recipe.yaml

quant_stage:
  quant_modifiers:
    QuantizationModifier:
      config_groups:
        group_0:
          targets: ['re:.*\.down_proj$']
          weights:
            num_bits: 8
            type: float
            symmetric: true
            group_size: null
            strategy: channel
            block_structure: null
            dynamic: false
            actorder: null
            scale_dtype: null
            zp_dtype: null
            observer: memoryless_minmax
            observer_kwargs: {}
          input_activations:
            num_bits: 8
            type: float
            symmetric: true
            group_size: null
            strategy: token
            block_structure: null
            dynamic: true
            actorder: null
            scale_dtype: null
            zp_dtype: null
            observer: null
            observer_kwargs: {}
          output_activations: null
          format: null
        group_1:
          targets: ['re:.*self_attn\.(k|o|q|v)_proj$', 're:.*\.(gate|up)_proj$']
          weights:
            num_bits: 4
            type: float
            symmetric: true
            group_size: 16
            strategy: tensor_group
            block_structure: null
            dynamic: false
            actorder: null
            scale_dtype: torch.float8_e4m3fn
            zp_dtype: null
            observer: memoryless_minmax
            observer_kwargs: {}
          input_activations:
            num_bits: 4
            type: float
            symmetric: true
            group_size: 16
            strategy: tensor_group
            block_structure: null
            dynamic: local
            actorder: null
            scale_dtype: null
            zp_dtype: null
            observer: minmax
            observer_kwargs: {}
          output_activations: null
          format: null
      targets: [Linear]
      ignore: [lm_head]
      bypass_divisibility_checks: false