Paper: https://arxiv.org/pdf/2310.06694.pdf
Code: https://github.com/princeton-nlp/LLM-Shearing
Models: Sheared-LLaMA-1.3B, Sheared-LLaMA-2.7B

License: Must comply with license of Llama2 since it's a model derived from Llama2.

Sheared-LLaMA-1.3B is a model pruned and further pre-trained from meta-llama/Llama-2-7b-hf. We dynamically load data from different domains in the RedPajama dataset to prune and contune pre-train the model. We use 0.4B tokens for pruning and 50B tokens for continued pre-training the pruned model. This model can be loaded with HuggingFace via

model = AutoModelForCausalLM.from_pretrained("princeton-nlp/Sheared-LLaMA-1.3B")

• Smaller-scale
• Same vocabulary as LLaMA1 and LLaMA2
• Derived with a budget of 50B tokens by utilizing existing strong LLMs

Downstream Tasks

We evaluate on an extensive set of downstream tasks including reasoning, reading comprehension, language modeling and knowledge intensive tasks. Our Sheared-LLaMA models outperform existing large language models.

1.3B

3B

Bibtex

@article{xia2023sheared,
   title={Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning},
   author={Xia, Mengzhou and Gao, Tianyu, and Zeng, Zhiyuan and Chen, Danqi},
   year={2023}
}