Upload folder using huggingface_hub

controlnet/config.json

@@ -0,0 +1,113 @@
+{
+  "_class_name": "BEVControlNetModel",
+  "_diffusers_version": "0.17.1",
+  "act_fn": "silu",
+  "attention_head_dim": [
+    5,
+    10,
+    20,
+    20
+  ],
+  "bbox_embedder_cls": "xscene.networks.bbox_embedder.ContinuousBBoxWithTextEmbedding",
+  "bbox_embedder_param": {
+    "class_token_dim": 1024,
+    "embedder_num_freq": 4,
+    "minmax_normalize": false,
+    "mode": "all-xyz",
+    "n_classes": 10,
+    "proj_dims": [
+      1024,
+      512,
+      512,
+      1024
+    ],
+    "trainable_class_token": false,
+    "use_text_encoder_init": true
+  },
+  "block_out_channels": [
+    320,
+    640,
+    1280,
+    1280
+  ],
+  "cam_embedder_param": {
+    "include_input": true,
+    "input_dims": 3,
+    "log_sampling": true,
+    "num_freqs": 4
+  },
+  "camera_in_dim": 189,
+  "camera_out_dim": 1024,
+  "canvas_conditioning_channels": 14,
+  "canvas_size": [
+    14,
+    224,
+    400
+  ],
+  "class_embed_type": null,
+  "conditioning_embedding_out_channels": [
+    16,
+    32,
+    96,
+    256
+  ],
+  "controlnet_conditioning_channel_order": "rgb",
+  "cross_attention_dim": 1024,
+  "down_block_types": [
+    "CrossAttnDownBlock2D",
+    "CrossAttnDownBlock2D",
+    "CrossAttnDownBlock2D",
+    "DownBlock2D"
+  ],
+  "downsample_padding": 1,
+  "drop_cam_num": 6,
+  "drop_cam_with_box": false,
+  "drop_cond_ratio": 0.25,
+  "flip_sin_to_cos": true,
+  "freq_shift": 0,
+  "global_pool_conditions": false,
+  "in_channels": 4,
+  "layers_per_block": 2,
+  "map_embedder_cls": null,
+  "map_embedder_param": null,
+  "map_size": [
+    4,
+    200,
+    200
+  ],
+  "mid_block_scale_factor": 1,
+  "norm_eps": 1e-05,
+  "norm_num_groups": 32,
+  "num_class_embeds": null,
+  "occrender_conditioning_channels": 20,
+  "occrender_embedding_out_channels": [
+    16,
+    32,
+    64,
+    96,
+    256
+  ],
+  "occrender_output_size": null,
+  "only_cross_attention": false,
+  "projection_class_embeddings_input_dim": null,
+  "render_depth_size": [
+    1,
+    224,
+    400
+  ],
+  "render_img_size": [
+    20,
+    224,
+    400
+  ],
+  "resnet_time_scale_shift": "default",
+  "uncond_cam_in_dim": [
+    3,
+    7
+  ],
+  "upcast_attention": false,
+  "use_linear_projection": true,
+  "use_uncond_map": null,
+  "with_layout_canvas": true,
+  "with_occ_render_img": false
+}

controlnet/diffusion_pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:048ee50cc1f6cb1df0c3bca48fd9b37b5da538ded2dda3a34b397bd79f5084e1
+size 1655442195

