999))
d_optim = Adam(D.parameters(),lr=0.0002,betas=(0.5,0.999))
#Define the loss function
loss_fn = nn.BCELoss()
if __name__ == "__main__":
#Create an instance of the StandardGANModel
trainer = tfgan.StandardGANModel(G,D,gen_loss_fn=loss_fn,disc_loss_fn=loss_fn)
#Train the two models
trainer.train(train_data,source_data,g_optim,d_optim,num_epochs=200,disc_steps=1,save_interval=3000)
ImageNet Inference
import torchfusion as tf
import torch
import torchvision.transforms as transforms
from torch.utils.data import DataLoader
from torchvision.models.squeezenet import squeezenet1_1
from PIL import Image
INFER_FOLDER = r"./images"
MODEL_PATH = r"squeezenet.pth"
transformations = transforms.Compose([
transforms.Resize((224,224)),
transforms.ToTensor(),
transforms.Normalize((0.5,0.5,0.5),(0.5,0.5,0.5))
])
infer_set = tf.ImagesFromPaths([INFER_FOLDER],recursive=False,transformations=transformations)
infer_loader = DataLoader(infer_set,batch_size=10)
net = squeezenet1_1()
model = tf.StandardModel(net)
model.load_model(MODEL_PATH)
def predict_loader(data_loader):
predictions = model.predict(data_loader,apply_softmax=True)
print(len(predictions))
for pred in predictions:
class_index = torch.argmax(pred)
class_name = tf.decode_imagenet(class_index)
confidence = torch.max(pred)
print("Prediction: {} , Accuracy: {} ".format(class_name, confidence))
def predict_image(image_path):
img = Image.open(image_path).convert("RGB")
img = transformations(img)
img = img.unsqueeze(0)
pred = model.predict(img,apply_softmax=True)
class_index = torch.argmax(pred)
class_name = tf.decode_imagenet(class_index)
confidence = torch.max(pred)
print("Prediction: {} , Accuracy: {} ".format(class_name, confidence))
if __name__ == "__main__":
predict_loader(infer_loader)
predict_image(r"sample.jpg")
Tutorials
See more at Examples
Contact Developers
John Olafenwa
Email: [email protected]
Website: https://john.specpal.science
Twitter: @johnolafenwa
Medium : @johnolafenwa
Facebook : olafenwajohn
Moses Olafenwa
Email: [email protected]
Website: https://moses.specpal.science
Twitter: @OlafenwaMoses
Medium : @guymodscientist
Facebook : moses.olafenwa
Summary of Resnet50 generated by TorchFusion Model Summary Name Input Size Output Size Parameters Multiply Adds (Flops) Conv2d_1 [1, 3, 224, 224] [1, 64, 112, 112] 9408 118013952 BatchNorm2d_1 [1, 64, 112, 112] [1, 64, 112, 112] 128 Not Available ReLU_1 [1, 64, 112, 112] [1, 64, 112, 112] 0 Not Available MaxPool2d_1 [1, 64, 112, 112] [1, 64, 56, 56] 0 Not Available Conv2d_2 [1, 64, 56, 56] [1, 64, 56, 56] 4096 12845056 BatchNorm2d_2 [1, 64, 56, 56] [1, 64, 56, 56] 128 Not Available ReLU_2 [1, 64, 56, 56] [1, 64, 56, 56] 0 Not Available Conv2d_3 [1, 64, 56, 56] [1, 64, 56, 56] 36864 115605504 BatchNorm2d_3 [1, 64, 56, 56] [1, 64, 56, 56] 128 Not Available ReLU_3 [1, 64, 56, 56] [1, 64, 56, 56] 0 Not Available Conv2d_4 [1, 64, 56, 56] [1, 256, 56, 56] 16384 51380224 BatchNorm2d_4 [1, 256, 56, 56] [1, 256, 56, 56] 512 Not Available Conv2d_5 [1, 64, 56, 56] [1, 256, 56, 56] 16384 51380224 BatchNorm2d_5 [1, 256, 56, 56] [1, 256, 56, 56] 512 Not Available ReLU_4 [1, 256, 56, 56] [1, 256, 56, 56] 0 Not Available Bottleneck_1 [1, 64, 56, 56] [1, 256, 56, 56] 0 Not Available Conv2d_6 [1, 256, 56, 56] [1, 64, 56, 56] 16384 51380224 BatchNorm2d_6 [1, 64, 56, 56] [1, 