GitLab

# Conflicts:
#	lm_eval/models/huggingface.py

* download nltk `punkt_tab` on LOCAL_RANK=0

* remove print

* remove `time`

* nit

* switch `max_tokens` for `max_completion_tokens`. OpenAI ChatCompletions

* remove stop, temp=1 for o1

* add chat assertion

* HF_DATASETS_TRUST_REMOTE_CODE = True for task tests

* move warning

* pass device_map other than auto for parallelize

* add jaqket_v2 and jcommonsenseqa

* remove comments

* remove num_beams as it is incompatible with vllm

* add jnli + refactor

* rename jnla -> jnli

* add jsquad + replace colon chars with the Japanese unicode

* ignore whitespaces in generation tasks

* add marc_ja

* add xwinograd + simplify other yamls

* add mgsm and xlsum

* refactor xlsum

* add ja_leaderboard tag

* edit README.md

* update README.md

* add credit + minor changes

* run ruff format

* address review comments + add group

* remove aggregate_metric_list

* remove tags

* update tasks/README.md

* Modify label errors in catcola and paws

* Update version to 1.0 in pawsx_template_yaml

* add changelog

---------
Co-authored-by: Baber <baber@hey.com>

* support gptqmodel

* code opt

* add gptqmodel option

* Update huggingface.py

* Update pyproject.toml

* gptqmodel version upgraded to 1.0.6

* GPTQModel version upgraded to 1.0.8

* Update pyproject.toml

* fix ruff-format error

* add gptqmodel test

* Update gptqmodel test model

* skip cuda

* python3.8 compatible

* Update README.md

* Update README.md

---------
Co-authored-by: CL-ModelCloud <cl@modelcloud.ai>

* Add xquad task

* Update general README

* Run pre-commit

* fix: use lora_request for data parallel vllm evals

* fix(docs): include type hint

* chore: lint, et pre-commit al

---------
Co-authored-by: Chris Kerwell Gresla <chris@wafer.systems>

* Update pyproject.toml with watsonx package extra
Signed-off-by: kiersten-stokes <kierstenstokes@gmail.com>

* Remove unused function
Signed-off-by: kiersten-stokes <kierstenstokes@gmail.com>

---------
Signed-off-by: kiersten-stokes <kierstenstokes@gmail.com>

* add support for IBM watsonx_llm

* add ibm_watsonx_ai package to optional-dependencies

* move global scope imports to inner scope

* change cache to lru_cache

* fix circular import

* use 3.8 typing

* use 3.8 typing

---------
Co-authored-by: Baber <baber@hey.com>

* Replace generic exception classes with a more specific ones

* rerun pre-commit to pass linter tests

* Revert "rerun pre-commit to pass linter tests"

This reverts commit 67f88ccf144469853217704520e613196042d859.

* reduce repetitions in errors or so

* Replace generic exception class with a more specific one

Update prompt according to: 
https://github.com/ikergarcia1996/NoticIA/blob/main/prompts.py

* Fix Regex Pattern for CoT experiments

---------

I find out this bug by comparing the code between hendrycks_math and minerva_math.

* added tasks to spanish_bench

* fixed capitalization in escola and run pre-commit

* Update _flores_common_yaml

* Update _flores_common_yaml

* Update direct_yaml

* Update cot_yaml

* Update cot_yaml

* Update _flores_common_yaml

---------
Co-authored-by: Hailey Schoelkopf <65563625+haileyschoelkopf@users.noreply.github.com>

* Add Unitxt Multimodality Support
Signed-off-by: elronbandel <elronbandel@gmail.com>

* Update
Signed-off-by: elronbandel <elronbandel@gmail.com>

* Fix formatting
Signed-off-by: elronbandel <elronbandel@gmail.com>

---------
Signed-off-by: elronbandel <elronbandel@gmail.com>

* max_images are passed on to vllms `limit_mm_per_prompt`

* replace max image placeholders in string

* handle chat_template error

* move `fewshot_random_seed` to global

* switch conditional checks to `self.backend`

* nit

* nit

* commit feedback

* fix test; update precommit hooks

* add escape hatch for custom self.AUTO_MODEL_CLASS

* add escape hatch for custom self.AUTO_MODEL_CLASS

* fix

* move assertion

* add logging messages

* update AUTO_MODEL_CLASS behavior in _get_backend

---------
Co-authored-by: haileyschoelkopf <hailey@eleuther.ai>

* tokenizer: trust-remote-code

* pre-commit

---------
Co-authored-by: Hailey Schoelkopf <65563625+haileyschoelkopf@users.noreply.github.com>

* Fix float limit override

See: https://github.com/EleutherAI/lm-evaluation-harness/issues/2324

The float limit will be override with the previous int limit of multiple tasks are triggered together.

This PR fix this issue

* Update evaluator.py

* Update evaluator.py

* Fixing scoring bugs for smaller models

* Catching another error type in parsing

* bugfix

* pre-commit

* Add catalan_bench

* added flores_ca.yaml

* Updated some task groupings and readme

* Fix create_yamls_flores_ca.py

---------
Co-authored-by: Hailey Schoelkopf <65563625+haileyschoelkopf@users.noreply.github.com>

* Add basque_bench

* Add flores_eu group

* Update _flores_common_yaml

* Run linters, updated flores, mgsm, copa, and readme

* Apply suggestions from code review
Co-authored-by: Baber Abbasi <92168766+baberabb@users.noreply.github.com>

---------
Co-authored-by: Baber Abbasi <92168766+baberabb@users.noreply.github.com>
Co-authored-by: Hailey Schoelkopf <65563625+haileyschoelkopf@users.noreply.github.com>

* Add galician_bench

* Update xnli_gl path

* Add flores_gl group

* Update _flores_common_yaml

* Updated some task groupings and readme

---------

* Add spanish_bench

* Add flores_es group

* Update _flores_common_yaml

* Delete lm_eval/tasks/spanish_bench/escola.yaml

* Delete escola from spanish_bench.yaml

* Delete escola from README.md

* pre-commit run --all-files

* Updated some task groupings and readme

---------