quant_stage:
  quant_modifiers:
    AWQModifier:
      config_groups:
        group_0:
          targets: [Linear]
          weights:
            num_bits: 4
            type: int
            symmetric: true
            group_size: 32
            strategy: group
            block_structure: null
            dynamic: false
            actorder: null
            observer: minmax
            observer_kwargs: {}
          input_activations: null
          output_activations: null
          format: null
      targets: [Linear]
      ignore: [lm_head, model.embed_tokens, 're:mtp.*', 're:.*input_layernorm$', 're:.*post_attention_layernorm$',
        model.norm, 're:.*mlp[.]gate$', 're:.*shared_expert_gate$', 're:.*shared_expert.*']
      mappings:
      - smooth_layer: re:.*input_layernorm$
        balance_layers: ['re:.*self_attn[.]q_proj$', 're:.*self_attn[.]k_proj$', 're:.*self_attn[.]v_proj$',
          're:.*linear_attn[.]in_proj_qkvz$', 're:.*linear_attn[.]in_proj_ba$']
      - smooth_layer: re:.*self_attn[.]v_proj$
        balance_layers: ['re:.*self_attn[.]o_proj$']
      - smooth_layer: re:.*post_attention_layernorm$
        balance_layers: ['re:.*gate_proj$', 're:.*up_proj$']
      - smooth_layer: re:.*up_proj$
        balance_layers: ['re:.*down_proj$']
      duo_scaling: true