issues
search
bigscience-workshop
/
petals
🌸 Run LLMs at home, BitTorrent-style. Fine-tuning and inference up to 10x faster than offloading
https://petals.dev
MIT License
8.89k
stars
490
forks
source link
Peft update
#557
Open
artek0chumak
opened
4 months ago
artek0chumak
commented
4 months ago
Make possible to use up-to-date PEFT library.
TODO:
[x] Fix initialization
[x] Fix inference without adapters
[x] Fix init for quantized modules
[x] Fix debug messages while server initialization
Make possible to use up-to-date PEFT library.
TODO: