TinyLlama-1.1B-Chat-v0.3-GPTQ/quantize_config.json

{
  "bits": 4,
  "group_size": 128,
  "damp_percent": 0.1,
  "desc_act": true,
  "sym": true,
  "true_sequential": true,
  "model_name_or_path": null,
  "model_file_base_name": "model"
}