Pytorch版本代碼修正

Pytorch版本代碼修正

概述：老版本代碼（0.3.0）修正爲（0.4.0）

1.UserWarning: nn.init.xavier_uniform is now deprecated in favor of nn.init.xavier_uniform_.init.xavier_uniform(param)

注：init.xavier_uniform 改爲 init.xavier_uniform_

2.UserWarning: volatile was removed and now has no effect. Use `with torch.no_grad():` instead. targets = [Variable(anno.cuda(), volatile=True) for anno in targets]

注：im = Variable(im.cuda()，volatile=True）
應該改成im = Variable(im.cuda())

3.UserWarning: invalid index of a 0-dim tensor. This will be an error in PyTorch 0.5. Use tensor.item() to convert a 0-dim tensor to a Python number

注：loss.data[0]
應該改成loss.item()

4.UserWarning: nn.functional.upsample is deprecated. Use nn.functional.interpolate instead.

注：upsample 替換爲interpolate

5.UserWarning: size_average and reduce args will be deprecated, please use reduction='sum' instead.warnings.warn(warning.format(ret))

注：size_average=False 改爲 reduction='sum’