hydra/config.yaml

@@ -0,0 +1,474 @@
+task_id: 224x400
+log_root_prefix: ./work_dirs/x-scene-img_224x400
+projname: ${model.name}
+try_run: false
+debug: false
+log_root: ./work_dirs
+init_method: env://
+seed: 42
+fix_seed_within_batch: false
+resume_from_checkpoint: work_dirs/x-scene-img_224x400/img_unet_2025-07-16_10-33_224x400/checkpoint-80000
+resume_only_model: false
+resume_reset_scheduler: false
+validation_only: false
+model:
+  name: img_unet
+  pretrained_model_name_or_path: pretrained/stable-diffusion-v2-1/
+  pretrained_t5_path: pretrained/t5-large/
+  bbox_mode: all-xyz
+  bbox_view_shared: false
+  crossview_attn_type: t5_crossview
+  train_with_same_noise: false
+  train_with_same_t: true
+  runner_module: xscene.runner.multiview_runner.MultiviewRunner
+  pipe_module: xscene.pipeline.pipeline_bev_controlnet.StableDiffusionBEVControlNetPipeline
+  unet_module: xscene.networks.unet_2d_condition_multiview.UNet2DConditionModelMultiview
+  use_fp32_for_unet_trainable: true
+  unet_dir: unet
+  unet:
+    trainable_state: only_new
+    neighboring_view_pair: ${dataset.neighboring_view_pair}
+    neighboring_attn_type: add
+    zero_module_type: zero_linear
+    crossview_attn_type: ${..crossview_attn_type}
+    img_size: ${dataset.image_size}
+    scene_channels: 320
+    attn1_q_trainable: true
+  scene_embedder_cls: xscene.networks.scene_position_embedder.ScenePositionEmbedding
+  scene_embedder_dir: scene_embedder
+  scene_embedder:
+    embed_dims: 320
+    LID: false
+  model_module: xscene.networks.unet_addon_rawbox.BEVControlNetModel
+  controlnet_dir: controlnet
+  controlnet:
+    camera_in_dim: 189
+    camera_out_dim: 1024
+    map_size:
+    - 4
+    - 200
+    - 200
+    conditioning_embedding_out_channels:
+    - 16
+    - 32
+    - 96
+    - 256
+    uncond_cam_in_dim:
+    - 3
+    - 7
+    use_uncond_map: null
+    drop_cond_ratio: 0.25
+    drop_cam_num: 6
+    drop_cam_with_box: false
+    cam_embedder_param:
+      input_dims: 3
+      num_freqs: 4
+      include_input: true
+      log_sampling: true
+    bbox_embedder_cls: xscene.networks.bbox_embedder.ContinuousBBoxWithTextEmbedding
+    bbox_embedder_param:
+      n_classes: 10
+      class_token_dim: 1024
+      trainable_class_token: false
+      use_text_encoder_init: true
+      embedder_num_freq: 4
+      proj_dims:
+      - 1024
+      - 512
+      - 512
+      - 1024
+      mode: ${...bbox_mode}
+      minmax_normalize: false
+    with_layout_canvas: true
+    canvas_conditioning_channels: 14
+    canvas_size:
+    - 14
+    - 224
+    - 400
+    with_occ_render_img: false
+    occrender_conditioning_channels: 20
+    render_img_size:
+    - 20
+    - 224
+    - 400
+    occrender_embedding_out_channels:
+    - 16
+    - 32
+    - 64
+    - 96
+    - 256
+dataset:
+  dataset_type: NuScenesDatasetM
+  occ_dataset_type: Occ3D-nuScenes
+  dataset_root: data/nuscenes/
+  triplane_root: data/nuscenes/nuscenes_triplane
+  dataset_process_root: data/nuscenes/nuscenes_mmdet3d-keyframes/
+  dataset_cache_file_tag: 200x200_12Hz_interp
+  dataset_cache_file:
+  - ${dataset.dataset_process_root}../nuscenes_map_aux_12Hz_interp/train_${dataset.dataset_cache_file_tag}.h5
+  - ${dataset.dataset_process_root}../nuscenes_map_aux_12Hz_interp/val_${dataset.dataset_cache_file_tag}.h5
+  template_clip: A driving scene image at {location}. {description}.
+  template_t5: A driving scene at {location}. {description}. {detailed_description}
+  image_size:
+  - 224
+  - 400
+  map_bound:
+    x:
+    - -40.0
+    - 40.0
+    - 0.4
+    'y':
+    - -40.0
+    - 40.0
+    - 0.4
+    z:
+    - -1.0
+    - 5.4
+    - 0.4
+  tri_size:
+  - 100
+  - 100
+  - 16
+  view_order:
+  - CAM_FRONT_LEFT
+  - CAM_FRONT
+  - CAM_FRONT_RIGHT
+  - CAM_BACK_RIGHT
+  - CAM_BACK
+  - CAM_BACK_LEFT
+  neighboring_view_pair:
+    0:
+    - 5
+    - 1
+    1:
+    - 0
+    - 2
+    2:
+    - 1
+    - 3
+    3:
+    - 2
+    - 4
+    4:
+    - 3
+    - 5
+    5:
+    - 4
+    - 0
+  back_resize:
+  - 896
+  - 1600
+  back_pad:
+  - 0
+  - 4
+  - 0
+  - 0
+  augment2d:
+    resize:
+    - - 0.25
+      - 0.25
+    rotate: null
+  aux_data:
+  - visibility
+  - center_offset
+  - center_ohw
+  - height
+  augment3d:
+    scale:
+    - 1.0
+    - 1.0
+    rotate:
+    - 0.0
+    - 0.0
+    translate: 0
+    flip_ratio: 0.0
+    flip_direction: null
+  object_classes:
+  - barrier
+  - bicycle
+  - bus
+  - car
+  - construction_vehicle
+  - motorcycle
+  - pedestrian
+  - traffic_cone
+  - trailer
+  - truck
+  map_classes:
+  - drivable_area
+  - ped_crossing
+  - walkway
+  - stop_line
+  - carpark_area
+  - road_divider
+  - lane_divider
+  - road_block
+  input_modality:
+    use_lidar: false
+    use_camera: true
+    use_radar: false
+    use_map: false
+    use_external: false
+  train_pipeline:
+  - type: LoadMultiViewImageFromFiles
+    to_float32: true
+  - type: LoadAnnotations3D
+    with_bbox_3d: true
+    with_label_3d: true
+    with_attr_label: false
+  - type: ImageAug3D
+    final_dim: ${...image_size}
+    resize_lim: ${...augment2d.resize[0]}
+    bot_pct_lim:
+    - 0.0
+    - 0.0
+    rot_lim: ${...augment2d.rotate}
+    rand_flip: false
+    is_train: false
+  - type: GlobalRotScaleTrans
+    resize_lim: ${...augment3d.scale}
+    rot_lim: ${...augment3d.rotate}
+    trans_lim: ${...augment3d.translate}
+    is_train: true
+  - type: ObjectNameFilterM
+    classes: ${...object_classes}
+  - type: LoadBEVSegmentationM
+    dataset_root: ${...dataset_root}
+    xbound: ${...map_bound.x}
+    ybound: ${...map_bound.y}
+    classes: ${...map_classes}
+    object_classes: ${...object_classes}
+    aux_data: ${...aux_data}
+    cache_file: ${...dataset_cache_file.0}
+  - type: LoadBEVHDMap
+    dataset_root: ${...dataset_root}
+    xbound: ${...map_bound.x}
+    ybound: ${...map_bound.y}
+    image_size: ${...image_size}
+    object_classes: ${...object_classes}
+  - type: RandomFlip3DwithViews
+    flip_ratio: ${...augment3d.flip_ratio}
+    direction: ${...augment3d.flip_direction}
+  - type: LoadDescription
+    dataset_root: ${...dataset_root}
+    dataset_type: ${...occ_dataset_type}
+  - type: ReorderMultiViewImagesM
+    order: ${...view_order}
+    safe: false
+  - type: ImageNormalize
+    mean:
+    - 0.5
+    - 0.5
+    - 0.5
+    std:
+    - 0.5
+    - 0.5
+    - 0.5
+  - type: DefaultFormatBundle3D
+    classes: ${...object_classes}
+  - type: Collect3D
+    keys:
+    - img
+    - gt_bboxes_3d
+    - gt_labels_3d
+    - gt_masks_bev
+    - gt_aux_bev
+    - bev_hdmap
+    - bev_hdmap_w_box
+    - layout_canvas
+    meta_keys:
+    - camera_intrinsics
+    - lidar2ego
+    - lidar2camera
+    - camera2lidar
+    - lidar2image
+    - img_aug_matrix
+    meta_lis_keys:
+    - timeofday
+    - location
+    - description
+    - detailed_description
+    - filename
+    - token
+  test_pipeline:
+  - type: LoadMultiViewImageFromFiles
+    to_float32: true
+  - type: LoadAnnotations3D
+    with_bbox_3d: true
+    with_label_3d: true
+    with_attr_label: false
+  - type: ImageAug3D
+    final_dim: ${...image_size}
+    resize_lim: ${...augment2d.resize[0]}
+    bot_pct_lim:
+    - 0.0
+    - 0.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    rand_flip: false
+    is_train: false
+  - type: GlobalRotScaleTrans
+    resize_lim: ${...augment3d.scale}
+    rot_lim: ${...augment3d.rotate}
+    trans_lim: ${...augment3d.translate}
+    is_train: true
+  - type: ObjectNameFilterM
+    classes: ${...object_classes}
+  - type: LoadBEVSegmentationM
+    dataset_root: ${...dataset_root}
+    xbound: ${...map_bound.x}
+    ybound: ${...map_bound.y}
+    classes: ${...map_classes}
+    object_classes: ${...object_classes}
+    aux_data: ${...aux_data}
+    cache_file: ${...dataset_cache_file.1}
+  - type: LoadBEVHDMap
+    dataset_root: ${...dataset_root}
+    xbound: ${...map_bound.x}
+    ybound: ${...map_bound.y}
+    image_size: ${...image_size}
+    object_classes: ${...object_classes}
+  - type: LoadDescription
+    dataset_root: ${...dataset_root}
+    dataset_type: ${...occ_dataset_type}
+  - type: ReorderMultiViewImagesM
+    order: ${...view_order}
+    safe: false
+  - type: ImageNormalize
+    mean:
+    - 0.5
+    - 0.5
+    - 0.5
+    std:
+    - 0.5
+    - 0.5
+    - 0.5
+  - type: DefaultFormatBundle3D
+    classes: ${...object_classes}
+  - type: Collect3D
+    keys:
+    - img
+    - gt_bboxes_3d
+    - gt_labels_3d
+    - gt_masks_bev
+    - gt_aux_bev
+    - bev_hdmap
+    - bev_hdmap_w_box
+    - layout_canvas
+    meta_keys:
+    - camera_intrinsics
+    - lidar2ego
+    - ego2global
+    - lidar2camera
+    - camera2lidar
+    - lidar2image
+    - img_aug_matrix
+    meta_lis_keys:
+    - timeofday
+    - location
+    - description
+    - detailed_description
+    - filename
+    - token
+    - lidar_token
+    - scene_name
+    - timestamp
+  data:
+    train:
+      type: ${...dataset_type}
+      dataset_root: ${...dataset_root}
+      ann_file: ${...dataset_process_root}nuscenes_infos_train.pkl
+      pipeline: ${...train_pipeline}
+      object_classes: ${...object_classes}
+      map_classes: ${...map_classes}
+      modality: ${...input_modality}
+      test_mode: false
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+    val:
+      type: ${...dataset_type}
+      dataset_root: ${...dataset_root}
+      ann_file: ${...dataset_process_root}nuscenes_infos_val.pkl
+      pipeline: ${...test_pipeline}
+      object_classes: ${...object_classes}
+      map_classes: ${...map_classes}
+      modality: ${...input_modality}
+      test_mode: false
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+    test:
+      type: ${...dataset_type}
+      dataset_root: ${...dataset_root}
+      ann_file: ${...dataset_process_root}nuscenes_infos_val.pkl
+      pipeline: ${...test_pipeline}
+      object_classes: ${...object_classes}
+      map_classes: ${...map_classes}
+      modality: ${...input_modality}
+      test_mode: true
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+occ_render_path: data/nuscenes/occ_render_map/
+accelerator:
+  gradient_accumulation_steps: 1
+  mixed_precision: fp16
+  report_to: tensorboard
+runner:
+  foreground_loss_weight: 0.0
+  bbox_drop_ratio: 0
+  bbox_add_ratio: 0.1
+  bbox_add_num: 3
+  keyframe_rate: 1
+  num_train_epochs: 115
+  train_batch_size: 10
+  max_train_steps: null
+  num_workers: 8
+  prefetch_factor: 4
+  display_per_epoch: 20
+  display_per_n_min: 10
+  max_grad_norm: 1.0
+  set_grads_to_none: true
+  enable_xformers_memory_efficient_attention: true
+  unet_in_fp16: true
+  enable_unet_checkpointing: true
+  enable_controlnet_checkpointing: true
+  noise_offset: 0.0
+  train_with_same_offset: true
+  use_8bit_adam: false
+  adam_beta1: 0.9
+  adam_beta2: 0.999
+  adam_weight_decay: 0.01
+  adam_epsilon: 1.0e-08
+  learning_rate: 8.0e-05
+  lr_scheduler: constant_with_warmup
+  gradient_accumulation_steps: 1
+  lr_num_cycles: 1
+  lr_power: 1.0
+  lr_warmup_steps: 3000
+  checkpointing_steps: 5000
+  validation_steps: 20000
+  save_model_per_epoch: null
+  validation_before_run: false
+  validation_index:
+  - 204
+  - 912
+  - 1828
+  - 2253
+  - 4467
+  - 5543
+  validation_times: 4
+  validation_batch_size: 1
+  validation_show_box: true
+  validation_show_line: true
+  validation_seed_global: false
+  pipeline_param:
+    guidance_scale: 1.2
+    num_inference_steps: 20
+    eta: 0.0
+    controlnet_conditioning_scale: 1.0
+    guess_mode: false
+    use_zero_map_as_unconditional: false
+    bbox_max_length: null

hydra/hydra.yaml

@@ -0,0 +1,162 @@
+hydra:
+  run:
+    dir: ${log_root_prefix}/${projname}_${now:%Y-%m-%d}_${now:%H-%M}_${task_id}
+  sweep:
+    dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task:
+    - +exp=224x400
+    - runner=4gpus
+  job:
+    name: train
+    chdir: null
+    override_dirname: +exp=224x400,runner=4gpus
+    id: ???
+    num: ???
+    config_name: config_img
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.3.0
+    version_base: '1.3'
+    cwd: /data/yyang/workspace/X-Scene
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /data/yyang/workspace/X-Scene/configs
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /data/yyang/workspace/X-Scene/work_dirs/x-scene-img_224x400/img_unet_2025-07-18_17-19_224x400
+    choices:
+      exp: 224x400
+      exp/model@model: ../../model/img_unet
+      runner: 4gpus
+      accelerator: default
+      dataset: Nuscenes_map_cache_box
+      model: img_unet
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

hydra/overrides.yaml

@@ -0,0 +1,2 @@
+- +exp=224x400
+- runner=4gpus_img

scene_embedder/scene_embedder_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f01fee2a89693024a951b8a43c8ab20eb32ef494923c7124076bb94e8dae6089
+size 9185335

tb-224x400/events.out.tfevents.1752830422.cvrp-gpu-11.2082924.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b537ad395beff95ff4f5c2a6edbd5c00e589ea02e06a19165406d5d9521e7131
+size 83560

