clefourrier's picture
clefourrier HF staff
Update benchmark count and fix typo (`inetuning->finetuning`) (#395)
7abc6a7
raw history blame
No virus
14.1 kB
from src.display.utils import ModelType
TITLE = """<h1 align="center" id="space-title">🤗 Open LLM Leaderboard</h1>"""
INTRODUCTION_TEXT = """
📐 The 🤗 Open LLM Leaderboard aims to track, rank and evaluate open LLMs and chatbots.
🤗 Submit a model for automated evaluation on the 🤗 GPU cluster on the "Submit" page!
The leaderboard's backend runs the great [Eleuther AI Language Model Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness) - read more details in the "About" page!
"""
LLM_BENCHMARKS_TEXT = f"""
Useful links: [FAQ](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard/discussions/179), [Community resources](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard/discussions/174), [Collection of best models](https://huggingface.co/collections/open-llm-leaderboard/llm-leaderboard-best-models-652d6c7965a4619fb5c27a03).
# Context
With the plethora of large language models (LLMs) and chatbots being released week upon week, often with grandiose claims of their performance, it can be hard to filter out the genuine progress that is being made by the open-source community and which model is the current state of the art.
## Icons
- {ModelType.PT.to_str(" : ")} model: new, base models, trained on a given corpora
- {ModelType.FT.to_str(" : ")} model: pretrained models finetuned on more data
Specific fine-tune subcategories (more adapted to chat):
- {ModelType.IFT.to_str(" : ")} model: instruction fine-tunes, which are model fine-tuned specifically on datasets of task instruction
- {ModelType.RL.to_str(" : ")} model: reinforcement fine-tunes, which usually change the model loss a bit with an added policy.
If there is no icon, we have not uploaded the information on the model yet, feel free to open an issue with the model information!
"Flagged" indicates that this model has been flagged by the community, and should probably be ignored! Clicking the link will redirect you to the discussion about the model.
(For ex, the model was trained on the evaluation data, and is therefore cheating on the leaderboard.)
## How it works
📈 We evaluate models on 7 key benchmarks using the <a href="https://github.com/EleutherAI/lm-evaluation-harness" target="_blank"> Eleuther AI Language Model Evaluation Harness </a>, a unified framework to test generative language models on a large number of different evaluation tasks.
- <a href="https://arxiv.org/abs/1803.05457" target="_blank"> AI2 Reasoning Challenge </a> (25-shot) - a set of grade-school science questions.
- <a href="https://arxiv.org/abs/1905.07830" target="_blank"> HellaSwag </a> (10-shot) - a test of commonsense inference, which is easy for humans (~95%) but challenging for SOTA models.
- <a href="https://arxiv.org/abs/2009.03300" target="_blank"> MMLU </a> (5-shot) - a test to measure a text model's multitask accuracy. The test covers 57 tasks including elementary mathematics, US history, computer science, law, and more.
- <a href="https://arxiv.org/abs/2109.07958" target="_blank"> TruthfulQA </a> (0-shot) - a test to measure a model's propensity to reproduce falsehoods commonly found online. Note: TruthfulQA in the Harness is actually a minima a 6-shots task, as it is prepended by 6 examples systematically, even when launched using 0 for the number of few-shot examples.
- <a href="https://arxiv.org/abs/1907.10641" target="_blank"> Winogrande </a> (5-shot) - an adversarial and difficult Winograd benchmark at scale, for commonsense reasoning.
- <a href="https://arxiv.org/abs/2110.14168" target="_blank"> GSM8k </a> (5-shot) - diverse grade school math word problems to measure a model's ability to solve multi-step mathematical reasoning problems.
- <a href="https://arxiv.org/abs/1903.00161" target="_blank"> DROP </a> (3-shot) - English reading comprehension benchmark requiring Discrete Reasoning Over the content of Paragraphs.
For all these evaluations, a higher score is a better score.
We chose these benchmarks as they test a variety of reasoning and general knowledge across a wide variety of fields in 0-shot and few-shot settings.
## Details and logs
You can find:
- detailed numerical results in the `results` Hugging Face dataset: https://huggingface.co/datasets/open-llm-leaderboard/results
- details on the input/outputs for the models in the `details` of each model, that you can access by clicking the 📄 emoji after the model name
- community queries and running status in the `requests` Hugging Face dataset: https://huggingface.co/datasets/open-llm-leaderboard/requests
## Reproducibility
To reproduce our results, here is the commands you can run, using [this version](https://github.com/EleutherAI/lm-evaluation-harness/tree/b281b0921b636bc36ad05c0b0b0763bd6dd43463) of the Eleuther AI Harness:
`python main.py --model=hf-causal --model_args="pretrained=<your_model>,use_accelerate=True,revision=<your_model_revision>"`
` --tasks=<task_list> --num_fewshot=<n_few_shot> --batch_size=2 --output_path=<output_path>`
The total batch size we get for models which fit on one A100 node is 16 (8 GPUs * 2). If you don't use parallelism, adapt your batch size to fit.
