Llama-2-70B-Chat-GPTQ / quantize_config.json
x0001's picture
Duplicate from localmodels/LLM
0cbd8a0
raw
history blame
183 Bytes
{
"bits": 4,
"group_size": -1,
"damp_percent": 0.01,
"desc_act": true,
"sym": true,
"true_sequential": true,
"model_name_or_path": null,
"model_file_base_name": null
}