From 93a056fe4249281107baafc2a936e34f7f59e111 Mon Sep 17 00:00:00 2001 From: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com> Date: Thu, 12 Dec 2024 16:30:45 +0100 Subject: [PATCH 1/2] Fix imports from model_config (#443) --- docs/source/using-the-python-api.mdx | 2 +- src/lighteval/main_endpoint.py | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/docs/source/using-the-python-api.mdx b/docs/source/using-the-python-api.mdx index 2e160a679..8c44050f4 100644 --- a/docs/source/using-the-python-api.mdx +++ b/docs/source/using-the-python-api.mdx @@ -11,7 +11,7 @@ After that, simply run the pipeline and save the results. ```python import lighteval from lighteval.logging.evaluation_tracker import EvaluationTracker -from lighteval.models.model_config import VLLMModelConfig +from lighteval.models.vllm.vllm_model import VLLMModelConfig from lighteval.pipeline import ParallelismManager, Pipeline, PipelineParameters from lighteval.utils.utils import EnvConfig from lighteval.utils.imports import is_accelerate_available diff --git a/src/lighteval/main_endpoint.py b/src/lighteval/main_endpoint.py index 952aae074..be75b711a 100644 --- a/src/lighteval/main_endpoint.py +++ b/src/lighteval/main_endpoint.py @@ -93,7 +93,7 @@ def openai( Evaluate OPENAI models. """ from lighteval.logging.evaluation_tracker import EvaluationTracker - from lighteval.models.model_config import OpenAIModelConfig + from lighteval.models.endpoints.openai_model import OpenAIModelConfig from lighteval.pipeline import EnvConfig, ParallelismManager, Pipeline, PipelineParameters env_config = EnvConfig(token=TOKEN, cache_dir=cache_dir) @@ -317,7 +317,7 @@ def tgi( import yaml from lighteval.logging.evaluation_tracker import EvaluationTracker - from lighteval.models.model_config import TGIModelConfig + from lighteval.models.endpoints.tgi_model import TGIModelConfig from lighteval.pipeline import EnvConfig, ParallelismManager, Pipeline, PipelineParameters env_config = EnvConfig(token=TOKEN, cache_dir=cache_dir) From 500632a097499fc7a2002ad682d605cb03072302 Mon Sep 17 00:00:00 2001 From: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com> Date: Mon, 16 Dec 2024 14:10:21 +0100 Subject: [PATCH 2/2] Fix wrong instructions and code for custom tasks (#450) * Delete wrong instruction in custom task docs * Delete wrong code for custom tasks * Delete wrong code for extended tasks * Delete wrong code for community tasks * Delete unnecessary code for community tasks --- community_tasks/_template.py | 7 ------- community_tasks/aimo_evals.py | 8 -------- community_tasks/arabic_evals.py | 4 ---- community_tasks/german_rag_evals.py | 8 -------- community_tasks/oz_evals.py | 5 ----- community_tasks/serbian_eval.py | 4 ---- docs/source/adding-a-custom-task.mdx | 11 ----------- examples/nanotron/custom_evaluation_tasks.py | 4 ---- src/lighteval/tasks/extended/ifeval/main.py | 5 ----- src/lighteval/tasks/extended/mix_eval/main.py | 5 ----- src/lighteval/tasks/extended/mt_bench/main.py | 4 ---- src/lighteval/tasks/extended/tiny_benchmarks/main.py | 8 -------- 12 files changed, 73 deletions(-) diff --git a/community_tasks/_template.py b/community_tasks/_template.py index d0099ba26..2db28e340 100644 --- a/community_tasks/_template.py +++ b/community_tasks/_template.py @@ -116,10 +116,3 @@ def __init__( sample_level_fn=lambda x: x, # how to compute score for one sample corpus_level_fn=np.mean, # aggregation ) - -# MODULE LOGIC -# You should not need to touch this -# Convert to dict for lighteval -if __name__ == "__main__": - print(t.name for t in TASKS_TABLE) - print(len(TASKS_TABLE)) diff --git a/community_tasks/aimo_evals.py b/community_tasks/aimo_evals.py index be59950bd..885ffd8da 100644 --- a/community_tasks/aimo_evals.py +++ b/community_tasks/aimo_evals.py @@ -56,11 +56,3 @@ def aimo_prompt(line, task_name: str = None): # STORE YOUR EVALS TASKS_TABLE = [task] - - -# MODULE LOGIC -# You should not need to touch this - -if __name__ == "__main__": - print(t.name for t in TASKS_TABLE) - print(len(TASKS_TABLE)) diff --git a/community_tasks/arabic_evals.py b/community_tasks/arabic_evals.py index 07a096eca..4408f22fa 100644 --- a/community_tasks/arabic_evals.py +++ b/community_tasks/arabic_evals.py @@ -856,7 +856,3 @@ def __init__( + [toxigen_ar_task] + [sciq_ar_task] ) - -if __name__ == "__main__": - print(t.name for t in TASKS_TABLE) - print(len(TASKS_TABLE)) diff --git