*You can expect results to vary slightly for different batch sizes because of padding.*
The tasks and few shots parameters are:
- ARC: 25-shot, *arc-challenge* (`acc_norm`)
- HellaSwag: 10-shot, *hellaswag* (`acc_norm`)
- TruthfulQA: 0-shot, *truthfulqa-mc* (`mc2`)
- MMLU: 5-shot, *hendrycksTest-abstract_algebra,hendrycksTest-anatomy,hendrycksTest-astronomy,hendrycksTest-business_ethics,hendrycksTest-clinical_knowledge,hendrycksTest-college_biology,hendrycksTest-college_chemistry,hendrycksTest-college_computer_science,hendrycksTest-college_mathematics,hendrycksTest-college_medicine,hendrycksTest-college_physics,hendrycksTest-computer_security,hendrycksTest-conceptual_physics,hendrycksTest-econometrics,hendrycksTest-electrical_engineering,hendrycksTest-elementary_mathematics,hendrycksTest-formal_logic,hendrycksTest-global_facts,hendrycksTest-high_school_biology,hendrycksTest-high_school_chemistry,hendrycksTest-high_school_computer_science,hendrycksTest-high_school_european_history,hendrycksTest-high_school_geography,hendrycksTest-high_school_government_and_politics,hendrycksTest-high_school_macroeconomics,hendrycksTest-high_school_mathematics,hendrycksTest-high_school_microeconomics,hendrycksTest-high_school_physics,hendrycksTest-high_school_psychology,hendrycksTest-high_school_statistics,hendrycksTest-high_school_us_history,hendrycksTest-high_school_world_history,hendrycksTest-human_aging,hendrycksTest-human_sexuality,hendrycksTest-international_law,hendrycksTest-jurisprudence,hendrycksTest-logical_fallacies,hendrycksTest-machine_learning,hendrycksTest-management,hendrycksTest-marketing,hendrycksTest-medical_genetics,hendrycksTest-miscellaneous,hendrycksTest-moral_disputes,hendrycksTest-moral_scenarios,hendrycksTest-nutrition,hendrycksTest-philosophy,hendrycksTest-prehistory,hendrycksTest-professional_accounting,hendrycksTest-professional_law,hendrycksTest-professional_medicine,hendrycksTest-professional_psychology,hendrycksTest-public_relations,hendrycksTest-security_studies,hendrycksTest-sociology,hendrycksTest-us_foreign_policy,hendrycksTest-virology,hendrycksTest-world_religions* (average of all the results `acc`)
- Winogrande: 5-shot, *winogrande* (`acc`)
- GSM8k: 5-shot, *gsm8k* (`acc`)
- DROP: 3-shot, *drop* (`f1`)
Side note on the baseline scores:
- for log-likelihood evaluation, we select the random baseline
- for DROP, we select the best submission score according to [their leaderboard](https://leaderboard.allenai.org/drop/submissions/public) when the paper came out (NAQANet score)
- for GSM8K, we select the score obtained in the paper after finetuning a 6B model on the full GSM8K training set for 50 epochs
## Quantization
To get more information about quantization, see:
- 8 bits: [blog post](https://huggingface.co/blog/hf-bitsandbytes-integration), [paper](https://arxiv.org/abs/2208.07339)
- 4 bits: [blog post](https://huggingface.co/blog/4bit-transformers-bitsandbytes), [paper](https://arxiv.org/abs/2305.14314)
"""
EVALUATION_QUEUE_TEXT = """
# Evaluation Queue for the 🤗 Open LLM Leaderboard
Models added here will be automatically evaluated on the 🤗 cluster.
## Some good practices before submitting a model
### 1) Make sure you can load your model and tokenizer using AutoClasses:
```python
from transformers import AutoConfig, AutoModel, AutoTokenizer
config = AutoConfig.from_pretrained("your model name", revision=revision)
model = AutoModel.from_pretrained("your model name", revision=revision)
tokenizer = AutoTokenizer.from_pretrained("your model name", revision=revision)
```
If this step fails, follow the error messages to debug your model before submitting it. It's likely your model has been improperly uploaded.