train.1.log

@@ -0,0 +1,1504 @@
+[2025-07-18 17:19:48,751][root][INFO] - using data cache from: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+[2025-07-18 17:19:48,983][root][INFO] - [RandomFlip3DwithViews] ratio=0.0, direction=None, reorder=True
+[2025-07-18 17:19:51,665][root][INFO] - using data cache from: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+[2025-07-18 17:20:16,341][root][INFO] - [UNet2DConditionModelMultiview] load pretrained with missing_keys: ['down_blocks.0.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.0.attentions.0.scene_proj_in.weight', 'down_blocks.0.attentions.0.scene_proj_in.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.0.attentions.1.scene_proj_in.weight', 'down_blocks.0.attentions.1.scene_proj_in.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.1.attentions.0.scene_proj_in.weight', 'down_blocks.1.attentions.0.scene_proj_in.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.1.attentions.1.scene_proj_in.weight', 'down_blocks.1.attentions.1.scene_proj_in.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.2.attentions.0.scene_proj_in.weight', 'down_blocks.2.attentions.0.scene_proj_in.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.2.attentions.1.scene_proj_in.weight', 'down_blocks.2.attentions.1.scene_proj_in.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.0.scene_proj_in.weight', 'up_blocks.1.attentions.0.scene_proj_in.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.1.scene_proj_in.weight', 'up_blocks.1.attentions.1.scene_proj_in.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.2.scene_proj_in.weight', 'up_blocks.1.attentions.2.scene_proj_in.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.0.scene_proj_in.weight', 'up_blocks.2.attentions.0.scene_proj_in.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.1.scene_proj_in.weight', 'up_blocks.2.attentions.1.scene_proj_in.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.2.scene_proj_in.weight', 'up_blocks.2.attentions.2.scene_proj_in.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.0.scene_proj_in.weight', 'up_blocks.3.attentions.0.scene_proj_in.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.1.scene_proj_in.weight', 'up_blocks.3.attentions.1.scene_proj_in.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.2.scene_proj_in.weight', 'up_blocks.3.attentions.2.scene_proj_in.bias', 'mid_block.attentions.0.transformer_blocks.0.norm4.weight', 'mid_block.attentions.0.transformer_blocks.0.norm4.bias', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'mid_block.attentions.0.transformer_blocks.0.norm5.weight', 'mid_block.attentions.0.transformer_blocks.0.norm5.bias', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'mid_block.attentions.0.transformer_blocks.0.connector_4.weight', 'mid_block.attentions.0.transformer_blocks.0.connector_4.bias', 'mid_block.attentions.0.transformer_blocks.0.connector_5.weight', 'mid_block.attentions.0.transformer_blocks.0.connector_5.bias', 'mid_block.attentions.0.scene_proj_in.weight', 'mid_block.attentions.0.scene_proj_in.bias']; unexpected_keys: []
+[2025-07-18 17:20:16,343][root][DEBUG] - [BEVControlNetModel] instantiating your own version of controlnet.
+[2025-07-18 17:20:16,344][root][DEBUG] - embedder out dim = 27
+[2025-07-18 17:20:16,355][root][DEBUG] - [BEVControlNetModel] map_embedder: BEVControlNetConditioningEmbedding(
+  (conv_in): Conv2d(4, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+  (blocks): ModuleList(
+    (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (1): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(2, 1))
+    (2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (3): Conv2d(32, 96, kernel_size=(3, 3), stride=(2, 2), padding=(2, 1))
+    (4): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(2, 1))
+    (5): Conv2d(96, 256, kernel_size=(3, 3), stride=(2, 1), padding=(2, 1))
+  )
+  (conv_out): Conv2d(256, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+)
+[2025-07-18 17:20:16,359][root][DEBUG] - [BEVControlNetModel] canvas_embedder: ControlNetConditioningEmbedding(
+  (conv_in): Conv2d(14, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+  (blocks): ModuleList(
+    (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (1): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+    (2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (3): Conv2d(32, 96, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+    (4): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (5): Conv2d(96, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+  )
+  (conv_out): Conv2d(256, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+)
+[2025-07-18 17:20:16,361][root][DEBUG] - embedder out dim = 27
+[2025-07-18 17:20:16,361][root][INFO] - [ContinuousBBoxWithTextEmbedding] bbox embedder has 27 dims.
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,728][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,729][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set position_encoder.0.weight to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set position_encoder.0.bias to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set position_encoder.2.weight to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set position_encoder.2.bias to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set conv_in.weight to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set conv_in.bias to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set blocks.0.weight to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set blocks.0.bias to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set blocks.1.weight to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set blocks.1.bias to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set conv_out.weight to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set conv_out.bias to requires_grad = True
+[2025-07-18 17:20:21,193][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,194][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,194][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,194][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'diffusers.models.unet_2d_blocks.DownBlock2D'>] to gradient_checkpointing
+[2025-07-18 17:20:21,194][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'diffusers.models.unet_2d_blocks.UpBlock2D'>] to gradient_checkpointing
+[2025-07-18 17:20:21,194][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,194][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,194][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,196][root][INFO] - [BaseValidator] Validator use model_param: dict_keys(['vae', 'text_encoder', 'text_encoder_t5', 'tokenizer', 'tokenizer_t5'])
+[2025-07-18 17:20:21,197][root][INFO] - [MultiviewRunner] add 130.82 M params from unet to optimizer.
+[2025-07-18 17:20:21,198][root][INFO] - [MultiviewRunner] have total 525.45 M params from unet and controlnet to optimizer.
+[2025-07-18 17:20:21,198][root][INFO] - [MultiviewRunner] add 2.19 M params from scene_embedder to optimizer.
+[2025-07-18 17:20:22,559][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,560][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,561][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,562][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,563][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,567][root][INFO] - [ContinuousBBoxWithTextEmbedding] Initialzing your class_tokens with text_encoder
+[2025-07-18 17:20:22,763][root][DEBUG] - Current config:
+task_id: 224x400
+log_root_prefix: ./work_dirs/x-scene-img_224x400
+projname: img_unet
+try_run: false
+debug: false
+log_root: /data/yyang/workspace/X-Scene/work_dirs/x-scene-img_224x400/img_unet_2025-07-18_17-19_224x400
+init_method: env://
+seed: 42
+fix_seed_within_batch: false
+resume_from_checkpoint: work_dirs/x-scene-img_224x400/img_unet_2025-07-16_10-33_224x400/checkpoint-80000
+resume_only_model: false
+resume_reset_scheduler: false
+validation_only: false
+model:
+  name: img_unet
+  pretrained_model_name_or_path: pretrained/stable-diffusion-v2-1/
+  pretrained_t5_path: pretrained/t5-large/
+  bbox_mode: all-xyz
+  bbox_view_shared: false
+  crossview_attn_type: t5_crossview
+  train_with_same_noise: false
+  train_with_same_t: true
+  runner_module: xscene.runner.multiview_runner.MultiviewRunner
+  pipe_module: xscene.pipeline.pipeline_bev_controlnet.StableDiffusionBEVControlNetPipeline
+  unet_module: xscene.networks.unet_2d_condition_multiview.UNet2DConditionModelMultiview
+  use_fp32_for_unet_trainable: true
+  unet_dir: unet
+  unet:
+    trainable_state: only_new
+    neighboring_view_pair:
+      0:
+      - 5
+      - 1
+      1:
+      - 0
+      - 2
+      2:
+      - 1
+      - 3
+      3:
+      - 2
+      - 4
+      4:
+      - 3
+      - 5
+      5:
+      - 4
+      - 0
+    neighboring_attn_type: add
+    zero_module_type: zero_linear
+    crossview_attn_type: t5_crossview
+    img_size:
+    - 224
+    - 400
+    scene_channels: 320
+    attn1_q_trainable: true
+  scene_embedder_cls: xscene.networks.scene_position_embedder.ScenePositionEmbedding
+  scene_embedder_dir: scene_embedder
+  scene_embedder:
+    embed_dims: 320
+    LID: false
+  model_module: xscene.networks.unet_addon_rawbox.BEVControlNetModel
+  controlnet_dir: controlnet
+  controlnet:
+    camera_in_dim: 189
+    camera_out_dim: 1024
+    map_size:
+    - 4
+    - 200
+    - 200
+    conditioning_embedding_out_channels:
+    - 16
+    - 32
+    - 96
+    - 256
+    uncond_cam_in_dim:
+    - 3
+    - 7
+    use_uncond_map: null
+    drop_cond_ratio: 0.25
+    drop_cam_num: 6
+    drop_cam_with_box: false
+    cam_embedder_param:
+      input_dims: 3
+      num_freqs: 4
+      include_input: true
+      log_sampling: true
+    bbox_embedder_cls: xscene.networks.bbox_embedder.ContinuousBBoxWithTextEmbedding
+    bbox_embedder_param:
+      n_classes: 10
+      class_token_dim: 1024
+      trainable_class_token: false
+      use_text_encoder_init: true
+      embedder_num_freq: 4
+      proj_dims:
+      - 1024
+      - 512
+      - 512
+      - 1024
+      mode: all-xyz
+      minmax_normalize: false
+    with_layout_canvas: true
+    canvas_conditioning_channels: 14
+    canvas_size:
+    - 14
+    - 224
