config.json

{
  "_moe_implementation": "fused",
  "action_dim": 55,
  "action_fp32": false,
  "action_head_dim": 128,
  "action_num_attention_heads": 32,
  "action_num_key_value_heads": 8,
  "adanorm_time": true,
  "align_params": {
    "depth": {
      "block_future_depth_to_action": false,
      "detach_future_image_feats": true,
      "dim_head": 32,
      "dim_out": 1024,
      "ff_mult": 1,
      "input_size": 224,
      "model_type": "MoRGBD",
      "moge_path": "assets/pretrained/moge-2-vitb-normal/model.pt",
      "morgbd_path": "assets/pretrained/lingbot-vla-v2-6b/depth/model.pt",
      "num_backbone_tokens": 256,
      "num_heads": 4,
      "num_layers": 1,
      "token_size": 16,
      "use_future_depth": true
    },
    "depth_loss_weight": 0.004,
    "future_depth_loss_weight": 0.004,
    "llm": {
      "dim_out": 2560,
      "image_input_size": 224,
      "image_token_size": 8
    },
    "mode": "query",
    "num_task_tokens": 8,
    "use_future_video": true,
    "video": {
      "attention_mode": "flex_block_causal",
      "block_suffix_to_future_video": false,
      "ckpt_path": "assets/pretrained/lingbot-vla-v2-6b/dino_video/teacher_step_10000.pth",
      "cls_loss_type": "mse",
      "cls_loss_weight": 0.2,
      "cls_pool": "last",
      "config_path": "assets/pretrained/lingbot-vla-v2-6b/dino_video/config.yaml",
      "cosine_loss_weight": 0.2,
      "detach_image_feats": true,
      "dim_head": 32,
      "dim_out": 1024,
      "effective_fps": 1.0,
      "ff_mult": 1,
      "future_video_loss_weight": 0.004,
      "input_size": 256,
      "log_max_samples": 32,
      "log_scale": 16,
      "mse_loss_weight": 1.0,
      "n_blocks": 1,
      "num_backbone_tokens": 256,
      "num_future_frames": 1,
      "num_heads": 4,
      "num_layers": 1,
      "share_future_depth_query": true,
      "use_cls_loss": false,
      "use_cosine_loss": true,
      "use_current_patch_loss": true,
      "use_current_shared_task_proj": true,
      "use_mse_loss": true,
      "use_patch_loss": true,
      "use_shared_future_task_proj": true,
      "use_smooth_l1_loss": false,
      "use_warmup_frame": true
    },
    "visual_dir": "outputs/images",
    "visual_steps": 5000
  },
  "architectures": [
    "LingbotVlaV2Policy"
  ],
  "attention_implementation": "flex_cached",
  "bias_update_speed": 0.0,
  "chunk_size": 50,
  "depth_incremental_training": false,
  "enable_expert_vision": false,
  "expert_hidden_size": 768,
  "expert_intermediate_size": 2752,
  "expert_vision_path": null,
  "expert_vision_type": null,
  "final_norm_adanorm": false,
  "freeze_vision_encoder": false,
  "incremental_training": false,
  "initializer_range": 0.02,
  "loss_type": "L1_fm",
  "max_action_dim": 55,
  "max_state_dim": 55,
  "model_type": "lingbotvla",
  "moe_implementation": "fused",
  "n_action_steps": 50,
  "n_obs_steps": 1,
  "norm_qkv": false,
  "nosplit_gate_liner": false,
  "num_steps": 10,
  "post_training": true,
  "precompute_grid_thw": true,
  "proj_width": 768,
  "qwen3vl_use_vision_boundaries": true,
  "reinit_mismatched_weights": false,
  "return_image_grid_thw": true,
  "routed_scaling_factor": 4.0,
  "router_activation": "sigmoid",
  "router_z_loss_coeff": 0.0001,
  "separate_time_proj": false,
  "sequence_wise_loss_coeff": 0.001,
  "sequence_wise_mode": "per_sequence",
  "split_fused_experts_from_decoder_fsdp": false,
  "split_gate_liner": false,
  "token_moe_intermediate_size": 512,
  "token_moe_layers": [
    0,
    1,
    2,
    3,
    4,
    5,
    6,
    7,
    8,
    9,
    10,
    11,
    12,
    13,
    14,
    15,
    16,
    17,
    18,
    19,
    20,
    21,
    22,
    23,
    24,
    25,
    26,
    27,
    28,
    29,
    30,
    31,
    32,
    33,
    34,
    35
  ],
  "token_num_experts": 32,
  "token_shared_intermediate_size": 704,
  "token_top_k": 4,
  "tokenizer_max_length": 72,
  "tokenizer_path": "assets/pretrained/Qwen3-VL-4B-Instruct",
  "train_expert_only": false,
  "train_state_proj": true,
  "transformers_version": "4.57.3",
  "use_cache": false,
  "use_compile": true,
  "use_lm_head": false,
  "use_moe": true,
  "use_qwen3_chat_template": true,
  "use_qwen3_fixed_grid_cache": true,
  "use_shared_expert_gate": false,
  "vit_attn_implementation": "flash_attention_2",
  "vlm_causal": true,
  "vlm_family": "qwen3_vl",
  "vlm_repo_id": null,
  "vocab_size": 257152
}