default_stage:
  default_modifiers:
    AWQModifier:
      targets: ['re:.*gate_proj$', 're:.*up_proj$', 're:.*down_proj$']
      ignore: [lm_head]
      scheme: W4A16
      mappings:
      - smooth_layer: re:.*post_attention_layernorm$
        balance_layers: ['re:.*gate_proj$', 're:.*up_proj$']
      - smooth_layer: re:.*up_proj$
        balance_layers: ['re:.*down_proj$']
      duo_scaling: both
      n_grid: 40
    GPTQModifier:
      targets: ['re:.*q_proj$', 're:.*k_proj$', 're:.*v_proj$', 're:.*o_proj$']
      ignore: [embed_tokens, lm_head]
      scheme: W4A16
      kv_cache_scheme:
        num_bits: 8
        type: float
        symmetric: true
        group_size: null
        strategy: tensor
        block_structure: null
        dynamic: false
        actorder: null
        scale_dtype: null
        zp_dtype: null
        observer: minmax
        observer_kwargs: {}
      block_size: 128
      dampening_frac: 0.01
      actorder: static
      offload_hessians: false