+    - 400
+    with_occ_render_img: false
+    occrender_conditioning_channels: 20
+    render_img_size:
+    - 20
+    - 224
+    - 400
+    occrender_embedding_out_channels:
+    - 16
+    - 32
+    - 64
+    - 96
+    - 256
+dataset:
+  dataset_type: NuScenesDatasetM
+  occ_dataset_type: Occ3D-nuScenes
+  dataset_root: data/nuscenes/
+  triplane_root: data/nuscenes/nuscenes_triplane
+  dataset_process_root: data/nuscenes/nuscenes_mmdet3d-keyframes/
+  dataset_cache_file_tag: 200x200_12Hz_interp
+  dataset_cache_file:
+  - data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+  - data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+  template_clip: A driving scene image at {location}. {description}.
+  template_t5: A driving scene at {location}. {description}. {detailed_description}
+  image_size:
+  - 224
+  - 400
+  map_bound:
+    x:
+    - -40.0
+    - 40.0
+    - 0.4
+    'y':
+    - -40.0
+    - 40.0
+    - 0.4
+    z:
+    - -1.0
+    - 5.4
+    - 0.4
+  tri_size:
+  - 100
+  - 100
+  - 16
+  view_order:
+  - CAM_FRONT_LEFT
+  - CAM_FRONT
+  - CAM_FRONT_RIGHT
+  - CAM_BACK_RIGHT
+  - CAM_BACK
+  - CAM_BACK_LEFT
+  neighboring_view_pair:
+    0:
+    - 5
+    - 1
+    1:
+    - 0
+    - 2
+    2:
+    - 1
+    - 3
+    3:
+    - 2
+    - 4
+    4:
+    - 3
+    - 5
+    5:
+    - 4
+    - 0
+  back_resize:
+  - 896
+  - 1600
+  back_pad:
+  - 0
+  - 4
+  - 0
+  - 0
+  augment2d:
+    resize:
+    - - 0.25
+      - 0.25
+    rotate: null
+  aux_data:
+  - visibility
+  - center_offset
+  - center_ohw
+  - height
+  augment3d:
+    scale:
+    - 1.0
+    - 1.0
+    rotate:
+    - 0.0
+    - 0.0
+    translate: 0
+    flip_ratio: 0.0
+    flip_direction: null
+  object_classes:
+  - barrier
+  - bicycle
+  - bus
+  - car
+  - construction_vehicle
+  - motorcycle
+  - pedestrian
+  - traffic_cone
+  - trailer
+  - truck
+  map_classes:
+  - drivable_area
+  - ped_crossing
+  - walkway
+  - stop_line
+  - carpark_area
+  - road_divider
+  - lane_divider
+  - road_block
+  input_modality:
+    use_lidar: false
+    use_camera: true
+    use_radar: false
+    use_map: false
+    use_external: false
+  train_pipeline:
+  - type: LoadMultiViewImageFromFiles
+    to_float32: true
+  - type: LoadAnnotations3D
+    with_bbox_3d: true
+    with_label_3d: true
+    with_attr_label: false
+  - type: ImageAug3D
+    final_dim:
+    - 224
+    - 400
+    resize_lim:
+    - 0.25
+    - 0.25
+    bot_pct_lim:
+    - 0.0
+    - 0.0
+    rot_lim: null
+    rand_flip: false
+    is_train: false
+  - type: GlobalRotScaleTrans
+    resize_lim:
+    - 1.0
+    - 1.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    trans_lim: 0
+    is_train: true
+  - type: ObjectNameFilterM
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadBEVSegmentationM
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    classes:
+    - drivable_area
+    - ped_crossing
+    - walkway
+    - stop_line
+    - carpark_area
+    - road_divider
+    - lane_divider
+    - road_block
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+    aux_data:
+    - visibility
+    - center_offset
+    - center_ohw
+    - height
+    cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+  - type: LoadBEVHDMap
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    image_size:
+    - 224
+    - 400
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: RandomFlip3DwithViews
+    flip_ratio: 0.0
+    direction: null
+  - type: LoadDescription
+    dataset_root: data/nuscenes/
+    dataset_type: Occ3D-nuScenes
+  - type: ReorderMultiViewImagesM
+    order:
+    - CAM_FRONT_LEFT
+    - CAM_FRONT
+    - CAM_FRONT_RIGHT
+    - CAM_BACK_RIGHT
+    - CAM_BACK
+    - CAM_BACK_LEFT
+    safe: false
+  - type: ImageNormalize
+    mean:
+    - 0.5
+    - 0.5
+    - 0.5
+    std:
+    - 0.5
+    - 0.5
+    - 0.5
+  - type: DefaultFormatBundle3D
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: Collect3D
+    keys:
+    - img
+    - gt_bboxes_3d
+    - gt_labels_3d
+    - gt_masks_bev
+    - gt_aux_bev
+    - bev_hdmap
+    - bev_hdmap_w_box
+    - layout_canvas
+    meta_keys:
+    - camera_intrinsics
+    - lidar2ego
+    - lidar2camera
+    - camera2lidar
+    - lidar2image
+    - img_aug_matrix
+    meta_lis_keys:
+    - timeofday
+    - location
+    - description
+    - detailed_description
+    - filename
+    - token
+  test_pipeline:
+  - type: LoadMultiViewImageFromFiles
+    to_float32: true
+  - type: LoadAnnotations3D
+    with_bbox_3d: true
+    with_label_3d: true
+    with_attr_label: false
+  - type: ImageAug3D
+    final_dim:
+    - 224
+    - 400
+    resize_lim:
+    - 0.25
+    - 0.25
+    bot_pct_lim:
+    - 0.0
+    - 0.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    rand_flip: false
+    is_train: false
+  - type: GlobalRotScaleTrans
+    resize_lim:
+    - 1.0
+    - 1.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    trans_lim: 0
+    is_train: true
+  - type: ObjectNameFilterM
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadBEVSegmentationM
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    classes:
+    - drivable_area
+    - ped_crossing
+    - walkway
+    - stop_line
+    - carpark_area
+    - road_divider
+    - lane_divider
+    - road_block
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+    aux_data:
+    - visibility
+    - center_offset
+    - center_ohw
+    - height
+    cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+  - type: LoadBEVHDMap
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    image_size:
+    - 224
+    - 400
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadDescription
+    dataset_root: data/nuscenes/
+    dataset_type: Occ3D-nuScenes
+  - type: ReorderMultiViewImagesM
+    order:
+    - CAM_FRONT_LEFT
+    - CAM_FRONT
+    - CAM_FRONT_RIGHT
+    - CAM_BACK_RIGHT
+    - CAM_BACK
+    - CAM_BACK_LEFT
+    safe: false
+  - type: ImageNormalize
+    mean:
+    - 0.5
+    - 0.5
+    - 0.5
+    std:
+    - 0.5
+    - 0.5
+    - 0.5
+  - type: DefaultFormatBundle3D
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: Collect3D
+    keys:
+    - img
+    - gt_bboxes_3d
+    - gt_labels_3d
+    - gt_masks_bev
+    - gt_aux_bev
+    - bev_hdmap
+    - bev_hdmap_w_box
+    - layout_canvas
+    meta_keys:
+    - camera_intrinsics
+    - lidar2ego
+    - ego2global
+    - lidar2camera
+    - camera2lidar
+    - lidar2image
+    - img_aug_matrix
+    meta_lis_keys:
+    - timeofday
+    - location
+    - description
+    - detailed_description
+    - filename
+    - token
+    - lidar_token
+    - scene_name
+    - timestamp
+  data:
+    train:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_train.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim: null
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: RandomFlip3DwithViews
+        flip_ratio: 0.0
+        direction: null
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: false
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+    val:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_val.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - ego2global
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+        - lidar_token
+        - scene_name
+        - timestamp
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: false
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+    test:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_val.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - ego2global
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+        - lidar_token
+        - scene_name
+        - timestamp
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: true
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+occ_render_path: data/nuscenes/occ_render_map/
+accelerator:
+  gradient_accumulation_steps: 1
+  mixed_precision: fp16
+  report_to: tensorboard
+runner:
+  foreground_loss_weight: 0.0
+  bbox_drop_ratio: 0
+  bbox_add_ratio: 0.1
+  bbox_add_num: 3
+  keyframe_rate: 1
+  num_train_epochs: 115
+  train_batch_size: 10
+  max_train_steps: 80960
+  num_workers: 8
+  prefetch_factor: 4
+  display_per_epoch: 20
+  display_per_n_min: 10
+  max_grad_norm: 1.0
+  set_grads_to_none: true
+  enable_xformers_memory_efficient_attention: true
+  unet_in_fp16: true
+  enable_unet_checkpointing: true
+  enable_controlnet_checkpointing: true
+  noise_offset: 0.0
+  train_with_same_offset: true
+  use_8bit_adam: false
+  adam_beta1: 0.9
+  adam_beta2: 0.999
+  adam_weight_decay: 0.01
+  adam_epsilon: 1.0e-08
+  learning_rate: 8.0e-05
+  lr_scheduler: constant_with_warmup
+  gradient_accumulation_steps: 1
+  lr_num_cycles: 1
+  lr_power: 1.0
+  lr_warmup_steps: 3000
+  checkpointing_steps: 5000
+  validation_steps: 20000
+  save_model_per_epoch: null
+  validation_before_run: false
+  validation_index:
+  - 204
+  - 912
+  - 1828
+  - 2253
+  - 4467
+  - 5543
+  validation_times: 4
+  validation_batch_size: 1
+  validation_show_box: true
+  validation_show_line: true
+  validation_seed_global: false
+  pipeline_param:
+    guidance_scale: 1.2
+    num_inference_steps: 20
+    eta: 0.0
+    controlnet_conditioning_scale: 1.0
+    guess_mode: false
+    use_zero_map_as_unconditional: false
+    bbox_max_length: null
+
+[2025-07-18 17:20:22,763][root][DEBUG] - start!
+[2025-07-18 17:20:22,763][root][INFO] - ***** Running training *****
+[2025-07-18 17:20:22,763][root][INFO] -   Num examples = 28130
+[2025-07-18 17:20:22,763][root][INFO] -   Num batches each epoch = 704
+[2025-07-18 17:20:22,763][root][INFO] -   Num Epochs = 115
+[2025-07-18 17:20:22,763][root][INFO] -   Instantaneous batch size per device = 10
+[2025-07-18 17:20:22,763][root][INFO] -   Total train batch size (w. parallel, distributed & accumulation) = 40
+[2025-07-18 17:20:22,763][root][INFO] -   Gradient Accumulation steps = 1
+[2025-07-18 17:20:22,764][root][INFO] -   Total optimization steps = 80960
+[2025-07-18 17:20:25,708][root][INFO] - Starting from epoch 113 to 115
+[2025-07-18 17:20:39,371][root][WARNING] - [UNet2DConditionModelMultiview] Forward upsample size to force interpolation output size.
+[2025-07-18 18:00:02,390][root][INFO] - Save your model to: /data/yyang/workspace/X-Scene/work_dirs/x-scene-img_224x400/img_unet_2025-07-18_17-19_224x400

