| { | |
| "bits": 4, | |
| "group_size": 128, | |
| "sym": true, | |
| "data_type": "int", | |
| "enable_quanted_input": true, | |
| "enable_minmax_tuning": true, | |
| "seqlen": 2048, | |
| "batch_size": 8, | |
| "scale_dtype": "torch.float16", | |
| "lr": 0.001, | |
| "minmax_lr": 0.001, | |
| "gradient_accumulate_steps": 1, | |
| "iters": 1000, | |
| "amp": true, | |
| "nsamples": 512, | |
| "low_gpu_mem_usage": false, | |
| "enable_norm_bias_tuning": false, | |
| "autoround_version": "0.4.4", | |
| "block_name_to_quantize": "model.transformer.blocks", | |
| "quant_method": "gptq", | |
| "desc_act": false, | |
| "true_sequential": false, | |
| "damp_percent": 0.01 | |
| } |