mirror of
https://github.com/huggingface/peft.git
synced 2025-10-20 15:33:48 +08:00
Currently, the CI fails because awq tries to import a non-existing function from transformers (presumably it was there at one point but no longer is): > from transformers.modeling_utils import shard_checkpoint > E ImportError: cannot import name 'shard_checkpoint' from 'transformers.modeling_utils' (/opt/conda/envs/peft/lib/python3.11/site-packages/transformers/modeling_utils.py) This has been fixed in awq v0.2.7. Therefore, this newer version is now used in CI.
PEFT Docker images
Here we store all PEFT Docker images used in our testing infrastructure. We use python 3.11 for now on all our images.
peft-cpu
: PEFT compiled on CPU with all other HF libraries installed on main branchpeft-gpu
: PEFT complied for NVIDIA GPUs wih all other HF libraries installed on main branchpeft-gpu-bnb-source
: PEFT complied for NVIDIA GPUs withbitsandbytes
and all other HF libraries installed from main branchpeft-gpu-bnb-latest
: PEFT complied for NVIDIA GPUs withbitsandbytes
complied from main and all other HF libraries installed from latest PyPipeft-gpu-bnb-multi-source
: PEFT complied for NVIDIA GPUs withbitsandbytes
complied frommulti-backend
branch and all other HF libraries installed from main branch
peft-gpu-bnb-source
and peft-gpu-bnb-multi-source
are essentially the same, with the only difference being bitsandbytes
compiled on another branch. Make sure to propagate the changes you applied on one file to the other!