a/community_tasks/german_rag_evals.py b/community_tasks/german_rag_evals.py index 30d6dcb4a..78af6794e 100644 --- a/community_tasks/german_rag_evals.py +++ b/community_tasks/german_rag_evals.py @@ -221,11 +221,3 @@ def prompt_fn_context_question_match(line, task_name: str = None): # STORE YOUR EVALS TASKS_TABLE = [task1, task2, task3, task4] - - -# MODULE LOGIC -# You should not need to touch this - -if __name__ == "__main__": - print(t.name for t in TASKS_TABLE) - print(len(TASKS_TABLE)) diff --git a/community_tasks/oz_evals.py b/community_tasks/oz_evals.py index 6252a20a0..4ec70e291 100644 --- a/community_tasks/oz_evals.py +++ b/community_tasks/oz_evals.py @@ -87,8 +87,3 @@ def prompt_fn_oz_eval_task(line, task_name: str = None): # STORE YOUR EVALS TASKS_TABLE = [oz_eval_task] - - -if __name__ == "__main__": - print(t["name"] for t in TASKS_TABLE) - print(len(TASKS_TABLE)) diff --git a/community_tasks/serbian_eval.py b/community_tasks/serbian_eval.py index 3b49c4cb0..d972ac69c 100644 --- a/community_tasks/serbian_eval.py +++ b/community_tasks/serbian_eval.py @@ -784,7 +784,3 @@ def create_task_config( mmlu_world_religions, mmlu_all, ] - -if __name__ == "__main__": - print(t.name for t in TASKS_TABLE) - print(len(TASKS_TABLE)) diff --git a/docs/source/adding-a-custom-task.mdx b/docs/source/adding-a-custom-task.mdx index e1823b7b9..e5160024d 100644 --- a/docs/source/adding-a-custom-task.mdx +++ b/docs/source/adding-a-custom-task.mdx @@ -167,17 +167,6 @@ TASKS_TABLE = SUBSET_TASKS # TASKS_TABLE = [task] ``` -Finally, you need to add a module logic to convert your task to a dict for lighteval. - -```python -# MODULE LOGIC -# You should not need to touch this -# Convert to dict for lighteval -if __name__ == "__main__": - print(t.name for t in TASKS_TABLE) - print(len(TASKS_TABLE)) -``` - Once your file is created you can then run the evaluation with the following command: ```bash diff --git a/examples/nanotron/custom_evaluation_tasks.py b/examples/nanotron/custom_evaluation_tasks.py index 78c354916..e4b1b5ca6 100644 --- a/examples/nanotron/custom_evaluation_tasks.py +++ b/examples/nanotron/custom_evaluation_tasks.py @@ -671,7 +671,3 @@ def __init__( "all": ",".join(t[1] for t in _TASKS_STRINGS), "early-signal": EARLY_SIGNAL_TASKS, } - -if __name__ == "__main__": - print(t["name"] for t in TASKS_TABLE) - print(len(TASKS_TABLE)) diff --git a/src/lighteval/tasks/extended/ifeval/main.py b/src/lighteval/tasks/extended/ifeval/main.py index e6947bb6e..60d1be5fa 100644 --- a/src/lighteval/tasks/extended/ifeval/main.py +++ b/src/lighteval/tasks/extended/ifeval/main.py @@ -160,8 +160,3 @@ def agg_inst_level_acc(items): TASKS_TABLE = [ifeval] extend_enum(Metrics, "ifeval_metric", ifeval_metrics) - -if __name__ == "__main__": - # Adds the metric to the metric list! - print(t["name"] for t in TASKS_TABLE) - print(len(TASKS_TABLE)) diff --git a/src/lighteval/tasks/extended/mix_eval/main.py b/src/lighteval/tasks/extended/mix_eval/main.py index 15d7490bc..8684e910c 100644 --- a/src/lighteval/tasks/extended/mix_eval/main.py +++ b/src/lighteval/tasks/extended/mix_eval/main.py @@ -228,8 +228,3 @@ def mean_dv_5(x): TASKS_TABLE = [mixeval_multichoice_easy, mixeval_freeform_easy, mixeval_multichoice_hard, mixeval_freeform_hard] - -if __name__ == "__main__": - # Adds the metric to the metric list! - print(t["name"] for t in TASKS_TABLE) - print(len(TASKS_TABLE)) diff --git a/src/lighteval/tasks/extended/mt_bench/main.py b/src/lighteval/tasks/extended/mt_bench/main.py index e5b209982..117e363dd 100644 --- a/src/lighteval/tasks/extended/mt_bench/main.py +++ b/src/lighteval/tasks/extended/mt_bench/main.py @@ -95,7 +95,3 @@ def flow_judge_mt_bench_prompt(question, answer, options, gold): TASKS_TABLE = [task] - -if __name__ == "__main__": - print(t["name"] for t in TASKS_TABLE) - print(len(TASKS_TABLE)) diff --git a/src/lighteval/tasks/extended/tiny_benchmarks/main.py b/src/lighteval/tasks/extended/tiny_benchmarks/main.py index b283921f2..fae6e89df 100644 --- a/src/lighteval/tasks/extended/tiny_benchmarks/main.py +++ b/src/lighteval/tasks/extended/tiny_benchmarks/main.py @@ -283,11 +283,3 @@ def aggregate(self, y_input): corpus_level_fn=TinyCorpusAggregator(name).aggregate, ), ) - - -# MODULE LOGIC -# You should not need to touch this -# Convert to dict for lighteval -if __name__ == "__main__": - print(t["name"] for t in TASKS_TABLE) - print(len(TASKS_TABLE))