Llama-2-7b-chat-quantized.w8a8/recipe.yaml

quant_stage:
  quant_modifiers:
    GPTQModifier:
      sequential_update: false
      dampening_frac: 1
      ignore: [lm_head]
      config_groups:
        group_0:
          targets: [Linear]
          weights: {num_bits: 8, type: int, symmetric: true, strategy: channel}
          input_activations: {num_bits: 8, type: int, symmetric: true, dynamic: true, strategy: token}