Note: make sure your model is public!
Note: if your model needs `use_remote_code=True`, we do not support this option yet but we are working on adding it, stay posted!
### 2) Convert your model weights to [safetensors](https://huggingface.co/docs/safetensors/index)
It's a new format for storing weights which is safer and faster to load and use. It will also allow us to add the number of parameters of your model to the `Extended Viewer`!
### 3) Make sure your model has an open license!
This is a leaderboard for Open LLMs, and we'd love for as many people as possible to know they can use your model 🤗
### 4) Fill up your model card
When we add extra information about models to the leaderboard, it will be automatically taken from the model card
## In case of model failure
If your model is displayed in the `FAILED` category, its execution stopped.
Make sure you have followed the above steps first.
If everything is done, check you can launch the EleutherAIHarness on your model locally, using the above command without modifications (you can add `--limit` to limit the number of examples per task).
"""
CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results"
CITATION_BUTTON_TEXT = r"""
@misc{open-llm-leaderboard,
author = {Edward Beeching and Clémentine Fourrier and Nathan Habib and Sheon Han and Nathan Lambert and Nazneen Rajani and Omar Sanseviero and Lewis Tunstall and Thomas Wolf},
title = {Open LLM Leaderboard},
year = {2023},
publisher = {Hugging Face},
howpublished = "\url{https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard}"
}
@software{eval-harness,
author = {Gao, Leo and
Tow, Jonathan and
Biderman, Stella and
Black, Sid and
DiPofi, Anthony and
Foster, Charles and
Golding, Laurence and
Hsu, Jeffrey and
McDonell, Kyle and
Muennighoff, Niklas and
Phang, Jason and
Reynolds, Laria and
Tang, Eric and
Thite, Anish and
Wang, Ben and
Wang, Kevin and
Zou, Andy},
title = {A framework for few-shot language model evaluation},
month = sep,
year = 2021,
publisher = {Zenodo},
version = {v0.0.1},
doi = {10.5281/zenodo.5371628},
url = {https://doi.org/10.5281/zenodo.5371628}
}
@misc{clark2018think,
title={Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge},
author={Peter Clark and Isaac Cowhey and Oren Etzioni and Tushar Khot and Ashish Sabharwal and Carissa Schoenick and Oyvind Tafjord},
year={2018},
eprint={1803.05457},
archivePrefix={arXiv},
primaryClass={cs.AI}
}
@misc{zellers2019hellaswag,
title={HellaSwag: Can a Machine Really Finish Your Sentence?},
author={Rowan Zellers and Ari Holtzman and Yonatan Bisk and Ali Farhadi and Yejin Choi},
year={2019},
eprint={1905.07830},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
@misc{hendrycks2021measuring,
title={Measuring Massive Multitask Language Understanding},
author={Dan Hendrycks and Collin Burns and Steven Basart and Andy Zou and Mantas Mazeika and Dawn Song and Jacob Steinhardt},
year={2021},
eprint={2009.03300},
archivePrefix={arXiv},
primaryClass={cs.CY}
}
@misc{lin2022truthfulqa,
title={TruthfulQA: Measuring How Models Mimic Human Falsehoods},
author={Stephanie Lin and Jacob Hilton and Owain Evans},
year={2022},
eprint={2109.07958},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
@misc{DBLP:journals/corr/abs-1907-10641,
title={{WINOGRANDE:} An Adversarial Winograd Schema Challenge at Scale},
author={Keisuke Sakaguchi and Ronan Le Bras and Chandra Bhagavatula and Yejin Choi},
year={2019},
eprint={1907.10641},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
@misc{DBLP:journals/corr/abs-2110-14168,
title={Training Verifiers to Solve Math Word Problems},
author={Karl Cobbe and
Vineet Kosaraju and
Mohammad Bavarian and
Mark Chen and
Heewoo Jun and
Lukasz Kaiser and
Matthias Plappert and
Jerry Tworek and
Jacob Hilton and
Reiichiro Nakano and
Christopher Hesse and
John Schulman},
year={2021},
eprint={2110.14168},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
@misc{DBLP:journals/corr/abs-1903-00161,
title={{DROP:} {A} Reading Comprehension Benchmark Requiring Discrete Reasoning
Over Paragraphs},
author={Dheeru Dua and
Yizhong Wang and
Pradeep Dasigi and
Gabriel Stanovsky and
Sameer Singh and
Matt Gardner},
year={2019},
eprinttype={arXiv},
eprint={1903.00161},
primaryClass={cs.CL}
}"""