| sparsity_stage: | |
| sparsity_modifiers: | |
| SparseGPTModifier: {sparsity: 0.5, mask_structure: '2:4', sequential_update: false} | |
| run_type: &id001 !!python/object/apply:llmcompressor.recipe.stage.StageRunType [oneshot] | |
| quantization_stage: | |
| quantization_modifiers: | |
| ConstantPruningModifier: | |
| targets: ['re:.*q_proj.weight', 're:.*k_proj.weight', 're:.*v_proj.weight', 're:.*o_proj.weight', | |
| 're:.*gate_proj.weight', 're:.*up_proj.weight', 're:.*down_proj.weight'] | |
| start: 0 | |
| QuantizationModifier: | |
| targets: [Linear] | |
| ignore: [lm_head] | |
| scheme: FP8_DYNAMIC | |
| run_type: | |