nm-testing/kv_cache_gptq_tinyllama-e2e

1quant_stage:2  quant_modifiers:3    GPTQModifier:4      config_groups:5        group_0:6          targets: [Linear]7          weights:8            num_bits: 49            type: int10            symmetric: true11            group_size: null12            strategy: channel13            block_structure: null14            dynamic: false15            actorder: null16            scale_dtype: null17            zp_dtype: null18            observer: memoryless_minmax19            observer_kwargs: {}20          input_activations: null21          output_activations: null22          format: null23      targets: [Linear]24      ignore: [lm_head]25      kv_cache_scheme:26        num_bits: 827        type: float28        symmetric: true29        group_size: null30        strategy: tensor31        block_structure: null32        dynamic: false33        actorder: null34        scale_dtype: null35        zp_dtype: null36        observer: null37        observer_kwargs: {}38      bypass_divisibility_checks: false39      requires_calibration_data: true40      block_size: 12841      dampening_frac: 0.0142      batched_quantization: auto43