num_img_tr = len(trainloader) num_img_ts = len(testloader) running_loss_tr = 0.0 running_loss_ts = 0.0 aveGrad = 0 global_step = 0 print("Training Network") min_mean_error = 1000.0 # Main Training and Testing Loop for epoch in range(resume_epoch, nEpochs): # print(epoch) start_time = timeit.default_timer() if epoch % p['epoch_size'] == p['epoch_size'] - 1: lr_ = utils.lr_poly(p['lr'], epoch, nEpochs, 0.9) print('(poly lr policy) learning rate: ', lr_) optimizer = optim.Adam(net.parameters(), lr=p['lr'], weight_decay=p['wd']) net.train() for ii, sample_batched in enumerate(trainloader): # inputs, labels = sample_batched['image'], sample_batched['label'] inputs, labels = sample_batched[0], sample_batched[1] # Forward-Backward of the mini-batch inputs, labels = Variable(inputs, requires_grad=True), Variable(labels) global_step += inputs.data.shape[0]
utils.generate_param_report(os.path.join(save_dir, exp_name + '.txt'), p) num_img_tr = len(trainloader) num_img_ts = len(testloader) running_loss_tr = 0.0 running_loss_ts = 0.0 aveGrad = 0 print("Training Network") # Main Training and Testing Loop for epoch in range(resume_epoch, nEpochs): start_time = timeit.default_timer() if epoch % p['epoch_size'] == p['epoch_size'] - 1: lr_ = utils.lr_poly(p['lr'], epoch, nEpochs, 0.9) print('(poly lr policy) learning rate: ', lr_) optimizer = optim.SGD(net.parameters(), lr=lr_, momentum=p['momentum'], weight_decay=p['wd']) net.train() for ii, sample_batched in enumerate(trainloader): inputs, gts = sample_batched['image'], sample_batched['gt'] # Forward-Backward of the mini-batch inputs, gts = Variable(inputs, requires_grad=True), Variable(gts) if gpu_id >= 0: inputs, gts = inputs.cuda(), gts.cuda() output = net.forward(inputs)