Aryanne/Sheared-LLaMA-2.7B-gguf

Some GGUF v2 quantizations of the model princeton-nlp/Sheared-LLaMA-2.7B

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

Sheared-LLaMA-2.7B 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. We use 0.4B tokens for pruning and 50B tokens for continued pre-training the pruned model.

• 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.

Model	# Pre-training Tokens	Average Performance
LLaMA2-7B	2T	64.6

1.3B

Model	# Pre-training Tokens	Average Performance
OPT-1.3B	300B	48.2
Pythia-1.4B	300B	48.9
Sheared-LLaMA-1.3B	50B	51.0

3B

Model	# Pre-training Tokens	Average Performance
OPT-2.7B	300B	51.4
Pythia-2.8B	300B	52.5
INCITE-Base-3B	800B	54.7
Open-LLaMA-3B-v1	1T	55.1
Open-LLaMA-3B-v2	1T	55.7
Sheared-LLaMA-2.7B	50B	56.7