train.2.log

@@ -0,0 +1,1504 @@
+[2025-07-18 17:19:48,677][root][INFO] - using data cache from: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+[2025-07-18 17:19:48,908][root][INFO] - [RandomFlip3DwithViews] ratio=0.0, direction=None, reorder=True
+[2025-07-18 17:19:51,571][root][INFO] - using data cache from: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+[2025-07-18 17:20:16,346][root][INFO] - [UNet2DConditionModelMultiview] load pretrained with missing_keys: ['down_blocks.0.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.0.attentions.0.scene_proj_in.weight', 'down_blocks.0.attentions.0.scene_proj_in.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.0.attentions.1.scene_proj_in.weight', 'down_blocks.0.attentions.1.scene_proj_in.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.1.attentions.0.scene_proj_in.weight', 'down_blocks.1.attentions.0.scene_proj_in.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.1.attentions.1.scene_proj_in.weight', 'down_blocks.1.attentions.1.scene_proj_in.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.2.attentions.0.scene_proj_in.weight', 'down_blocks.2.attentions.0.scene_proj_in.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.2.attentions.1.scene_proj_in.weight', 'down_blocks.2.attentions.1.scene_proj_in.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.0.scene_proj_in.weight', 'up_blocks.1.attentions.0.scene_proj_in.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.1.scene_proj_in.weight', 'up_blocks.1.attentions.1.scene_proj_in.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.2.scene_proj_in.weight', 'up_blocks.1.attentions.2.scene_proj_in.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.0.scene_proj_in.weight', 'up_blocks.2.attentions.0.scene_proj_in.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.1.scene_proj_in.weight', 'up_blocks.2.attentions.1.scene_proj_in.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.2.scene_proj_in.weight', 'up_blocks.2.attentions.2.scene_proj_in.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.0.scene_proj_in.weight', 'up_blocks.3.attentions.0.scene_proj_in.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.1.scene_proj_in.weight', 'up_blocks.3.attentions.1.scene_proj_in.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.2.scene_proj_in.weight', 'up_blocks.3.attentions.2.scene_proj_in.bias', 'mid_block.attentions.0.transformer_blocks.0.norm4.weight', 'mid_block.attentions.0.transformer_blocks.0.norm4.bias', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'mid_block.attentions.0.transformer_blocks.0.norm5.weight', 'mid_block.attentions.0.transformer_blocks.0.norm5.bias', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'mid_block.attentions.0.transformer_blocks.0.connector_4.weight', 'mid_block.attentions.0.transformer_blocks.0.connector_4.bias', 'mid_block.attentions.0.transformer_blocks.0.connector_5.weight', 'mid_block.attentions.0.transformer_blocks.0.connector_5.bias', 'mid_block.attentions.0.scene_proj_in.weight', 'mid_block.attentions.0.scene_proj_in.bias']; unexpected_keys: []
+[2025-07-18 17:20:16,347][root][DEBUG] - [BEVControlNetModel] instantiating your own version of controlnet.
+[2025-07-18 17:20:16,348][root][DEBUG] - embedder out dim = 27
+[2025-07-18 17:20:16,360][root][DEBUG] - [BEVControlNetModel] map_embedder: BEVControlNetConditioningEmbedding(
+  (conv_in): Conv2d(4, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+  (blocks): ModuleList(
+    (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (1): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(2, 1))
+    (2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (3): Conv2d(32, 96, kernel_size=(3, 3), stride=(2, 2), padding=(2, 1))
+    (4): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(2, 1))
+    (5): Conv2d(96, 256, kernel_size=(3, 3), stride=(2, 1), padding=(2, 1))
+  )
+  (conv_out): Conv2d(256, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+)
+[2025-07-18 17:20:16,364][root][DEBUG] - [BEVControlNetModel] canvas_embedder: ControlNetConditioningEmbedding(
+  (conv_in): Conv2d(14, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+  (blocks): ModuleList(
+    (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (1): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+    (2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (3): Conv2d(32, 96, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+    (4): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (5): Conv2d(96, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+  )
+  (conv_out): Conv2d(256, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+)
+[2025-07-18 17:20:16,365][root][DEBUG] - embedder out dim = 27
+[2025-07-18 17:20:16,365][root][INFO] - [ContinuousBBoxWithTextEmbedding] bbox embedder has 27 dims.
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set position_encoder.0.weight to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set position_encoder.0.bias to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set position_encoder.2.weight to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set position_encoder.2.bias to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set conv_in.weight to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set conv_in.bias to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set blocks.0.weight to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set blocks.0.bias to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set blocks.1.weight to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set blocks.1.bias to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set conv_out.weight to requires_grad = True
+[2025-07-18 17:20:20,733][root][DEBUG] - [MultiviewRunner] set conv_out.bias to requires_grad = True
+[2025-07-18 17:20:21,174][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,174][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,174][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,174][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'diffusers.models.unet_2d_blocks.DownBlock2D'>] to gradient_checkpointing
+[2025-07-18 17:20:21,174][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'diffusers.models.unet_2d_blocks.UpBlock2D'>] to gradient_checkpointing
+[2025-07-18 17:20:21,174][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,175][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,175][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,176][root][INFO] - [BaseValidator] Validator use model_param: dict_keys(['vae', 'text_encoder', 'text_encoder_t5', 'tokenizer', 'tokenizer_t5'])
+[2025-07-18 17:20:21,178][root][INFO] - [MultiviewRunner] add 130.82 M params from unet to optimizer.
+[2025-07-18 17:20:21,178][root][INFO] - [MultiviewRunner] have total 525.45 M params from unet and controlnet to optimizer.
+[2025-07-18 17:20:21,178][root][INFO] - [MultiviewRunner] add 2.19 M params from scene_embedder to optimizer.
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,564][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,565][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,566][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,567][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,568][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,572][root][INFO] - [ContinuousBBoxWithTextEmbedding] Initialzing your class_tokens with text_encoder
+[2025-07-18 17:20:22,769][root][DEBUG] - Current config:
+task_id: 224x400
+log_root_prefix: ./work_dirs/x-scene-img_224x400
+projname: img_unet
+try_run: false
+debug: false
+log_root: /data/yyang/workspace/X-Scene/work_dirs/x-scene-img_224x400/img_unet_2025-07-18_17-19_224x400
+init_method: env://
+seed: 42
+fix_seed_within_batch: false
+resume_from_checkpoint: work_dirs/x-scene-img_224x400/img_unet_2025-07-16_10-33_224x400/checkpoint-80000
+resume_only_model: false
+resume_reset_scheduler: false
+validation_only: false
+model:
+  name: img_unet
+  pretrained_model_name_or_path: pretrained/stable-diffusion-v2-1/
+  pretrained_t5_path: pretrained/t5-large/
+  bbox_mode: all-xyz
+  bbox_view_shared: false
+  crossview_attn_type: t5_crossview
+  train_with_same_noise: false
+  train_with_same_t: true
+  runner_module: xscene.runner.multiview_runner.MultiviewRunner
+  pipe_module: xscene.pipeline.pipeline_bev_controlnet.StableDiffusionBEVControlNetPipeline
+  unet_module: xscene.networks.unet_2d_condition_multiview.UNet2DConditionModelMultiview
+  use_fp32_for_unet_trainable: true
+  unet_dir: unet
+  unet:
+    trainable_state: only_new
+    neighboring_view_pair:
+      0:
+      - 5
+      - 1
+      1:
+      - 0
+      - 2
+      2:
+      - 1
+      - 3
+      3:
+      - 2
+      - 4
+      4:
+      - 3
+      - 5
+      5:
+      - 4
+      - 0
+    neighboring_attn_type: add
+    zero_module_type: zero_linear
+    crossview_attn_type: t5_crossview
+    img_size:
+    - 224
+    - 400
+    scene_channels: 320
+    attn1_q_trainable: true
+  scene_embedder_cls: xscene.networks.scene_position_embedder.ScenePositionEmbedding
+  scene_embedder_dir: scene_embedder
+  scene_embedder:
+    embed_dims: 320
+    LID: false
+  model_module: xscene.networks.unet_addon_rawbox.BEVControlNetModel
+  controlnet_dir: controlnet
+  controlnet:
+    camera_in_dim: 189
+    camera_out_dim: 1024
+    map_size:
+    - 4
+    - 200
+    - 200
+    conditioning_embedding_out_channels:
+    - 16
+    - 32
+    - 96
+    - 256
+    uncond_cam_in_dim:
+    - 3
+    - 7
+    use_uncond_map: null
+    drop_cond_ratio: 0.25
+    drop_cam_num: 6
+    drop_cam_with_box: false
+    cam_embedder_param:
+      input_dims: 3
+      num_freqs: 4
+      include_input: true
+      log_sampling: true
+    bbox_embedder_cls: xscene.networks.bbox_embedder.ContinuousBBoxWithTextEmbedding
+    bbox_embedder_param:
+      n_classes: 10
+      class_token_dim: 1024
+      trainable_class_token: false
+      use_text_encoder_init: true
+      embedder_num_freq: 4
+      proj_dims:
+      - 1024
+      - 512
+      - 512
+      - 1024
+      mode: all-xyz
+      minmax_normalize: false
+    with_layout_canvas: true
+    canvas_conditioning_channels: 14
+    canvas_size:
+    - 14
+    - 224
+    - 400
+    with_occ_render_img: false
