baseline-gpt2-quantized / quantize_config.json
Luca-Engel's picture
AutoGPTQ model for gpt2-dpo: 8bits, gr128, desc_act=False
63a7a7a verified
raw
history blame contribute delete
266 Bytes
{
"bits": 8,
"group_size": 128,
"damp_percent": 0.01,
"desc_act": false,
"static_groups": false,
"sym": true,
"true_sequential": true,
"model_name_or_path": null,
"model_file_base_name": null,
"is_marlin_format": false,
"quant_method": "gptq"
}