After scouring the internet for a couple of days looking for a code to run alexnet on VOC12 data for classification ending in vain, I had to write the integrated code myself and I a sharing the same for anyone to reuse.
Due to lack of time I am posting the codes for sorting & training separately. I will make an update when I get time, to write a single unified code. Meanwhile if anyone faces any error, just drop a comment (or email to anurag@anuragmeena.com) with details (including pytorh and torchvision versions) and I will look into it
I had a choice to write a custom dataloader or convert the data into a pre-defined data loader structure. I chose the latter and the image_sort.py code basically converts the VOC12 data into Imgefolder data structure. Just change the root directory string in the image_sort.py and run it.
This would create the following folders with hierarchvaliy:
|classes |train (training set) |<folders with class names> |val (validation set) |<folders with classnames>
After this go to main.py . It contains the actual code to initialize your neeural network and configure your net (whether you want to download pre trained model, what is your input, how many output classes you have, )
Download Link for Code files
image_sort.py
<code>
## Place this code in the same folder where Imagefolder exists
# creating data folder
import shutil
from shutil import copyfile
import os
all_files = os.listdir("Main/")
# print(all_files[0])
a=0
all_files_set = set(all_files)
# print(all_files_set)
print("_________________________________________________________________")
# To create all class directories
for name in all_files_set:
#print (name)
temp = name.split('_')
#print(temp[0])
if((os.path.isdir("classes_val/"+temp[0]))):
# if((os.path.isdir("classes/"+temp[0]))):
print("Directory already exists \n \n \n")
else:
os.makedirs("classes_val/"+temp[0])
# os.makedirs("classes/"+temp[0])
if(temp[0] == "train.txt" or temp[0] == "val.txt" or temp[0] =="trainval.txt"):
print("__")
for name in all_files_set:
temp_f = name.split('_')
if (temp_f[0] == "train.txt" or temp_f[0] == "val.txt" or temp_f[0] =="trainval.txt"):
print("\n ************** \n skipped train, val & trainval text files \n ************** \n ")
else:
if(temp_f[1] == "val.txt"):
# if(temp_f[1] == "train.txt"):
print("Opening Text file: "+name)
# print(name)
f = open("Main/"+name, "r")
lines = list(f)
# print(lines)
# lines = f.readlines()
f.close()
for line in lines:
temp = line.split(" ")
# print(temp[1])
print (temp)
if (temp[1] == '-1\n'):
print("skipfile")
elif (temp[2] == '1\n'):
# copyfile(src, dst)
print(temp[0]+".jpg")
shutil.copy("../JPEGImages/"+temp[0]+".jpg", "classes_val/"+temp_f[0]+"/")
# shutil.copy("../JPEGImages/"+temp[0]+".jpg", "classes/"+temp_f[0]+"/")
</code>
main.py
from __future__ import print_function
from __future__ import division
import torch
import torch.nn as nn
import torch.optim as optim
import numpy as np
import torchvision
from torchvision import datasets, models, transforms
import matplotlib.pyplot as plt
import time
import os
import copy
# import voc
def run():
print("PyTorch Version: ",torch.__version__)
print("Torchvision Version: ",torchvision.__version__)
# Top level data directory. Here we assume the format of the directory conforms
# to the ImageFolder structure
data_dir = "./classes/"
# Models to choose from [resnet, alexnet, vgg, squeezenet, densenet, inception]
model_name = "alexnet"
# Number of classes in the dataset
num_classes = 20
# Batch size for training (change depending on how much memory you have)
batch_size = 8
# Number of epochs to train for
num_epochs = 1
# Flag for feature extracting. When False, we finetune the whole model,
# when True we only update the reshaped layer params
feature_extract = True
def train_model(model, dataloaders, criterion, optimizer, num_epochs=25, is_inception=False):
since = time.time()
val_acc_history = []
best_model_wts = copy.deepcopy(model.state_dict())
best_acc = 0.0
for epoch in range(num_epochs):
print('Epoch {}/{}'.format(epoch, num_epochs - 1))
print('-' * 10)
# Each epoch has a training and validation phase
for phase in ['train', 'val']:
if phase == 'train':
model.train() # Set model to training mode
else:
model.eval() # Set model to evaluate mode
running_loss = 0.0
running_corrects = 0
# Iterate over data.
for inputs, labels in dataloaders[phase]:
inputs = inputs.to(device)
labels = labels.to(device)
# zero the parameter gradients
optimizer.zero_grad()
# forward
# track history if only in train
with torch.set_grad_enabled(phase == 'train'):
# Get model outputs and calculate loss
# Special case for inception because in training it has an auxiliary output. In train
# mode we calculate the loss by summing the final output and the auxiliary output
# but in testing we only consider the final output.
if is_inception and phase == 'train':
# From https://discuss.pytorch.org/t/how-to-optimize-inception-model-with-auxiliary-classifiers/7958
outputs, aux_outputs = model(inputs)
loss1 = criterion(outputs, labels)
loss2 = criterion(aux_outputs, labels)
loss = loss1 + 0.4*loss2
else:
outputs = model(inputs)
loss = criterion(outputs, labels)
_, preds = torch.max(outputs, 1)
# backward + optimize only if in training phase
if phase == 'train':
loss.backward()
optimizer.step()
# statistics
running_loss += loss.item() * inputs.size(0)
running_corrects += torch.sum(preds == labels.data)
epoch_loss = running_loss / len(dataloaders[phase].dataset)
epoch_acc = running_corrects.double() / len(dataloaders[phase].dataset)
print('{} Loss: {:.4f} Acc: {:.4f}'.format(phase, epoch_loss, epoch_acc))
# deep copy the model
if phase == 'val' and epoch_acc > best_acc:
best_acc = epoch_acc
best_model_wts = copy.deepcopy(model.state_dict())
if phase == 'val':
val_acc_history.append(epoch_acc)
print()
time_elapsed = time.time() - since
print('Training complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))
print('Best val Acc: {:4f}'.format(best_acc))
# load best model weights
model.load_state_dict(best_model_wts)
torch.save(model.state_dict(), "./classes/abc.pth")
return model, val_acc_history
"""
#saving a model weights
model.save(model.state_dict(), ./classes)
"""
def set_parameter_requires_grad(model, feature_extracting):
if feature_extracting:
for param in model.parameters():
param.requires_grad = False
# model.classifier[6] = nn.Linear(4096,num_classes)
def initialize_model(model_name, num_classes, feature_extract, use_pretrained=True):
