gemma3-e1b-sliced-4bit / generation_config.json
varshu23's picture
Add 4-bit quantized 2GB model
be502d6 verified
{
"_from_model_config": true,
"bos_token_id": 2,
"eos_token_id": 1,
"pad_token_id": 0,
"transformers_version": "4.57.6"
}