64, 56, 56] 128 Not Available ReLU_5 [1, 64, 56, 56] [1, 64, 56, 56] 0 Not Available Conv2d_7 [1, 64, 56, 56] [1, 64, 56, 56] 36864 115605504 BatchNorm2d_7 [1, 64, 56, 56] [1, 64, 56, 56] 128 Not Available ReLU_6 [1, 64, 56, 56] [1, 64, 56, 56] 0 Not Available Conv2d_8 [1, 64, 56, 56] [1, 256, 56, 56] 16384 51380224 BatchNorm2d_8 [1, 256, 56, 56] [1, 256, 56, 56] 512 Not Available ReLU_7 [1, 256, 56, 56] [1, 256, 56, 56] 0 Not Available Bottleneck_2 [1, 256, 56, 56] [1, 256, 56, 56] 0 Not Available Conv2d_9 [1, 256, 56, 56] [1, 64, 56, 56] 16384 51380224 BatchNorm2d_9 [1, 64, 56, 56] [1, 64, 56, 56] 128 Not Available ReLU_8 [1, 64, 56, 56] [1, 64, 56, 56] 0 Not Available Conv2d_10 [1, 64, 56, 56] [1, 64, 56, 56] 36864 115605504 BatchNorm2d_10 [1, 64, 56, 56] [1, 64, 56, 56] 128 Not Available ReLU_9 [1, 64, 56, 56] [1, 64, 56, 56] 0 Not Available Conv2d_11 [1, 64, 56, 56] [1, 256, 56, 56] 16384 51380224 BatchNorm2d_11 [1, 256, 56, 56] [1, 256, 56, 56] 512 Not Available ReLU_10 [1, 256, 56, 56] [1, 256, 56, 56] 0 Not Available Bottleneck_3 [1, 256, 56, 56] [1, 256, 56, 56] 0 Not Available Conv2d_12 [1, 256, 56, 56] [1, 128, 56, 56] 32768 102760448 BatchNorm2d_12 [1, 128, 56, 56] [1, 128, 56, 56] 256 Not Available ReLU_11 [1, 128, 56, 56] [1, 128, 56, 56] 0 Not Available Conv2d_13 [1, 128, 56, 56] [1, 128, 28, 28] 147456 115605504 BatchNorm2d_13 [1, 128, 28, 28] [1, 128, 28, 28] 256 Not Available ReLU_12 [1, 128, 28, 28] [1, 128, 28, 28] 0 Not Available Conv2d_14 [1, 128, 28, 28] [1, 512, 28, 28] 65536 51380224 BatchNorm2d_14 [1, 512, 28, 28] [1, 512, 28, 28] 1024 Not Available Conv2d_15 [1, 256, 56, 56] [1, 512, 28, 28] 131072 102760448 BatchNorm2d_15 [1, 512, 28, 28] [1, 512, 28, 28] 1024 Not Available ReLU_13 [1, 512, 28, 28] [1, 512, 28, 28] 0 Not Available Bottleneck_4 [1, 256, 56, 56] [1, 512, 28, 28] 0 Not Available Conv2d_16 [1, 512, 28, 28] [1, 128, 28, 28] 65536 51380224 BatchNorm2d_16 [1, 128, 28, 28] [1, 128, 28, 28] 256 Not Available ReLU_14 [1, 128, 28, 28] [1, 128, 28, 28] 0 Not Available Conv2d_17 [1, 128, 28, 28] [1, 128, 28, 28] 147456 115605504 BatchNorm2d_17 [1, 128, 28, 28] [1, 128, 28, 28] 256 Not Available ReLU_15 [1, 128, 28, 28] [1, 128, 28, 28] 0 Not Available Conv2d_18 [1, 128, 28, 28] [1, 512, 28, 28] 65536 51380224 BatchNorm2d_18 [1, 512, 28, 28] [1, 512, 28, 28] 1024 Not Available ReLU_16 [1, 512, 28, 28] [1, 512, 28, 28] 0 Not Available Bottleneck_5 [1, 512, 28, 28] [1, 512, 28, 28] 0 Not Available Conv2d_19 [1, 512, 28, 28] [1, 128, 28, 28] 65536 51380224 BatchNorm2d_19 [1, 128, 28, 28] [1, 128, 28, 28] 256 Not Available ReLU_17 [1, 128, 28, 28] [1, 128, 28, 28] 0 Not Available Conv2d_20 [1, 128, 28, 28] [1, 128, 28, 28] 147456 115605504 BatchNorm2d_20 [1, 128, 28, 28] [1, 128, 28, 28] 256 Not Available ReLU_18 [1, 128, 28, 28] [1, 128, 28, 28] 0 Not Available Conv2d_21 [1, 128, 28, 28] [1, 512, 28, 28] 65536 51380224 BatchNorm2d_21 [1, 512, 28, 28] [1, 512, 28, 28] 1024 Not Available ReLU_19 [1, 512, 28, 28] [1, 512, 28, 28] 0 Not Available Bottleneck_6 [1, 512, 28, 28] [1, 512, 28, 28] 0 Not Available Conv2d_22 [1, 512, 28, 28] [1, 128, 28, 28] 65536 51380224 BatchNorm2d_22 [1, 128, 28, 28] [1, 128, 28, 28] 256 Not Available ReLU_20 [1, 128, 28, 28] [1, 128, 28, 28] 0 Not Available Conv2d_23 [1, 128, 28, 28] [1, 128, 28, 