# Initialize these variables which will be set in this if statement. Each of these
# variables is model specific.
model_ft = None
input_size = 0
if model_name == "resnet":
""" Resnet18
"""
model_ft = models.resnet18(pretrained=use_pretrained)
set_parameter_requires_grad(model_ft, feature_extract)
num_ftrs = model_ft.fc.in_features
model_ft.fc = nn.Linear(num_ftrs, num_classes)
input_size = 224
elif model_name == "alexnet":
""" Alexnet
"""
model_ft = models.alexnet(pretrained=use_pretrained)
set_parameter_requires_grad(model_ft, feature_extract)
num_ftrs = model_ft.classifier[6].in_features
model_ft.classifier[6] = nn.Linear(num_ftrs,num_classes)
input_size = 224
elif model_name == "vgg":
""" VGG11_bn
"""
model_ft = models.vgg11_bn(pretrained=use_pretrained)
set_parameter_requires_grad(model_ft, feature_extract)
num_ftrs = model_ft.classifier[6].in_features
model_ft.classifier[6] = nn.Linear(num_ftrs,num_classes)
input_size = 224
elif model_name == "squeezenet":
""" Squeezenet
"""
model_ft = models.squeezenet1_0(pretrained=use_pretrained)
set_parameter_requires_grad(model_ft, feature_extract)
model_ft.classifier[1] = nn.Conv2d(512, num_classes, kernel_size=(1,1), stride=(1,1))
model_ft.num_classes = num_classes
input_size = 224
elif model_name == "densenet":
""" Densenet
"""
model_ft = models.densenet121(pretrained=use_pretrained)
set_parameter_requires_grad(model_ft, feature_extract)
num_ftrs = model_ft.classifier.in_features
model_ft.classifier = nn.Linear(num_ftrs, num_classes)
input_size = 224
elif model_name == "inception":
""" Inception v3
Be careful, expects (299,299) sized images and has auxiliary output
"""
model_ft = models.inception_v3(pretrained=use_pretrained)
set_parameter_requires_grad(model_ft, feature_extract)
# Handle the auxilary net
num_ftrs = model_ft.AuxLogits.fc.in_features
model_ft.AuxLogits.fc = nn.Linear(num_ftrs, num_classes)
# Handle the primary net
num_ftrs = model_ft.fc.in_features
model_ft.fc = nn.Linear(num_ftrs,num_classes)
input_size = 299
else:
print("Invalid model name, exiting...")
exit()
return model_ft, input_size
# Initialize the model for this run
model_ft, input_size = initialize_model(model_name, num_classes, feature_extract, use_pretrained=True)
# Print the model we just instantiated
print(model_ft)
# Data augmentation and normalization for training
# Just normalization for validation
data_transforms = {
'train': transforms.Compose([
transforms.RandomResizedCrop(input_size),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
]),
'val': transforms.Compose([
transforms.Resize(input_size),
transforms.CenterCrop(input_size),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
]),
}
print("Initializing Datasets and Dataloaders...")
# Create training and validation datasets
image_datasets = {x: datasets.ImageFolder(os.path.join(data_dir, x), data_transforms[x]) for x in ['train', 'val']}
# Create training and validation dataloaders
dataloaders_dict = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=batch_size, shuffle=True, num_workers=4) for x in ['train', 'val']}
# Create training and validation datasets
#image_datasets = voc.VOCDetection(data_dir, year='2012', image_set='train', download=False, transform=transform, target_transform=None)
# Create training and validation dataloaders
#dataloaders_dict = torch.utils.data.DataLoader(trainset, batch_size=4, shuffle=True, num_workers=2)
# Detect if we have a GPU available
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
# device = "cpu"
# Send the model to GPU
model_ft = model_ft.to(device)
# Gather the parameters to be optimized/updated in this run. If we are
# finetuning we will be updating all parameters. However, if we are
# doing feature extract method, we will only update the parameters
# that we have just initialized, i.e. the parameters with requires_grad
# is True.
params_to_update = model_ft.parameters()
print("Params to learn:")
if feature_extract:
params_to_update = []
for name,param in model_ft.named_parameters():
if param.requires_grad == True:
params_to_update.append(param)
print("\t",name)
else:
for name,param in model_ft.named_parameters():
if param.requires_grad == True:
print("\t",name)
# Observe that all parameters are being optimized #lr=0.0001
optimizer_ft = optim.SGD(params_to_update, lr=0.9, momentum=0.9)
# Setup the loss fxn
criterion = nn.CrossEntropyLoss()
# Train and evaluate
model_ft, hist = train_model(model_ft, dataloaders_dict, criterion, optimizer_ft, num_epochs=num_epochs, is_inception=(model_name=="inception"))
print("Chekcpoint save started")
# torch.save(net.state_dict(), "model_{}.pt".format(epoch))
# torch.save(net, 'model.pt')
print("Chekcpoint saved")
if __name__ == '__main__':
run()