Yizhen Jia
Yizhen Jia
With tons of experiments and tests, we finally support iterative dpo within a python script. Other useful features come alongside with iterative dpo: 1. Multi instance vllm inference (using ray)...
This document includes the features in LMFlow's roadmap. We welcome any discuss or contribute to the specific features at related Issues/PRs. 🤗 ### Main Features * Data * [x] DPO...
Make lots of packages optional
# Description LISA now supports multi GPU. Key points: 1. When initialize models, ds will wrap all model params with optimizer, which uses huge amount of gpu mem. However, according...