{
  "source_directory": "/home/bellock/projects/internvl35-fp8/models/InternVL3_5-4B-HF",
  "output_directory": "/home/bellock/projects/internvl35-fp8/models/InternVL3_5-4B-AWQ-W4A16-G128",
  "source_revision": "model_id=OpenGVLab/InternVL3_5-4B-HF\nrevision=6bd4487402110ef9889ba50eb7aefeb302526fed",
  "quantization_scheme": "W4A16_ASYM",
  "quantization_algorithm": "AWQ",
  "weight_format": "pack-quantized",
  "weight_num_bits": 4,
  "weight_type": "int",
  "weight_strategy": "group",
  "weight_group_size": 128,
  "weight_symmetric": false,
  "input_activations": null,
  "target_module_type": "Linear (language decoder only, regex-scoped)",
  "target_linear_count": 252,
  "target_attention_linear": 144,
  "target_mlp_linear": 108,
  "protected_linear_count": 147,
  "protected_modules": [
    "vision_tower",
    "multi_modal_projector",
    "lm_head",
    "embed_tokens"
  ],
  "awq_duo_scaling": true,
  "awq_n_grid": 20,
  "calibration_dataset": "lmms-lab/flickr30k",
  "calibration_samples": 128,
  "calibration_skipped": 0,
  "calibration_image_size": "448x448",
  "calibration_image_patches_per_sample": 1,
  "calibration_sequence_length_range": "278-288",
  "source_dtype": "torch.bfloat16",
  "python_version": "3.12.3",
  "torch_version": "2.12.0+cu132",
  "transformers_version": "5.10.1",
  "llmcompressor_version": "0.12.0.1",
  "compressed_tensors_version": "0.17.1",
  "checkpoint_tensors_total": 1597,
  "checkpoint_packed_int4_modules": 252,
  "checkpoint_size_gib": 3.816
}