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

Update/adult gpt chat completion few shot #146

Merged
merged 15 commits into from
Aug 7, 2023

Conversation

firojalam
Copy link
Contributor

No description provided.

@firojalam firojalam requested a review from fdalvi July 10, 2023 13:41
ahmed451 and others added 11 commits July 23, 2023 14:34
* Add ArabicSequenceTagging

* push Segmentation task

* Add segmentation changes

* push segmentation and merge changes

* Fix evaluation

* change None data to unsegmented words

* fix none segmentation, re-format code

* fix segmentation except

* fix changes with upstream

* Format code

* Update segmentation task and assets from feat/POS branch

* Format code

* Fix evaluation and GPT4 asset

* Add latest assets from Ahmed

* Fix test for multi-config

* Add code to remove extra spaces in assets

---------

Co-authored-by: Fahim Imaduddin Dalvi <faimaduddin@hbku.edu.qa>
* Add diacritizaton module

* Update ArabicDiacritization.py

Use undiacritized tokens as fall back for None results.

* Format code

* Add comments and minor fixes

* More fixes to dataloader

---------

Co-authored-by: Ahmed Abdelali <ahmed.abdelali@gmail.com>
* Added few shot learning script for Arcd dataset

* Added few shot asset for ARCD data

* Added few shot script for MLQA

* added few shot script for TydiQA

* added few shot script for XQUAD

* Format code

* Fix zeroshot assets to not mask prediction failure

* Remove hardcoded engine names

* Fix MLQA paths

---------

Co-authored-by: Fahim Imaduddin Dalvi <faimaduddin@hbku.edu.qa>
* Added scripts for QA ZS

* Format code

* Save input along with fewshot samples

* Fix MLQA path

---------

Co-authored-by: Fahim Imaduddin Dalvi <faimaduddin@hbku.edu.qa>
Reset sentence/labels tokens correctly.
* Added BLOOOMZ implementation for STS Track 1

* changed the post processing sts_bloomz

* changed the prompt.

* edited the prompt

* changed the rating scale from 0-5 to 0-10 and divided the output by 2

* code formatting

* Remove dead code

* Added FS implementation for Hate Speech, to replace previous implementation

* Update data paths in STSTrack1 asset for BLOOM

* Format code

* Update data path in HS asset

* Update Adult assets hierarchy and data paths

---------

Co-authored-by: maramhasanain <maramhasanain@gmail.com>
Co-authored-by: Fahim Imaduddin Dalvi <faimaduddin@hbku.edu.qa>
Co-authored-by: sabdaljalil <sabdaljalil@hbku.edu.qa>
Co-authored-by: Fahim Imaduddin Dalvi <faimaduddin@hbku.edu.qa>
* Add MT GPT4 with multi-configs

* Make GPT3 asset multi-config

* Format code

* Fix prompt function content to match ChatCompletion API

---------

Co-authored-by: Ahmed Abdelali <ahmed.abdelali@gmail.com>
Co-authored-by: Ahmed Abdelali <aabdelali@hbku.edu.qa>
* DialectADI task added, updated init file

* Reorganize older DialectID assets

* Revert task to DialectID

---------

Co-authored-by: Fahim Imaduddin Dalvi <faimaduddin@hbku.edu.qa>
Co-authored-by: Fahim Imaduddin Dalvi <faimaduddin@hbku.edu.qa>
@fdalvi
Copy link
Collaborator

fdalvi commented Jul 23, 2023

Also the PR says "Update/adult gpt chat completion few shot" but is actually adding a BLOOMZ zeroshot asset for Adult task, just wanted to make sure some files were not missed?

@fdalvi fdalvi merged commit ce4aea4 into main Aug 7, 2023
4 checks passed
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.

6 participants