falcon-180B-chat-marlin / quantize_config.json
softmax's picture
Upload folder using huggingface_hub
ec95510 verified
raw
history blame
286 Bytes
{
"bits": 4,
"group_size": 128,
"damp_percent": 0.1,
"desc_act": false,
"static_groups": false,
"sym": true,
"true_sequential": true,
"model_name_or_path": "falcon-180B-chat-gptq",
"model_file_base_name": "model",
"is_marlin_format": true,
"quant_method": "gptq"
}