EfficientQAT: LLM Quantization, gets a 2-bit llama2-70B outperform regular 13B

Article URL: https://old.reddit.com/r/LocalLLaMA/comments/1e5x2k4/new_llms_quantization_algorithm_efficientqat/

Comments URL: https://news.ycombinator.com/item?id=40991588

Points: 7

# Comments: 0

Read more here: External Link