recipe.yaml

quant_stage:
  quant_modifiers:
    SmoothQuantModifier:
      smoothing_strength: 0.9
      mappings:
      - - ['re:.*q_proj', 're:.*k_proj', 're:.*v_proj']
        - re:.*input_layernorm
      ignore: []
    GPTQModifier:
      config_groups:
        group_0:
          targets: [Linear]
          weights:
            num_bits: 8
            type: int
            symmetric: true
            group_size: null
            strategy: channel
            block_structure: null
            dynamic: false
            actorder: null
            scale_dtype: null
            zp_dtype: null
            observer: mse
            observer_kwargs: {}
          input_activations:
            num_bits: 8
            type: int
            symmetric: true
            group_size: null
            strategy: token
            block_structure: null
            dynamic: true
            actorder: null
            scale_dtype: null
            zp_dtype: null
            observer: null
            observer_kwargs: {}
          output_activations: null
          format: null
      targets: [Linear]
      ignore: [lm_head, 're:.*mlp.gate$', 're:.*mlp.shared_expert_gate$']
      block_size: 128
      dampening_frac: 0.1
      actorder: static
      offload_hessians: false