+    occrender_conditioning_channels: 20
+    render_img_size:
+    - 20
+    - 224
+    - 400
+    occrender_embedding_out_channels:
+    - 16
+    - 32
+    - 64
+    - 96
+    - 256
+dataset:
+  dataset_type: NuScenesDatasetM
+  occ_dataset_type: Occ3D-nuScenes
+  dataset_root: data/nuscenes/
+  triplane_root: data/nuscenes/nuscenes_triplane
+  dataset_process_root: data/nuscenes/nuscenes_mmdet3d-keyframes/
+  dataset_cache_file_tag: 200x200_12Hz_interp
+  dataset_cache_file:
+  - data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+  - data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+  template_clip: A driving scene image at {location}. {description}.
+  template_t5: A driving scene at {location}. {description}. {detailed_description}
+  image_size:
+  - 224
+  - 400
+  map_bound:
+    x:
+    - -40.0
+    - 40.0
+    - 0.4
+    'y':
+    - -40.0
+    - 40.0
+    - 0.4
+    z:
+    - -1.0
+    - 5.4
+    - 0.4
+  tri_size:
+  - 100
+  - 100
+  - 16
+  view_order:
+  - CAM_FRONT_LEFT
+  - CAM_FRONT
+  - CAM_FRONT_RIGHT
+  - CAM_BACK_RIGHT
+  - CAM_BACK
+  - CAM_BACK_LEFT
+  neighboring_view_pair:
+    0:
+    - 5
+    - 1
+    1:
+    - 0
+    - 2
+    2:
+    - 1
+    - 3
+    3:
+    - 2
+    - 4
+    4:
+    - 3
+    - 5
+    5:
+    - 4
+    - 0
+  back_resize:
+  - 896
+  - 1600
+  back_pad:
+  - 0
+  - 4
+  - 0
+  - 0
+  augment2d:
+    resize:
+    - - 0.25
+      - 0.25
+    rotate: null
+  aux_data:
+  - visibility
+  - center_offset
+  - center_ohw
+  - height
+  augment3d:
+    scale:
+    - 1.0
+    - 1.0
+    rotate:
+    - 0.0
+    - 0.0
+    translate: 0
+    flip_ratio: 0.0
+    flip_direction: null
+  object_classes:
+  - barrier
+  - bicycle
+  - bus
+  - car
+  - construction_vehicle
+  - motorcycle
+  - pedestrian
+  - traffic_cone
+  - trailer
+  - truck
+  map_classes:
+  - drivable_area
+  - ped_crossing
+  - walkway
+  - stop_line
+  - carpark_area
+  - road_divider
+  - lane_divider
+  - road_block
+  input_modality:
+    use_lidar: false
+    use_camera: true
+    use_radar: false
+    use_map: false
+    use_external: false
+  train_pipeline:
+  - type: LoadMultiViewImageFromFiles
+    to_float32: true
+  - type: LoadAnnotations3D
+    with_bbox_3d: true
+    with_label_3d: true
+    with_attr_label: false
+  - type: ImageAug3D
+    final_dim:
+    - 224
+    - 400
+    resize_lim:
+    - 0.25
+    - 0.25
+    bot_pct_lim:
+    - 0.0
+    - 0.0
+    rot_lim: null
+    rand_flip: false
+    is_train: false
+  - type: GlobalRotScaleTrans
+    resize_lim:
+    - 1.0
+    - 1.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    trans_lim: 0
+    is_train: true
+  - type: ObjectNameFilterM
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadBEVSegmentationM
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    classes:
+    - drivable_area
+    - ped_crossing
+    - walkway
+    - stop_line
+    - carpark_area
+    - road_divider
+    - lane_divider
+    - road_block
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+    aux_data:
+    - visibility
+    - center_offset
+    - center_ohw
+    - height
+    cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+  - type: LoadBEVHDMap
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    image_size:
+    - 224
+    - 400
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: RandomFlip3DwithViews
+    flip_ratio: 0.0
+    direction: null
+  - type: LoadDescription
+    dataset_root: data/nuscenes/
+    dataset_type: Occ3D-nuScenes
+  - type: ReorderMultiViewImagesM
+    order:
+    - CAM_FRONT_LEFT
+    - CAM_FRONT
+    - CAM_FRONT_RIGHT
+    - CAM_BACK_RIGHT
+    - CAM_BACK
+    - CAM_BACK_LEFT
+    safe: false
+  - type: ImageNormalize
+    mean:
+    - 0.5
+    - 0.5
+    - 0.5
+    std:
+    - 0.5
+    - 0.5
+    - 0.5
+  - type: DefaultFormatBundle3D
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: Collect3D
+    keys:
+    - img
+    - gt_bboxes_3d
+    - gt_labels_3d
+    - gt_masks_bev
+    - gt_aux_bev
+    - bev_hdmap
+    - bev_hdmap_w_box
+    - layout_canvas
+    meta_keys:
+    - camera_intrinsics
+    - lidar2ego
+    - lidar2camera
+    - camera2lidar
+    - lidar2image
+    - img_aug_matrix
+    meta_lis_keys:
+    - timeofday
+    - location
+    - description
+    - detailed_description
+    - filename
+    - token
+  test_pipeline:
+  - type: LoadMultiViewImageFromFiles
+    to_float32: true
+  - type: LoadAnnotations3D
+    with_bbox_3d: true
+    with_label_3d: true
+    with_attr_label: false
+  - type: ImageAug3D
+    final_dim:
+    - 224
+    - 400
+    resize_lim:
+    - 0.25
+    - 0.25
+    bot_pct_lim:
+    - 0.0
+    - 0.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    rand_flip: false
+    is_train: false
+  - type: GlobalRotScaleTrans
+    resize_lim:
+    - 1.0
+    - 1.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    trans_lim: 0
+    is_train: true
+  - type: ObjectNameFilterM
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadBEVSegmentationM
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    classes:
+    - drivable_area
+    - ped_crossing
+    - walkway
+    - stop_line
+    - carpark_area
+    - road_divider
+    - lane_divider
+    - road_block
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+    aux_data:
+    - visibility
+    - center_offset
+    - center_ohw
+    - height
+    cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+  - type: LoadBEVHDMap
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    image_size:
+    - 224
+    - 400
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadDescription
+    dataset_root: data/nuscenes/
+    dataset_type: Occ3D-nuScenes
+  - type: ReorderMultiViewImagesM
+    order:
+    - CAM_FRONT_LEFT
+    - CAM_FRONT
+    - CAM_FRONT_RIGHT
+    - CAM_BACK_RIGHT
+    - CAM_BACK
+    - CAM_BACK_LEFT
+    safe: false
+  - type: ImageNormalize
+    mean:
+    - 0.5
+    - 0.5
+    - 0.5
+    std:
+    - 0.5
+    - 0.5
+    - 0.5
+  - type: DefaultFormatBundle3D
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: Collect3D
+    keys:
+    - img
+    - gt_bboxes_3d
+    - gt_labels_3d
+    - gt_masks_bev
+    - gt_aux_bev
+    - bev_hdmap
+    - bev_hdmap_w_box
+    - layout_canvas
+    meta_keys:
+    - camera_intrinsics
+    - lidar2ego
+    - ego2global
+    - lidar2camera
+    - camera2lidar
+    - lidar2image
+    - img_aug_matrix
+    meta_lis_keys:
+    - timeofday
+    - location
+    - description
+    - detailed_description
+    - filename
+    - token
+    - lidar_token
+    - scene_name
+    - timestamp
+  data:
+    train:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_train.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim: null
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: RandomFlip3DwithViews
+        flip_ratio: 0.0
+        direction: null
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: false
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+    val:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_val.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - ego2global
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+        - lidar_token
+        - scene_name
+        - timestamp
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: false
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+    test:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_val.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - ego2global
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+        - lidar_token
+        - scene_name
+        - timestamp
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: true
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+occ_render_path: data/nuscenes/occ_render_map/
+accelerator:
+  gradient_accumulation_steps: 1
+  mixed_precision: fp16
+  report_to: tensorboard
+runner:
+  foreground_loss_weight: 0.0
+  bbox_drop_ratio: 0
+  bbox_add_ratio: 0.1
+  bbox_add_num: 3
+  keyframe_rate: 1
+  num_train_epochs: 115
+  train_batch_size: 10
+  max_train_steps: 80960
+  num_workers: 8
+  prefetch_factor: 4
+  display_per_epoch: 20
+  display_per_n_min: 10
+  max_grad_norm: 1.0
+  set_grads_to_none: true
+  enable_xformers_memory_efficient_attention: true
+  unet_in_fp16: true
+  enable_unet_checkpointing: true
+  enable_controlnet_checkpointing: true
+  noise_offset: 0.0
+  train_with_same_offset: true
+  use_8bit_adam: false
+  adam_beta1: 0.9
+  adam_beta2: 0.999
+  adam_weight_decay: 0.01
+  adam_epsilon: 1.0e-08
+  learning_rate: 8.0e-05
+  lr_scheduler: constant_with_warmup
+  gradient_accumulation_steps: 1
+  lr_num_cycles: 1
+  lr_power: 1.0
+  lr_warmup_steps: 3000
+  checkpointing_steps: 5000
+  validation_steps: 20000
+  save_model_per_epoch: null
+  validation_before_run: false
+  validation_index:
+  - 204
+  - 912
+  - 1828
+  - 2253
+  - 4467
+  - 5543
+  validation_times: 4
+  validation_batch_size: 1
+  validation_show_box: true
+  validation_show_line: true
+  validation_seed_global: false
+  pipeline_param:
+    guidance_scale: 1.2
+    num_inference_steps: 20
+    eta: 0.0
+    controlnet_conditioning_scale: 1.0
+    guess_mode: false
+    use_zero_map_as_unconditional: false
+    bbox_max_length: null
+
+[2025-07-18 17:20:22,769][root][DEBUG] - start!
+[2025-07-18 17:20:22,769][root][INFO] - ***** Running training *****
+[2025-07-18 17:20:22,769][root][INFO] -   Num examples = 28130
+[2025-07-18 17:20:22,769][root][INFO] -   Num batches each epoch = 704
+[2025-07-18 17:20:22,769][root][INFO] -   Num Epochs = 115
+[2025-07-18 17:20:22,769][root][INFO] -   Instantaneous batch size per device = 10
+[2025-07-18 17:20:22,769][root][INFO] -   Total train batch size (w. parallel, distributed & accumulation) = 40
+[2025-07-18 17:20:22,769][root][INFO] -   Gradient Accumulation steps = 1
+[2025-07-18 17:20:22,769][root][INFO] -   Total optimization steps = 80960
+[2025-07-18 17:20:25,715][root][INFO] - Starting from epoch 113 to 115
+[2025-07-18 17:20:39,885][root][WARNING] - [UNet2DConditionModelMultiview] Forward upsample size to force interpolation output size.
+[2025-07-18 18:00:02,390][root][INFO] - Save your model to: /data/yyang/workspace/X-Scene/work_dirs/x-scene-img_224x400/img_unet_2025-07-18_17-19_224x400

