import numpy as np
|
import torch
|
import torch.nn as nn
|
from torch.autograd import Variable
|
from torch.utils.data import DataLoader
|
from torchvision import transforms
|
import torchvision
|
import torch.backends.cudnn as cudnn
|
import torch.nn.functional as F
|
|
import cv2
|
import matplotlib.pyplot as plt
|
import sys
|
import os
|
import argparse
|
|
import datasets
|
import hopenet
|
import torch.utils.model_zoo as model_zoo
|
|
model_urls = {
|
'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',
|
'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',
|
'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',
|
'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',
|
'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',
|
}
|
|
def parse_args():
|
"""Parse input arguments."""
|
parser = argparse.ArgumentParser(description='Head pose estimation using the Hopenet network.')
|
parser.add_argument('--gpu', dest='gpu_id', help='GPU device id to use [0]',
|
default=0, type=int)
|
parser.add_argument('--num_epochs', dest='num_epochs', help='Maximum number of training epochs.',
|
default=5, type=int)
|
parser.add_argument('--batch_size', dest='batch_size', help='Batch size.',
|
default=16, type=int)
|
parser.add_argument('--lr', dest='lr', help='Base learning rate.',
|
default=0.001, type=float)
|
parser.add_argument('--data_dir', dest='data_dir', help='Directory path for data.',
|
default='', type=str)
|
parser.add_argument('--filename_list', dest='filename_list', help='Path to text file containing relative paths for every example.',
|
default='', type=str)
|
parser.add_argument('--finetune', dest='finetune', help='Boolean: finetune or from Imagenet pretrain.',
|
default=False, type=bool)
|
parser.add_argument('--snapshot', dest='snapshot', help='Path to finetune snapshot.',
|
default='', type=str)
|
|
args = parser.parse_args()
|
|
return args
|
|
def get_ignored_params(model):
|
# Generator function that yields ignored params.
|
b = []
|
b.append(model.conv1)
|
b.append(model.bn1)
|
b.append(model.layer1)
|
b.append(model.layer2)
|
b.append(model.layer3)
|
b.append(model.layer4)
|
for i in range(len(b)):
|
for j in b[i].modules():
|
for k in j.parameters():
|
yield k
|
|
def get_non_ignored_params(model):
|
# Generator function that yields params that will be optimized.
|
b = []
|
b.append(model.fc_yaw)
|
b.append(model.fc_pitch)
|
b.append(model.fc_roll)
|
for i in range(len(b)):
|
for j in b[i].modules():
|
for k in j.parameters():
|
yield k
|
|
def load_filtered_state_dict(model, snapshot):
|
# By user apaszke from discuss.pytorch.org
|
model_dict = model.state_dict()
|
# 1. filter out unnecessary keys
|
snapshot = {k: v for k, v in snapshot.items() if k in model_dict}
|
# 2. overwrite entries in the existing state dict
|
model_dict.update(snapshot)
|
# 3. load the new state dict
|
model.load_state_dict(model_dict)
|
|
if __name__ == '__main__':
|
args = parse_args()
|
|
cudnn.enabled = True
|
num_epochs = args.num_epochs
|
batch_size = args.batch_size
|
gpu = args.gpu_id
|
|
if not os.path.exists('output/snapshots'):
|
os.makedirs('output/snapshots')
|
|
# ResNet101 with 3 outputs
|
# model = hopenet.Hopenet(torchvision.models.resnet.Bottleneck, [3, 4, 23, 3], 66)
|
# ResNet50
|
model = hopenet.Hopenet(torchvision.models.resnet.Bottleneck, [3, 4, 6, 3], 66)
|
# ResNet18
|
# model = hopenet.Hopenet(torchvision.models.resnet.BasicBlock, [2, 2, 2, 2], 66)
|
|
if args.finetune:
|
print 'Finetuning.'
|
model.load_state_dict(torch.load(args.snapshot))
|
else:
|
load_filtered_state_dict(model, model_zoo.load_url(model_urls['resnet50']))
|
|
print 'Loading data.'
