opt-125m-bnb-4bit / README.md
poedator's picture
Create README.md
d418914
|
raw
history blame
No virus
273 Bytes
facebook/OPT-125m quantized using bitsandbytes 4-bit NF4 quantization.
All license matters are set based on the underlying facebook/OPT-125m model.
This model is work in progress, use it for testing https://github.com/huggingface/transformers/pull/26037 pull request only.