train.3.log

@@ -0,0 +1,1504 @@
+[2025-07-18 17:19:48,725][root][INFO] - using data cache from: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+[2025-07-18 17:19:48,958][root][INFO] - [RandomFlip3DwithViews] ratio=0.0, direction=None, reorder=True
+[2025-07-18 17:19:51,636][root][INFO] - using data cache from: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+[2025-07-18 17:20:16,343][root][INFO] - [UNet2DConditionModelMultiview] load pretrained with missing_keys: ['down_blocks.0.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.0.attentions.0.scene_proj_in.weight', 'down_blocks.0.attentions.0.scene_proj_in.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.0.attentions.1.scene_proj_in.weight', 'down_blocks.0.attentions.1.scene_proj_in.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.1.attentions.0.scene_proj_in.weight', 'down_blocks.1.attentions.0.scene_proj_in.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.1.attentions.1.scene_proj_in.weight', 'down_blocks.1.attentions.1.scene_proj_in.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.2.attentions.0.scene_proj_in.weight', 'down_blocks.2.attentions.0.scene_proj_in.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.2.attentions.1.scene_proj_in.weight', 'down_blocks.2.attentions.1.scene_proj_in.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.0.scene_proj_in.weight', 'up_blocks.1.attentions.0.scene_proj_in.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.1.scene_proj_in.weight', 'up_blocks.1.attentions.1.scene_proj_in.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.2.scene_proj_in.weight', 'up_blocks.1.attentions.2.scene_proj_in.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.0.scene_proj_in.weight', 'up_blocks.2.attentions.0.scene_proj_in.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.1.scene_proj_in.weight', 'up_blocks.2.attentions.1.scene_proj_in.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.2.scene_proj_in.weight', 'up_blocks.2.attentions.2.scene_proj_in.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.0.scene_proj_in.weight', 'up_blocks.3.attentions.0.scene_proj_in.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.1.scene_proj_in.weight', 'up_blocks.3.attentions.1.scene_proj_in.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.2.scene_proj_in.weight', 'up_blocks.3.attentions.2.scene_proj_in.bias', 'mid_block.attentions.0.transformer_blocks.0.norm4.weight', 'mid_block.attentions.0.transformer_blocks.0.norm4.bias', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'mid_block.attentions.0.transformer_blocks.0.norm5.weight', 'mid_block.attentions.0.transformer_blocks.0.norm5.bias', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'mid_block.attentions.0.transformer_blocks.0.connector_4.weight', 'mid_block.attentions.0.transformer_blocks.0.connector_4.bias', 'mid_block.attentions.0.transformer_blocks.0.connector_5.weight', 'mid_block.attentions.0.transformer_blocks.0.connector_5.bias', 'mid_block.attentions.0.scene_proj_in.weight', 'mid_block.attentions.0.scene_proj_in.bias']; unexpected_keys: []
+[2025-07-18 17:20:16,344][root][DEBUG] - [BEVControlNetModel] instantiating your own version of controlnet.
+[2025-07-18 17:20:16,345][root][DEBUG] - embedder out dim = 27
+[2025-07-18 17:20:16,357][root][DEBUG] - [BEVControlNetModel] map_embedder: BEVControlNetConditioningEmbedding(
+  (conv_in): Conv2d(4, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+  (blocks): ModuleList(
+    (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (1): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(2, 1))
+    (2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (3): Conv2d(32, 96, kernel_size=(3, 3), stride=(2, 2), padding=(2, 1))
+    (4): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(2, 1))
+    (5): Conv2d(96, 256, kernel_size=(3, 3), stride=(2, 1), padding=(2, 1))
+  )
+  (conv_out): Conv2d(256, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+)
+[2025-07-18 17:20:16,361][root][DEBUG] - [BEVControlNetModel] canvas_embedder: ControlNetConditioningEmbedding(
+  (conv_in): Conv2d(14, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+  (blocks): ModuleList(
+    (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (1): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+    (2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (3): Conv2d(32, 96, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+    (4): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (5): Conv2d(96, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+  )
+  (conv_out): Conv2d(256, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+)
+[2025-07-18 17:20:16,362][root][DEBUG] - embedder out dim = 27
+[2025-07-18 17:20:16,362][root][INFO] - [ContinuousBBoxWithTextEmbedding] bbox embedder has 27 dims.
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,730][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,731][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set position_encoder.0.weight to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set position_encoder.0.bias to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set position_encoder.2.weight to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set position_encoder.2.bias to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set conv_in.weight to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set conv_in.bias to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set blocks.0.weight to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set blocks.0.bias to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set blocks.1.weight to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set blocks.1.bias to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set conv_out.weight to requires_grad = True
+[2025-07-18 17:20:20,732][root][DEBUG] - [MultiviewRunner] set conv_out.bias to requires_grad = True
+[2025-07-18 17:20:21,171][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,171][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,171][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,172][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'diffusers.models.unet_2d_blocks.DownBlock2D'>] to gradient_checkpointing
+[2025-07-18 17:20:21,172][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'diffusers.models.unet_2d_blocks.UpBlock2D'>] to gradient_checkpointing
+[2025-07-18 17:20:21,172][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,172][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,172][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,174][root][INFO] - [BaseValidator] Validator use model_param: dict_keys(['vae', 'text_encoder', 'text_encoder_t5', 'tokenizer', 'tokenizer_t5'])
+[2025-07-18 17:20:21,175][root][INFO] - [MultiviewRunner] add 130.82 M params from unet to optimizer.
+[2025-07-18 17:20:21,175][root][INFO] - [MultiviewRunner] have total 525.45 M params from unet and controlnet to optimizer.
+[2025-07-18 17:20:21,175][root][INFO] - [MultiviewRunner] add 2.19 M params from scene_embedder to optimizer.
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,569][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,570][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,571][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,572][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,573][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,574][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,575][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,576][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,576][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,576][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,576][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,576][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,576][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,576][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,576][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,576][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,577][root][INFO] - [ContinuousBBoxWithTextEmbedding] Initialzing your class_tokens with text_encoder
+[2025-07-18 17:20:22,773][root][DEBUG] - Current config:
+task_id: 224x400
+log_root_prefix: ./work_dirs/x-scene-img_224x400
+projname: img_unet
+try_run: false
+debug: false
+log_root: /data/yyang/workspace/X-Scene/work_dirs/x-scene-img_224x400/img_unet_2025-07-18_17-19_224x400
+init_method: env://
+seed: 42
+fix_seed_within_batch: false
+resume_from_checkpoint: work_dirs/x-scene-img_224x400/img_unet_2025-07-16_10-33_224x400/checkpoint-80000
+resume_only_model: false
+resume_reset_scheduler: false
+validation_only: false
+model:
+  name: img_unet
+  pretrained_model_name_or_path: pretrained/stable-diffusion-v2-1/
+  pretrained_t5_path: pretrained/t5-large/
+  bbox_mode: all-xyz
+  bbox_view_shared: false
+  crossview_attn_type: t5_crossview
+  train_with_same_noise: false
+  train_with_same_t: true
+  runner_module: xscene.runner.multiview_runner.MultiviewRunner
+  pipe_module: xscene.pipeline.pipeline_bev_controlnet.StableDiffusionBEVControlNetPipeline
+  unet_module: xscene.networks.unet_2d_condition_multiview.UNet2DConditionModelMultiview
+  use_fp32_for_unet_trainable: true
+  unet_dir: unet
+  unet:
+    trainable_state: only_new
+    neighboring_view_pair:
+      0:
+      - 5
+      - 1
+      1:
+      - 0
+      - 2
+      2:
+      - 1
+      - 3
+      3:
+      - 2
+      - 4
+      4:
+      - 3
+      - 5
+      5:
+      - 4
+      - 0
+    neighboring_attn_type: add
+    zero_module_type: zero_linear
+    crossview_attn_type: t5_crossview
+    img_size:
+    - 224
+    - 400
+    scene_channels: 320
+    attn1_q_trainable: true
