[GH-ISSUE #2762] Quip# support for better quantization and quality. #1663

Open
opened 2026-04-12 11:38:08 -05:00 by GiteaMirror · 0 comments
Owner

Originally created by @Mayorc1978 on GitHub (Feb 26, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2762

Quip# quantization methods are proving a drastic improvement in quantizing the models while retaining the quality of the model, especially with Quip# 3bit quantization getting better results than Theoretical Lossless (FP16) 4 Bit.

I think supporting these new quantization methods would benefit a lot the Ollama community.

Originally created by @Mayorc1978 on GitHub (Feb 26, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2762 Quip# quantization methods are proving a drastic improvement in quantizing the models while retaining the quality of the model, especially with **Quip# 3bit quantization getting better results than Theoretical Lossless (FP16) 4 Bit**. I think supporting these new quantization methods would benefit a lot the Ollama community.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#1663