Fix a bug caused by commas in the embeddings.w2v.txt file

[yanshengjia]

No description provided.

[yanshengjia]

I met this error when running nea with the --emb option.

Traceback (most recent call last): File "train_nea.py", line 162, in <module> model = create_model(args, train_y.mean(axis=0), overal_maxlen, vocab) File "/Users/yanshengjia/Desktop/aes/nea/nea/models.py", line 138, in create_model emb_reader = EmbReader(args.emb_path, emb_dim=args.emb_dim) File "/Users/yanshengjia/Desktop/aes/nea/nea/w2vEmbReader.py", line 39, in __init__ assert len(tokens) == self.emb_dim + 1, 'The number of dimensions does not match the header info' AssertionError: The number of dimensions does not match the header info

I found that this error is caused by commas in En_vectors.txt.

The float numbers in En_vectors.txt are supposed to be separated by space instead of comma.

And the original code can only deal with the situation that float numbers are separated by space.

[izk8]

Ran into the same issue. Thanks Yangshengjia!

[yanshengjia]

You are welcome!

[jkdufair]

👍 This would help a lot!

[nayna-porwal]

caught the same problem and have word2vec.bin file not .txt file? any solution !!!!