deeping-flow

deeping-flow copied to clipboard

Youly172

corpus文件在哪儿？

baiziyuandyufei

Hi， https://github.com/ne7ermore/deeping-flow/blob/master/reinforced-translate/model.py b_words = model.sample(self.prev) s_words, s_props = model.sample(self.prev, False) rewards = self.compute_levenshtein(model.tgt, s_words) baseline = self.compute_levenshtein(model.tgt, b_words) advantage = rewards - baseline 其中的 b-words 和 s-words 都是 Tensor 类型，...

JinmingZhao

Hi， https://github.com/ne7ermore/deeping-flow/blob/master/reinforced-translate/model.py#L175 mask = pad_mask(model.tgt, EOS, [args.batch_size, args.max_len]) 这里是不是应该是 MC 采样的作为target？ 应该是s_words? Thanks

JinmingZhao

Why do you not apply tf.nn.log_softmax() in to probs in sample function of reinforced-translate/model.py

1

pemywei

UnicodeDecodeError in corpus.py

1

Sorry, I have encountered a UnicodeDecodeError when I run the corpus.py with python3.6. The problem is following: Traceback (most recent call last): File "corpus.py", line 136, in Corpus() File "corpus.py",...

yuewang1402

dynamic_rnn的outputs似乎不能保存for循环前几步的输出

1

https://github.com/ne7ermore/deeping-flow/blob/433296c0a2cd1ebd6db8524aa9d20bbc59d0f31f/deep-reinforced-sum-model/attention.py#L46 这里说dec_out的形状是bsz * time * dec_hsz。而dec_out是model.py中tf.nn.dynamic_rnn的输出。因为你每个时间步的对dynamic_rnn的输入形状是bsz * 1 * emb_dim，所以dec_out的形状应该是bsz * 1 * dec_hsz，他不会保存前几个时间步的状态。麻烦您看看我的疑问是否正确。

theoqian