Llama-2-7b-NuGPTQ / quantize_config.json
smpanaro's picture
Initial commit
c2fa56c verified
raw
history blame
312 Bytes
{
"bits": 4,
"group_size": -1,
"damp_percent": 0.01,
"desc_act": true,
"static_groups": false,
"sym": true,
"true_sequential": true,
"model_name_or_path": "meta-llama/Llama-2-7b-hf-4bit-nuq",
"model_file_base_name": "gptq_model-4bit--1g",
"is_marlin_format": false,
"quant_method": "gptq"
}