Zhuoran Lyu

Results 7 comments of Zhuoran Lyu

I tried to use torch.nn.DataParallel(seq2seq_model).cuda() but it failed. It gives me this error: `TypeError Traceback (most recent call last) in () 17 batch_label = batch_label.cuda() 18 ---> 19 decoder_outputs, decoder_hidden,...

Same quesition, looking for a mathematical proof or some papers about this.

We can view conv op as matrix multiplication, then everything will be easy to understand.

> > 请问使用其他超分模型(如EDSR),会出现这种情况吗?还是这是所有超分问题的通病? > > 目前我们使用近似的FLOPS的其他模型暂时没有这种问题,感觉可能还是纯3x3卷积串联处理实际非bicubic LR的能力有限,可能采用作者table 4中融入其他模型的方案更有效一些。 说的非常对,这也是我们遇到的问题,生产环境中这些玩意儿没一个灵的,还是得靠edsr或者rrdb

@windj007 Hi, yes. I'm trying to fine-tune the model on my dataset. May I know the steps for that? THX