nouhadziri / dialogentailment Goto Github PK

View Code? Open in Web Editor NEW

74.0 7.0 5.0 82 KB

The implementation of the paper "Evaluating Coherence in Dialogue Systems using Entailment"

Home Page: https://arxiv.org/abs/1904.03371

License: MIT License

Python 95.86% Shell 1.25% Jsonnet 2.89%

evaluation-framework natural-language-inference dialogue-evaluation bert

dialogentailment's People

Contributors

Stargazers

Watchers

Forkers

syx528911137 hjw1 spuronlee e0397123 abisowri

dialogentailment's Issues

How to test with other metrics ?

Hi. I have already trained my BERT model and tested my responses with the entailment model.

But I didn't know how to test with other metrics, i.e., Semantic Similarity, Word-level metrics, Consistency by textual entailment.

Can you show a script as an example?
Thanks

Errors when training BERT

Hi @korymath @ehsk ,

Thanks for your great work. I followed your instruction to train BERT and I got the following error:

Better speed can be achieved with apex installed from https://www.github.com/nvidia/apex.
Namespace(bert_model='bert-base-uncased', cache_dir='', do_eval=True, do_lower_case=True, do_train=True, eval_batch_size=8, eval_dataset='/content/gdrive/MyDrive/A/factoid_one_focus/CoQA/Answer-unaware/Evaluation/DialogEntailment/convai_nli_valid_both_revised_no_cands_ctx2_v1.3.jsonl', gradient_accumulation_steps=1, learning_rate=5e-05, local_rank=-1, max_seq_length=128, model='bert-base-uncased', no_cuda=False, num_train_epochs=3.0, output_dir='/content/gdrive/MyDrive/A/factoid_one_focus/CoQA/Answer-unaware/Evaluation/DialogEntailment/output', seed=42, train_batch_size=32, train_dataset='/content/gdrive/MyDrive/A/factoid_one_focus/CoQA/Answer-unaware/Evaluation/DialogEntailment/convai_nli_train_both_revised_no_cands_ctx2_v1.3.jsonl', warmup_proportion=0.1)
08/02/2022 08:42:33 - INFO - dialogentail.huggingface.finetune_bert -   device: cpu n_gpu: 0
08/02/2022 08:42:34 - INFO - pytorch_pretrained_bert.file_utils -   https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-uncased-vocab.txt not found in cache, downloading to /tmp/tmpvr7fii78
100% 231508/231508 [00:00<00:00, 907140.45B/s]
08/02/2022 08:42:35 - INFO - pytorch_pretrained_bert.file_utils -   copying /tmp/tmpvr7fii78 to cache at /root/.pytorch_pretrained_bert/26bc1ad6c0ac742e9b52263248f6d0f00068293b33709fae12320c0e35ccfbbb.542ce4285a40d23a559526243235df47c5f75c197f04f37d1a0c124c32c9a084
08/02/2022 08:42:35 - INFO - pytorch_pretrained_bert.file_utils -   creating metadata file for /root/.pytorch_pretrained_bert/26bc1ad6c0ac742e9b52263248f6d0f00068293b33709fae12320c0e35ccfbbb.542ce4285a40d23a559526243235df47c5f75c197f04f37d1a0c124c32c9a084
08/02/2022 08:42:35 - INFO - pytorch_pretrained_bert.file_utils -   removing temp file /tmp/tmpvr7fii78
08/02/2022 08:42:35 - INFO - pytorch_pretrained_bert.tokenization -   loading vocabulary file https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-uncased-vocab.txt from cache at /root/.pytorch_pretrained_bert/26bc1ad6c0ac742e9b52263248f6d0f00068293b33709fae12320c0e35ccfbbb.542ce4285a40d23a559526243235df47c5f75c197f04f37d1a0c124c32c9a084
Traceback (most recent call last):
  File "/usr/lib/python3.7/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/usr/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/content/gdrive/MyDrive/A/factoid_one_focus/CoQA/Answer-unaware/Evaluation/DialogEntailment/dialogentail/huggingface/__main__.py", line 93, in <module>
    main()
  File "/content/gdrive/MyDrive/A/factoid_one_focus/CoQA/Answer-unaware/Evaluation/DialogEntailment/dialogentail/huggingface/__main__.py", line 89, in main
    bert_run(args)
  File "/content/gdrive/MyDrive/A/factoid_one_focus/CoQA/Answer-unaware/Evaluation/DialogEntailment/dialogentail/huggingface/finetune_bert.py", line 353, in run
    train_examples = processor.get_train_examples(args.train_dataset)
  File "/content/gdrive/MyDrive/A/factoid_one_focus/CoQA/Answer-unaware/Evaluation/DialogEntailment/dialogentail/huggingface/finetune_bert.py", line 111, in get_train_examples
    self._read_tsv(data_file), "train")
  File "/content/gdrive/MyDrive/A/factoid_one_focus/CoQA/Answer-unaware/Evaluation/DialogEntailment/dialogentail/huggingface/finetune_bert.py", line 129, in _create_examples
    text_a = line[8]
IndexError: list index out of range

I linked the train and valid paths to convai_nli_train_both_revised_no_cands_ctx2_v1.3.jsonl and convai_nli_valid_both_revised_no_cands_ctx2_v1.3.jsonl. Do you have any idea what happened?

Thanks!

Test script

Can you guide me how to test after training cause it described a little vaguely.
Please show a test script you've made.
Thanks,

could you please release the already trained model

Thanks for your great work, and I am impressed by the work!
could you please release the trained model for a test

Thanks

Where can I find the original dataset

Dear author,
I am impressed by the work! I have a few questions though.

It seems the proposed dataset is built upon a previous dataset called ConvAI. Is this dataset also named PERSONA-CHAT Dataset? In addition, where can I download this original dataset? i.e. the convai_file in convai_to_nli.py?
As I observed from the proposed dataset, each dialogue only has one turn. I wonder if you have tried multi-turn dialogue? If so, how is the performance?
Thanks!

How to obtain the results of Table 2 in the paper?

ESIM+ELMo and BERT both output labels like entailment, contradiction or neutral. But in Reddit and OpenSubtitles majority vote of the 4-scale human rating constitutes the labels(Excellent, good, poor, bad). How to calculate accuracy in this situation?
Thank you very much!

Can I use it directly to evaluate the coherence between the response and the history utterances?

After installing the DialogEntailment package, how can I use it to score a certain response given the conversation history in Python? Can you give me a simple demo?

Thank you very much!

How to use BERT to evaluate directly to my conversation?

Hi @ehsk , @korymath ,

Thanks for your great work. May I ask if there is any inference scripts so that I can run them to evaluate my generated dialog? Look forward hearing from you soon.

Thanks!

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.