PygmalionAI / aphrodite-engine

PygmalionAI's large-scale inference engine
https://pygmalion.chat
GNU Affero General Public License v3.0
606 stars 78 forks source link

Improve cohere model. #404

Closed sgsdxzy closed 1 month ago

sgsdxzy commented 1 month ago

Based on https://github.com/vllm-project/vllm/pull/3985 https://github.com/vllm-project/vllm/pull/3919 Fix weight loading for exl2.