Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Upstream/2 write a predictor class to mimic the gec model #172

Open
wants to merge 77 commits into
base: master
Choose a base branch
from

Conversation

damien2012eng
Copy link

As discussed before, our team has been working on updating GECToR dependencies to the latest versions (Inference only). This is the second PR, one more PR is coming soon.
Link to the successful run: here
We also checked the impact for the training process. No breaks were identified.

First PR: Add unit tests and regression test. Additionally, bump Python version.
Second PR: Write a predictor class to mimic the GEC_model.
Third PR: Unpin all the dependency requirements

How to review:

Unit test:
pytest -v tests
Regression test:
python regression_tests/test_gector_roberta.py
python regression_tests/test_regression_data_predictor.py
Our plan is to submit one more PR. Suggestions and comments are welcome!

Zhaoyang Xie and others added 30 commits September 14, 2022 10:04
…ckage

Create Python package & bump to Python 3.8
…t_test_for_tokenToIndexers_and_tokenization

Features/add unit test for token to indexers and tokenization
…y gec_model,

get_token_action removed and placed in model.
There's a bug in here somewhere though because the output isn't exactly the same. I suspect there's an off by one error happening when the indices for edit operations are generated.
Frost45 and others added 30 commits October 12, 2022 12:28
…r_from_gec_model

Allennlp Predictor class that mimics gec_model
…tingService/gector into feature/fix-gec-predictor
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants