File size: 1,385 Bytes
0f4f1cf a437579 ed635e9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
---
license: llama2
---
---
license: llama2
---
**Paper**: [https://arxiv.org/pdf/2310.06694.pdf](https://arxiv.org/pdf/2310.06694.pdf)
**Code**: https://github.com/princeton-nlp/LLM-Shearing
**Models**: [Sheared-LLaMA-1.3B](https://huggingface.co/princeton-nlp/Sheared-LLaMA-1.3B), [Sheared-LLaMA-2.7B](https://huggingface.co/princeton-nlp/Sheared-LLaMA-2.7B)
**Pruned Models without Continued Pre-training**: [Sheared-LLaMA-1.3B-Pruned](https://huggingface.co/princeton-nlp/Sheared-LLaMA-1.3B-Pruned), [Sheared-LLaMA-2.7B-Pruned](https://huggingface.co/princeton-nlp/Sheared-LLaMA-2.7B-Pruned)
**Instruction-tuned Models**: [Sheared-LLaMA-1.3B-ShareGPT](https://huggingface.co/princeton-nlp/Sheared-LLaMA-1.3B-ShareGPT), [Sheared-LLaMA-2.7B-ShareGPT](https://huggingface.co/princeton-nlp/Sheared-LLaMA-2.7B-ShareGPT)
**License**: Must comply with license of Llama2 since it's a model derived from Llama2.
Sheared-LLaMA-2.7B-Pruned is the model pruned from [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) **without continued pre-training**.
We used roughly 0.4B tokens to perform the pruning experiment. This model could be a good use to study
- effective data mixtures for continued pre-training
- comparisons to other pruning techniques
- extensive evaluations to understand how pruning affects knowledge and reasoning capabilities of LLMs |