| quantization_stage: | |
| sparsity_modifiers: | |
| SparseGPTModifier: | |
| sparsity: 0.5 | |
| mask_structure: '2:4' | |
| sequential_update: false | |
| sequential_targets: [LlamaDecoderLayer] | |
| targets: [Linear] | |
| ignore: ['re:.*lm_head'] | |
| block_size: 128 | |
| dampening_frac: 0.01 | |
| preserve_sparsity_mask: false | |
| offload_hessians: false | |
| quantization_modifiers: | |
| GPTQModifier: | |
| config_groups: | |
| group_0: | |
| targets: [Linear] | |
| weights: | |
| num_bits: 4 | |
| type: int | |
| symmetric: true | |
| group_size: null | |
| strategy: channel | |
| block_structure: null | |
| dynamic: false | |
| actorder: null | |
| observer: minmax | |
| observer_kwargs: {} | |
| input_activations: null | |
| output_activations: null | |
| targets: [Linear] | |
| ignore: [lm_head] | |
| sequential_update: true | |
| block_size: 128 | |
| dampening_frac: 0.01 | |
| offload_hessians: false | |