|
|
transformations = transforms.Compose([transforms.Scale(224),
|
transforms.RandomCrop(224),
|
transforms.ToTensor()])
|
|
transformations = transforms.Compose([transforms.Scale(224),
|
transforms.RandomCrop(224), transforms.ToTensor(),
|
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])])
|
|
pose_dataset = datasets.BIWI(args.data_dir, args.filename_list,
|
transformations)
|
train_loader = torch.utils.data.DataLoader(dataset=pose_dataset,
|
batch_size=batch_size,
|
shuffle=True,
|
num_workers=2)
|
|
model.cuda(gpu)
|
criterion = nn.CrossEntropyLoss().cuda()
|
reg_criterion = nn.MSELoss().cuda()
|
# Regression loss coefficient
|
alpha = 0.01
|
|
idx_tensor = [idx for idx in xrange(66)]
|
idx_tensor = torch.FloatTensor(idx_tensor).cuda(gpu)
|
|
# optimizer = torch.optim.Adam([{'params': get_ignored_params(model), 'lr': args.lr},
|
# {'params': get_non_ignored_params(model), 'lr': args.lr}],
|
# lr = args.lr)
|
|
optimizer = torch.optim.SGD([{'params': get_ignored_params(model), 'lr': args.lr},
|
{'params': get_non_ignored_params(model), 'lr': args.lr}],
|
lr = args.lr, momentum=0.9, weight_decay=0.01)
|
|
print 'Ready to train network.'
|
|
for epoch in range(num_epochs):
|
for i, (images, labels, name) in enumerate(train_loader):
|
images = Variable(images.cuda(gpu))
|
label_yaw = Variable(labels[:,0].cuda(gpu))
|
label_pitch = Variable(labels[:,1].cuda(gpu))
|
label_roll = Variable(labels[:,2].cuda(gpu))
|
|
optimizer.zero_grad()
|
model.zero_grad()
|
|
yaw, pitch, roll = model(images)
|
|
# Cross entropy loss
|
loss_yaw = criterion(yaw, label_yaw)
|
loss_pitch = criterion(pitch, label_pitch)
|
loss_roll = criterion(roll, label_roll)
|
|
# MSE loss
|
yaw_predicted = F.softmax(yaw)
|
pitch_predicted = F.softmax(pitch)
|
roll_predicted = F.softmax(roll)
|
|
yaw_predicted = torch.sum(yaw_predicted.data * idx_tensor, 1)
|
pitch_predicted = torch.sum(pitch_predicted.data * idx_tensor, 1)
|
roll_predicted = torch.sum(roll_predicted.data * idx_tensor, 1)
|
|
loss_reg_yaw = reg_criterion(yaw_predicted, label_yaw.float())
|
loss_reg_pitch = reg_criterion(pitch_predicted, label_pitch.float())
|
loss_reg_roll = reg_criterion(roll_predicted, label_roll.float())
|
|
# Total loss
|
loss_yaw += alpha * loss_reg_yaw
|
loss_pitch += alpha * loss_reg_pitch
|
loss_roll += alpha * loss_reg_roll
|
|
loss_seq = [loss_yaw, loss_pitch, loss_roll]
|
grad_seq = [torch.Tensor(1).cuda(gpu) for _ in range(len(loss_seq))]
|
torch.autograd.backward(loss_seq, grad_seq)
|
optimizer.step()
|
|
# print ('Epoch [%d/%d], Iter [%d/%d] Losses: Yaw %.4f, Pitch %.4f, Roll %.4f'
|
# %(epoch+1, num_epochs, i+1, len(pose_dataset)//batch_size, loss_yaw.data[0], loss_pitch.data[0], loss_roll.data[0]))
|
|
if (i+1) % 100 == 0:
|
print ('Epoch [%d/%d], Iter [%d/%d] Losses: Yaw %.4f, Pitch %.4f, Roll %.4f'
|
%(epoch+1, num_epochs, i+1, len(pose_dataset)//batch_size, loss_yaw.data[0], loss_pitch.data[0], loss_roll.data[0]))
|
# if epoch == 0:
|
# torch.save(model.state_dict(),
|
# 'output/snapshots/resnet50_ftbiwi_iter_'+ str(i+1) + '.pkl')
|
|
# Save models at numbered epochs.
|
if epoch % 1 == 0 and epoch < num_epochs - 1:
|
print 'Taking snapshot...'
|
torch.save(model.state_dict(),
|
'output/snapshots/resnet50_ftbiwi_epoch_'+ str(epoch+1) + '.pkl')
|
|
# Save the final Trained Model
|
torch.save(model.state_dict(), 'output/snapshots/resnet50_ftbiwi_epoch_' + str(epoch+1) + '.pkl')
|
