mossvl_fp8_dynamic_report.json

{
  "source_model": "/inspire/qb-ilm/project/video-understanding/public/train/moss_vl_streaming/8B/final_release/MOSS-VL-0708-Instruct",
  "output_dir": "/inspire/qb-ilm/project/video-understanding/public/train/moss_vl_streaming/8B/quant/MOSS-VL-0708-Instruct-FP8-Dynamic",
  "format": "compressed-tensors",
  "scheme": "FP8_DYNAMIC",
  "weights": "FP8 channel-wise static",
  "input_activations": "FP8 per-token dynamic",
  "quantized_linear_count": 252,
  "quantized_layers": [
    0,
    1,
    3,
    4,
    5,
    7,
    8,
    9,
    11,
    12,
    13,
    15,
    16,
    17,
    19,
    20,
    21,
    23,
    24,
    25,
    27,
    28,
    29,
    31,
    32,
    33,
    35,
    36,
    37,
    39,
    40,
    41,
    43,
    44,
    45,
    47
  ],
  "excluded_cross_attention_layers": [
    2,
    6,
    10,
    14,
    18,
    22,
    26,
    30,
    34,
    38,
    42,
    46
  ],
  "elapsed_seconds": 55.73,
  "torch": "2.10.0+cu128"
}