gemma-2b-bnb-4bit / README.md
evaldas-leliuga's picture
Added model
95788d1 verified
|
raw
history blame
616 Bytes
metadata
base_model: google/gemma-2b
license: other
license_name: gemma-terms-of-use
license_link: https://ai.google.dev/gemma/terms
inference: false
model_creator: Google
model_name: gemma-2b
quantized_by: Leliuga
pipeline_tag: text-generation
tags:
  - gemma
  - arxiv:1910.09700

gemma-2b - bnb 4bit

Description

This model is 4bit quantized version of gemma-2b using bitsandbytes. It's designed for fine-tuning! The PAD token is set as "".