Linkai

Can you explain how fuzzy match works?

1

That is, given a source and a translation from machine, and translation memory, how the TM assists in enhancing the translation?

有小数点数字的分词

1

分词的时候会将小数点单独成一个部分，这个不合理。

Intersect operation

9

The intersect operation always returns results, however, it is not true when two blocks are not overlapped.

I am sorry this issue is not directly related to the project. In MT, some words/phrases are not translated, but copied from source sentences, such as person names, company names,...

[Question] Inadequate translation

3

NMT is better than SMT in fluency, while it suffers from inadequate translation for long sentences. I have come across research where [coverage](http://www.aclweb.org/anthology/P16-1008) is modelled for NMT. Does Transformer have...

Sorry, but i am not sure that there is beam search algorithm according to your code. Could you please advise in which script the beam search is used? Thank you.

### Feature description There is only translation server. Will you consider add scorer function in the server, too? ### Example Add a usage examples for the new feature, e.g. a...

I suppose your sequential model has one-layer encoder and possible multi-layer decoder. however, in most cases encoder has also many layers. To me, for _ in range(num_layers): model.add(LSTM(hidden_size, return_sequences=True)) may...

The prefix label id is made -100, as `source_len` counts an extra `eos` token, see below code. Subtracting 1 from `source_len` is necessary. https://github.com/tatsu-lab/stanford_alpaca/blob/65512697dc67779a6e53c267488aba0ec4d7c02a/train.py#L123 Also add an extra `\n` at...