recipe.yaml

quant_stage:
  quant_modifiers:
    GPTQModifier:
      config_groups:
        group_0:
          targets: [Linear]
          weights:
            num_bits: 4
            type: int
            symmetric: false
            group_size: 128
            strategy: group
            block_structure: null
            dynamic: false
            actorder: static
            observer: mse
            observer_kwargs: {}
          input_activations: null
          output_activations: null
          format: null
      targets: [Linear]
      ignore: [lm_head]
      sequential_update: true
      block_size: 128
      dampening_frac: 0.1
      offload_hessians: false