How long does single LLM's tunning reuqired?

OpenLLMAI / OpenRLHF

An Easy-to-use, Scalable and High-performance RLHF Framework (70B+ PPO Full Tuning & Iterative DPO & LoRA & Mixtral)

https://openrlhf.readthedocs.io/

Apache License 2.0

1.72k stars 161 forks source link

Open alphahumancoder opened 2 months ago

alphahumancoder commented 2 months ago

I mean 7B and 13B models

wuxibin89 commented 2 months ago

It depends on your datasets and GPUs, we have some benchmark for reference: https://github.com/OpenLLMAI/OpenRLHF/tree/wuxibin/benchmark/benchmark

hijkzzz commented 2 months ago

It depends on your datasets and GPUs, we have some benchmark for reference: https://github.com/OpenLLMAI/OpenRLHF/tree/wuxibin/benchmark/benchmark

This performance data in this branch is not optimized and we will further adjust the configs in the official technical report