{
    "model_id": "qwen3_vl_moe",
    "init_from_hf_path": "./ckpt/hf_path/Qwen3-VL-30B-Instruct",
    "image_encoder": {
        "vision_encoder": {
            "model_id": "qwen3vit",
            "num_layers": 27,
            "hidden_size": 1152,
            "num_attention_heads": 16,
            "freeze": true,
            "attn_implementation": "sdpa"
        },
        "vision_projector": {
            "model_id": "lnmlp",
            "num_layers": 1,
            "freeze": true
        }
    },
    "text_decoder": {
        "model_id": "qwen3lm",
        "num_layers": 48,
        "hidden_size": 2048,
        "num_attention_heads": 32,
        "max_position_embeddings": 262144,
        "freeze": false,
        "use_npu_fused_moe": false,
        "attn_implementation": "sdpa",
        "activation_offload": false
    },
    "loss_cfg": {
        "compute_mode": "default",
        "chunk_size": 1024
    },
    "patch": {
        "clip_grad_async": true,
        "scale_grad": true
    }
}