28] 147456 115605504 BatchNorm2d_23 [1, 128, 28, 28] [1, 128, 28, 28] 256 Not Available ReLU_21 [1, 128, 28, 28] [1, 128, 28, 28] 0 Not Available Conv2d_24 [1, 128, 28, 28] [1, 512, 28, 28] 65536 51380224 BatchNorm2d_24 [1, 512, 28, 28] [1, 512, 28, 28] 1024 Not Available ReLU_22 [1, 512, 28, 28] [1, 512, 28, 28] 0 Not Available Bottleneck_7 [1, 512, 28, 28] [1, 512, 28, 28] 0 Not Available Conv2d_25 [1, 512, 28, 28] [1, 256, 28, 28] 131072 102760448 BatchNorm2d_25 [1, 256, 28, 28] [1, 256, 28, 28] 512 Not Available ReLU_23 [1, 256, 28, 28] [1, 256, 28, 28] 0 Not Available Conv2d_26 [1, 256, 28, 28] [1, 256, 14, 14] 589824 115605504 BatchNorm2d_26 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_24 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_27 [1, 256, 14, 14] [1, 1024, 14, 14] 262144 51380224 BatchNorm2d_27 [1, 1024, 14, 14] [1, 1024, 14, 14] 2048 Not Available Conv2d_28 [1, 512, 28, 28] [1, 1024, 14, 14] 524288 102760448 BatchNorm2d_28 [1, 1024, 14, 14] [1, 1024, 14, 14] 2048 Not Available ReLU_25 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Bottleneck_8 [1, 512, 28, 28] [1, 1024, 14, 14] 0 Not Available Conv2d_29 [1, 1024, 14, 14] [1, 256, 14, 14] 262144 51380224 BatchNorm2d_29 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_26 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_30 [1, 256, 14, 14] [1, 256, 14, 14] 589824 115605504 BatchNorm2d_30 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_27 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_31 [1, 256, 14, 14] [1, 1024, 14, 14] 262144 51380224 BatchNorm2d_31 [1, 1024, 14, 14] [1, 1024, 14, 14] 2048 Not Available ReLU_28 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Bottleneck_9 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Conv2d_32 [1, 1024, 14, 14] [1, 256, 14, 14] 262144 51380224 BatchNorm2d_32 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_29 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_33 [1, 256, 14, 14] [1, 256, 14, 14] 589824 115605504 BatchNorm2d_33 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_30 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_34 [1, 256, 14, 14] [1, 1024, 14, 14] 262144 51380224 BatchNorm2d_34 [1, 1024, 14, 14] [1, 1024, 14, 14] 2048 Not Available ReLU_31 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Bottleneck_10 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Conv2d_35 [1, 1024, 14, 14] [1, 256, 14, 14] 262144 51380224 BatchNorm2d_35 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_32 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_36 [1, 256, 14, 14] [1, 256, 14, 14] 589824 115605504 BatchNorm2d_36 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_33 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_37 [1, 256, 14, 14] [1, 1024, 14, 14] 262144 51380224 BatchNorm2d_37 [1, 1024, 14, 14] [1, 1024, 14, 14] 2048 Not Available ReLU_34 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Bottleneck_11 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Conv2d_38 [1, 1024, 14, 14] [1, 256, 14, 14] 262144 51380224 BatchNorm2d_38 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_35 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_39 [1, 256, 14, 14] [1, 256, 14, 14] 589824 115605504 BatchNorm2d_39 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_36 