llama3_8b_gptq_4bits / quantize_config.json
shihyangl
First model version
dd8dd6a
raw
history blame contribute delete
203 Bytes
{
"bits": 4,
"group_size": -1,
"damp_percent": 0.01,
"desc_act": true,
"sym": false,
"true_sequential": true,
"model_name_or_path": null,
"model_file_base_name": "model"
}