{
  "created_at": "2026-07-26T06:19:47.245758+00:00",
  "source": "/inspire/qb-ilm/project/video-understanding/public/train/moss_vl_streaming/8B/final_release/mossvl_streaming_final_release",
  "output": "/inspire/qb-ilm/project/video-understanding/public/train/moss_vl_streaming/8B/quant/mossvl_streaming_w4a16_nf4_keep_first4_last4",
  "variant": "w4_keep_ends",
  "method": "bitsandbytes NF4 W4A16 keep-first-last-4",
  "compute_dtype": "bfloat16",
  "skip_modules": [
    "model.visual",
    "cross_attn",
    "lm_head",
    "model.language_model.layers.0",
    "model.language_model.layers.1",
    "model.language_model.layers.2",
    "model.language_model.layers.3",
    "model.language_model.layers.44",
    "model.language_model.layers.45",
    "model.language_model.layers.46",
    "model.language_model.layers.47"
  ],
  "quantized_kind": "4bit",
  "quantized_linear_count": 240,
  "quantized_module_samples": [
    "model.language_model.layers.4.self_attn.q_proj",
    "model.language_model.layers.4.self_attn.k_proj",
    "model.language_model.layers.4.self_attn.v_proj",
    "model.language_model.layers.4.self_attn.o_proj",
    "model.language_model.layers.4.mlp.gate_proj",
    "model.language_model.layers.4.mlp.up_proj",
    "model.language_model.layers.4.mlp.down_proj",
    "model.language_model.layers.5.self_attn.q_proj",
    "model.language_model.layers.5.self_attn.k_proj",
    "model.language_model.layers.5.self_attn.v_proj",
    "model.language_model.layers.5.self_attn.o_proj",
    "model.language_model.layers.5.mlp.gate_proj",
    "model.language_model.layers.5.mlp.up_proj",
    "model.language_model.layers.5.mlp.down_proj",
    "model.language_model.layers.6.mlp.gate_proj",
    "model.language_model.layers.6.mlp.up_proj",
    "model.language_model.layers.6.mlp.down_proj",
    "model.language_model.layers.7.self_attn.q_proj",
    "model.language_model.layers.7.self_attn.k_proj",
    "model.language_model.layers.7.self_attn.v_proj"
  ],
  "peak_gpu_memory_gib_during_conversion": 11.038,
  "gpu": "NVIDIA H200",
  "versions": {
    "torch": "2.8.0+cu128",
    "transformers": "4.57.1",
    "accelerate": "1.12.0",
    "bitsandbytes": "0.49.2"
  },
  "reload_verified": true,
  "reload_quantized_kind": "4bit",
  "reload_quantized_linear_count": 240,
  "peak_gpu_memory_gib_after_reload": 14.549
}