GeneZC/MiniMA-2-1B

Model

MiniMA-2-1B

2

9 commits

1 linked in READMEs

updated Feb 10, 2025

See the code

README

MiniMA-2-1B

πŸ“‘ arXiv | πŸ‘» GitHub | πŸ€— HuggingFace-MiniMA-3B | πŸ€— HuggingFace-MiniChat-3B | πŸ€– ModelScope-MiniMA-3B | πŸ€– ModelScope-MiniChat-3B | πŸ€— HuggingFace-MiniChat-1.5-3B | πŸ€— HuggingFace-MiniMA-2-3B | πŸ€— HuggingFace-MiniChat-2-3B | πŸ€— HuggingFace-MiniMA-2-1B | πŸ€— HuggingFace-MiniLoong-3B | πŸ€— HuggingFace-MiniMix-2/4x3B

❗ Must comply with LICENSE of LLaMA-2 since it is derived from LLaMA-2.

teaser_a

Standard Benchmarks

MethodTFLOPsMMLU (5-shot)CEval (5-shot)DROP (3-shot)HumanEval (0-shot)BBH (3-shot)GSM8K (8-shot)
Mamba-2.8B4.6E925.5824.7415.727.3229.373.49
ShearedLLaMA-2.7B0.8E926.9722.8819.984.8830.483.56
BTLM-3B11.3E927.2026.0017.8410.9830.874.55
StableLM-3B72.0E944.7531.0522.3515.8532.5910.99
Qwen-1.8B23.8E944.0554.7512.9714.0230.8022.97
Phi-2-2.8B159.9E956.7434.0330.7446.9544.1355.42
LLaMA-2-7B84.0E946.0034.4031.5712.8032.0214.10
MiniMA-3B4.0E928.5128.2322.5010.9831.618.11
MiniMA-2-1B6.3E931.3434.9220.0810.3731.167.28
MiniMA-2-3B13.4E940.1444.6523.1014.6331.438.87
MiniMix-2/4x3B25.4E944.3545.7733.7818.2933.6021.61
MiniChat-3B4.0E938.4036.4822.5818.2931.3629.72
MiniChat-2-3B13.4E946.1743.9130.2622.5634.9538.13

Bibtex

@article{zhang2023law,
    title={Towards the Law of Capacity Gap in Distilling Language Models},
    author={Zhang, Chen and Song, Dawei and Ye, Zheyu and Gao, Yan},
    year={2023},
    url={https://arxiv.org/abs/2311.07052}
}
endpoints_compatible
llama
pytorch
safetensors
text-generation
text-generation-inference
transformers

Contributors

GeneZC

8 commits

SFconvertbot

1 commits

GeneZC/MiniMA-2-1B

Model

MiniMA-2-1B

2

9 commits

1 linked in READMEs

updated Feb 10, 2025

See the code

README

MiniMA-2-1B

πŸ“‘ arXiv | πŸ‘» GitHub | πŸ€— HuggingFace-MiniMA-3B | πŸ€— HuggingFace-MiniChat-3B | πŸ€– ModelScope-MiniMA-3B | πŸ€– ModelScope-MiniChat-3B | πŸ€— HuggingFace-MiniChat-1.5-3B | πŸ€— HuggingFace-MiniMA-2-3B | πŸ€— HuggingFace-MiniChat-2-3B | πŸ€— HuggingFace-MiniMA-2-1B | πŸ€— HuggingFace-MiniLoong-3B | πŸ€— HuggingFace-MiniMix-2/4x3B

❗ Must comply with LICENSE of LLaMA-2 since it is derived from LLaMA-2.

teaser_a

Standard Benchmarks

MethodTFLOPsMMLU (5-shot)CEval (5-shot)DROP (3-shot)HumanEval (0-shot)BBH (3-shot)GSM8K (8-shot)
Mamba-2.8B4.6E925.5824.7415.727.3229.373.49
ShearedLLaMA-2.7B0.8E926.9722.8819.984.8830.483.56
BTLM-3B11.3E927.2026.0017.8410.9830.874.55
StableLM-3B72.0E944.7531.0522.3515.8532.5910.99
Qwen-1.8B23.8E944.0554.7512.9714.0230.8022.97
Phi-2-2.8B159.9E956.7434.0330.7446.9544.1355.42
LLaMA-2-7B84.0E946.0034.4031.5712.8032.0214.10
MiniMA-3B4.0E928.5128.2322.5010.9831.618.11
MiniMA-2-1B6.3E931.3434.9220.0810.3731.167.28
MiniMA-2-3B13.4E940.1444.6523.1014.6331.438.87
MiniMix-2/4x3B25.4E944.3545.7733.7818.2933.6021.61
MiniChat-3B4.0E938.4036.4822.5818.2931.3629.72
MiniChat-2-3B13.4E946.1743.9130.2622.5634.9538.13

Bibtex

@article{zhang2023law,
    title={Towards the Law of Capacity Gap in Distilling Language Models},
    author={Zhang, Chen and Song, Dawei and Ye, Zheyu and Gao, Yan},
    year={2023},
    url={https://arxiv.org/abs/2311.07052}
}
endpoints_compatible
llama
pytorch
safetensors
text-generation
text-generation-inference
transformers

Contributors

GeneZC

8 commits

SFconvertbot

1 commits