hred

hred copied to clipboard

dict file and dataset

5

Can you please provide the dataset and dictionary file?

nikhiljaiswal

hi, thanks for your code and sharing. Because i don't have the Movie-triples dataset. I have some dialogs with different number of turns. So may i ask you, if the...

lytum

dataset needed

1

I cannot find the dataset file. please help me to get them.

xxgt

Lack of data files

1

There are some files which are missing. They are dict file ('/home/harshals/hed-dlg/Data/MovieTriples/Training.dict.pkl'), eval source ('/data2/chatbot_eval_issues/results/AMT_NCM_Test_NCM_Joao/neural_conv_model_eval_source.txt'). Please help me to get them.

khoaipx

Thanks so much for this repo! In `custom_collate_fn`: (https://github.com/hsgodhia/hred/blob/master/util.py#L72) We ``` return u1_batch[sort1, :], u1_lens[sort1], u2_batch[sort2, :], u2_lens[sort2], u3_batch[sort3, :], u3_lens[sort3] ``` Doesn't this imply that our `[d1,d2,d3]` `DialogTurns` will...

sshleifer

in util.py, you did "sort" for u1_batch, u2_batch, u3_batch Will this lead to u1, u2 and u3 don't match each other?

evelinehong

decode for generating only one response?

2

Thank you for your code! But in original HRED, for k context and one response, HRED generates k utterances. It seems that you only generate the final response in training...

luomuqinghan

dataloader and result

2

https://github.com/hsgodhia/hred/blob/3ec59b6ad29a6db1b77379ac31ccc94eea628adf/util.py#L69 https://github.com/hsgodhia/hred/blob/3ec59b6ad29a6db1b77379ac31ccc94eea628adf/util.py#L72 I think it breaks the alignment. for example there are [u0_1, u0_2, u0_3], [u1_1, u0_2, u0_3] data, but after the alignment, the data can be [u0_1, u1_2, u0_3],...

Ri0S