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_40 [1, 256, 14, 14] [1, 1024, 14, 14] 262144 51380224 BatchNorm2d_40 [1, 1024, 14, 14] [1, 1024, 14, 14] 2048 Not Available ReLU_37 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Bottleneck_12 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Conv2d_41 [1, 1024, 14, 14] [1, 256, 14, 14] 262144 51380224 BatchNorm2d_41 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_38 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_42 [1, 256, 14, 14] [1, 256, 14, 14] 589824 115605504 BatchNorm2d_42 [1, 256, 14, 14] [1, 256, 14, 14] 512 Not Available ReLU_39 [1, 256, 14, 14] [1, 256, 14, 14] 0 Not Available Conv2d_43 [1, 256, 14, 14] [1, 1024, 14, 14] 262144 51380224 BatchNorm2d_43 [1, 1024, 14, 14] [1, 1024, 14, 14] 2048 Not Available ReLU_40 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Bottleneck_13 [1, 1024, 14, 14] [1, 1024, 14, 14] 0 Not Available Conv2d_44 [1, 1024, 14, 14] [1, 512, 14, 14] 524288 102760448 BatchNorm2d_44 [1, 512, 14, 14] [1, 512, 14, 14] 1024 Not Available ReLU_41 [1, 512, 14, 14] [1, 512, 14, 14] 0 Not Available Conv2d_45 [1, 512, 14, 14] [1, 512, 7, 7] 2359296 115605504 BatchNorm2d_45 [1, 512, 7, 7] [1, 512, 7, 7] 1024 Not Available ReLU_42 [1, 512, 7, 7] [1, 512, 7, 7] 0 Not Available Conv2d_46 [1, 512, 7, 7] [1, 2048, 7, 7] 1048576 51380224 BatchNorm2d_46 [1, 2048, 7, 7] [1, 2048, 7, 7] 4096 Not Available Conv2d_47 [1, 1024, 14, 14] [1, 2048, 7, 7] 2097152 102760448 BatchNorm2d_47 [1, 2048, 7, 7] [1, 2048, 7, 7] 4096 Not Available ReLU_43 [1, 2048, 7, 7] [1, 2048, 7, 7] 0 Not Available Bottleneck_14 [1, 1024, 14, 14] [1, 2048, 7, 7] 0 Not Available Conv2d_48 [1, 2048, 7, 7] [1, 512, 7, 7] 1048576 51380224 BatchNorm2d_48 [1, 512, 7, 7] [1, 512, 7, 7] 1024 Not Available ReLU_44 [1, 512, 7, 7] [1, 512, 7, 7] 0 Not Available Conv2d_49 [1, 512, 7, 7] [1, 512, 7, 7] 2359296 115605504 BatchNorm2d_49 [1, 512, 7, 7] [1, 512, 7, 7] 1024 Not Available ReLU_45 [1, 512, 7, 7] [1, 512, 7, 7] 0 Not Available Conv2d_50 [1, 512, 7, 7] [1, 2048, 7, 7] 1048576 51380224 BatchNorm2d_50 [1, 2048, 7, 7] [1, 2048, 7, 7] 4096 Not Available ReLU_46 [1, 2048, 7, 7] [1, 2048, 7, 7] 0 Not Available Bottleneck_15 [1, 2048, 7, 7] [1, 2048, 7, 7] 0 Not Available Conv2d_51 [1, 2048, 7, 7] [1, 512, 7, 7] 1048576 51380224 BatchNorm2d_51 [1, 512, 7, 7] [1, 512, 7, 7] 1024 Not Available ReLU_47 [1, 512, 7, 7] [1, 512, 7, 7] 0 Not Available Conv2d_52 [1, 512, 7, 7] [1, 512, 7, 7] 2359296 115605504 BatchNorm2d_52 [1, 512, 7, 7] [1, 512, 7, 7] 1024 Not Available ReLU_48 [1, 512, 7, 7] [1, 512, 7, 7] 0 Not Available Conv2d_53 [1, 512, 7, 7] [1, 2048, 7, 7] 1048576 51380224 BatchNorm2d_53 [1, 2048, 7, 7] [1, 2048, 7, 7] 4096 Not Available ReLU_49 [1, 2048, 7, 7] [1, 2048, 7, 7] 0 Not Available Bottleneck_16 [1, 2048, 7, 7] [1, 2048, 7, 7] 0 Not Available AvgPool2d_1 [1, 2048, 7, 7] [1, 2048, 1, 1] 0 Not Available Linear_1 [1, 2048] [1, 1000] 2049000 2048000 Total Parameters: 25557032 Total Multiply Adds (For Convolution aand Linear Layers only): 4089184256 Number of Layers MaxPool2d : 1 layers Linear : 1 layers AvgPool2d : 1 layers Bottleneck : 16 layers ReLU : 49 layers Conv2d : 53 layers BatchNorm2d : 53 layers
Recommend
About Joyk
Aggregate valuable and interesting links.
Joyk means Joy of geeK