bug

[ddy-ddy]

hello,author.i seem to have found a bug.I don't know if it was caused by my improper operation.
I run predict.py with these parameter。
python predict.py --test_context ./data/test_context.txt \ --test_ques ./data/test_ques.txt \ --test_ids ./data/test_asin.txt \ --word_embeddings ./data/word_embeddings.p \ --vocab ./data/vocab.p \ --kwd_vocab ./data/train_kwd_vocab.txt \ --load_models_dir ./ckpt/##YOUR_MODEL##.models \ --load_hparams_dir ./hparams/##YOUR_MODEL##.json \ --load_filter_dir ./data/kwd_filter_dict.json \ --out_dir ./output \ --threshold 0.075 \ --batch_size 128 \ --diverse_beam \ --save_all_beam \ --beam_size 6 \ --diverse_group 3 \ --diverse_lambda 0.4 \ --batch_size 128
But it got this error.

[blmoistawinde]

Should be similar as #1 , seems the current pytorch version don't convert index to integer as my original version, now fixed.

[ddy-ddy]

Should be similar as #1 , seems the current pytorch version don't convert index to integer as my original version, now fixed.

ok，thank you so much.