+  scene_embedder_cls: xscene.networks.scene_position_embedder.ScenePositionEmbedding
+  scene_embedder_dir: scene_embedder
+  scene_embedder:
+    embed_dims: 320
+    LID: false
+  model_module: xscene.networks.unet_addon_rawbox.BEVControlNetModel
+  controlnet_dir: controlnet
+  controlnet:
+    camera_in_dim: 189
+    camera_out_dim: 1024
+    map_size:
+    - 4
+    - 200
+    - 200
+    conditioning_embedding_out_channels:
+    - 16
+    - 32
+    - 96
+    - 256
+    uncond_cam_in_dim:
+    - 3
+    - 7
+    use_uncond_map: null
+    drop_cond_ratio: 0.25
+    drop_cam_num: 6
+    drop_cam_with_box: false
+    cam_embedder_param:
+      input_dims: 3
+      num_freqs: 4
+      include_input: true
+      log_sampling: true
+    bbox_embedder_cls: xscene.networks.bbox_embedder.ContinuousBBoxWithTextEmbedding
+    bbox_embedder_param:
+      n_classes: 10
+      class_token_dim: 1024
+      trainable_class_token: false
+      use_text_encoder_init: true
+      embedder_num_freq: 4
+      proj_dims:
+      - 1024
+      - 512
+      - 512
+      - 1024
+      mode: all-xyz
+      minmax_normalize: false
+    with_layout_canvas: true
+    canvas_conditioning_channels: 14
+    canvas_size:
+    - 14
+    - 224
+    - 400
+    with_occ_render_img: false
+    occrender_conditioning_channels: 20
+    render_img_size:
+    - 20
+    - 224
+    - 400
+    occrender_embedding_out_channels:
+    - 16
+    - 32
+    - 64
+    - 96
+    - 256
+dataset:
+  dataset_type: NuScenesDatasetM
+  occ_dataset_type: Occ3D-nuScenes
+  dataset_root: data/nuscenes/
+  triplane_root: data/nuscenes/nuscenes_triplane
+  dataset_process_root: data/nuscenes/nuscenes_mmdet3d-keyframes/
+  dataset_cache_file_tag: 200x200_12Hz_interp
+  dataset_cache_file:
+  - data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+  - data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+  template_clip: A driving scene image at {location}. {description}.
+  template_t5: A driving scene at {location}. {description}. {detailed_description}
+  image_size:
+  - 224
+  - 400
+  map_bound:
+    x:
+    - -40.0
+    - 40.0
+    - 0.4
+    'y':
+    - -40.0
+    - 40.0
+    - 0.4
+    z:
+    - -1.0
+    - 5.4
+    - 0.4
+  tri_size:
+  - 100
+  - 100
+  - 16
+  view_order:
+  - CAM_FRONT_LEFT
+  - CAM_FRONT
+  - CAM_FRONT_RIGHT
+  - CAM_BACK_RIGHT
+  - CAM_BACK
+  - CAM_BACK_LEFT
+  neighboring_view_pair:
+    0:
+    - 5
+    - 1
+    1:
+    - 0
+    - 2
+    2:
+    - 1
+    - 3
+    3:
+    - 2
+    - 4
+    4:
+    - 3
+    - 5
+    5:
+    - 4
+    - 0
+  back_resize:
+  - 896
+  - 1600
+  back_pad:
+  - 0
+  - 4
+  - 0
+  - 0
+  augment2d:
+    resize:
+    - - 0.25
+      - 0.25
+    rotate: null
+  aux_data:
+  - visibility
+  - center_offset
+  - center_ohw
+  - height
+  augment3d:
+    scale:
+    - 1.0
+    - 1.0
+    rotate:
+    - 0.0
+    - 0.0
+    translate: 0
+    flip_ratio: 0.0
+    flip_direction: null
+  object_classes:
+  - barrier
+  - bicycle
+  - bus
+  - car
+  - construction_vehicle
+  - motorcycle
+  - pedestrian
+  - traffic_cone
+  - trailer
+  - truck
+  map_classes:
+  - drivable_area
+  - ped_crossing
+  - walkway
+  - stop_line
+  - carpark_area
+  - road_divider
+  - lane_divider
+  - road_block
+  input_modality:
+    use_lidar: false
+    use_camera: true
+    use_radar: false
+    use_map: false
+    use_external: false
+  train_pipeline:
+  - type: LoadMultiViewImageFromFiles
+    to_float32: true
+  - type: LoadAnnotations3D
+    with_bbox_3d: true
+    with_label_3d: true
+    with_attr_label: false
+  - type: ImageAug3D
+    final_dim:
+    - 224
+    - 400
+    resize_lim:
+    - 0.25
+    - 0.25
+    bot_pct_lim:
+    - 0.0
+    - 0.0
+    rot_lim: null
+    rand_flip: false
+    is_train: false
+  - type: GlobalRotScaleTrans
+    resize_lim:
+    - 1.0
+    - 1.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    trans_lim: 0
+    is_train: true
+  - type: ObjectNameFilterM
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadBEVSegmentationM
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    classes:
+    - drivable_area
+    - ped_crossing
+    - walkway
+    - stop_line
+    - carpark_area
+    - road_divider
+    - lane_divider
+    - road_block
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+    aux_data:
+    - visibility
+    - center_offset
+    - center_ohw
+    - height
+    cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+  - type: LoadBEVHDMap
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    image_size:
+    - 224
+    - 400
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: RandomFlip3DwithViews
+    flip_ratio: 0.0
+    direction: null
+  - type: LoadDescription
+    dataset_root: data/nuscenes/
+    dataset_type: Occ3D-nuScenes
+  - type: ReorderMultiViewImagesM
+    order:
+    - CAM_FRONT_LEFT
+    - CAM_FRONT
+    - CAM_FRONT_RIGHT
+    - CAM_BACK_RIGHT
+    - CAM_BACK
+    - CAM_BACK_LEFT
+    safe: false
+  - type: ImageNormalize
+    mean:
+    - 0.5
+    - 0.5
+    - 0.5
+    std:
+    - 0.5
+    - 0.5
+    - 0.5
+  - type: DefaultFormatBundle3D
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: Collect3D
+    keys:
+    - img
+    - gt_bboxes_3d
+    - gt_labels_3d
+    - gt_masks_bev
+    - gt_aux_bev
+    - bev_hdmap
+    - bev_hdmap_w_box
+    - layout_canvas
+    meta_keys:
+    - camera_intrinsics
+    - lidar2ego
+    - lidar2camera
+    - camera2lidar
+    - lidar2image
+    - img_aug_matrix
+    meta_lis_keys:
+    - timeofday
+    - location
+    - description
+    - detailed_description
+    - filename
+    - token
+  test_pipeline:
+  - type: LoadMultiViewImageFromFiles
+    to_float32: true
+  - type: LoadAnnotations3D
+    with_bbox_3d: true
+    with_label_3d: true
+    with_attr_label: false
+  - type: ImageAug3D
+    final_dim:
+    - 224
+    - 400
+    resize_lim:
+    - 0.25
+    - 0.25
+    bot_pct_lim:
+    - 0.0
+    - 0.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    rand_flip: false
+    is_train: false
+  - type: GlobalRotScaleTrans
+    resize_lim:
+    - 1.0
+    - 1.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    trans_lim: 0
+    is_train: true
+  - type: ObjectNameFilterM
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadBEVSegmentationM
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    classes:
+    - drivable_area
+    - ped_crossing
+    - walkway
+    - stop_line
+    - carpark_area
+    - road_divider
+    - lane_divider
+    - road_block
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+    aux_data:
+    - visibility
+    - center_offset
+    - center_ohw
+    - height
+    cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+  - type: LoadBEVHDMap
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    image_size:
+    - 224
+    - 400
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadDescription
+    dataset_root: data/nuscenes/
+    dataset_type: Occ3D-nuScenes
+  - type: ReorderMultiViewImagesM
+    order:
+    - CAM_FRONT_LEFT
+    - CAM_FRONT
+    - CAM_FRONT_RIGHT
+    - CAM_BACK_RIGHT
+    - CAM_BACK
+    - CAM_BACK_LEFT
+    safe: false
+  - type: ImageNormalize
+    mean:
+    - 0.5
+    - 0.5
+    - 0.5
+    std:
+    - 0.5
+    - 0.5
+    - 0.5
+  - type: DefaultFormatBundle3D
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: Collect3D
+    keys:
+    - img
+    - gt_bboxes_3d
+    - gt_labels_3d
+    - gt_masks_bev
+    - gt_aux_bev
+    - bev_hdmap
+    - bev_hdmap_w_box
+    - layout_canvas
+    meta_keys:
+    - camera_intrinsics
+    - lidar2ego
+    - ego2global
+    - lidar2camera
+    - camera2lidar
+    - lidar2image
+    - img_aug_matrix
+    meta_lis_keys:
+    - timeofday
+    - location
+    - description
+    - detailed_description
+    - filename
+    - token
+    - lidar_token
+    - scene_name
+    - timestamp
+  data:
+    train:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_train.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim: null
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: RandomFlip3DwithViews
+        flip_ratio: 0.0
+        direction: null
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: false
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+    val:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_val.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - ego2global
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+        - lidar_token
+        - scene_name
+        - timestamp
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: false
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+    test:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_val.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - ego2global
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+        - lidar_token
+        - scene_name
+        - timestamp
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: true
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+occ_render_path: data/nuscenes/occ_render_map/
+accelerator:
+  gradient_accumulation_steps: 1
+  mixed_precision: fp16
+  report_to: tensorboard
+runner:
+  foreground_loss_weight: 0.0
+  bbox_drop_ratio: 0
+  bbox_add_ratio: 0.1
+  bbox_add_num: 3
+  keyframe_rate: 1
+  num_train_epochs: 115
+  train_batch_size: 10
+  max_train_steps: 80960
+  num_workers: 8
+  prefetch_factor: 4
+  display_per_epoch: 20
+  display_per_n_min: 10
+  max_grad_norm: 1.0
+  set_grads_to_none: true
+  enable_xformers_memory_efficient_attention: true
+  unet_in_fp16: true
+  enable_unet_checkpointing: true
+  enable_controlnet_checkpointing: true
+  noise_offset: 0.0
+  train_with_same_offset: true
+  use_8bit_adam: false
+  adam_beta1: 0.9
+  adam_beta2: 0.999
+  adam_weight_decay: 0.01
+  adam_epsilon: 1.0e-08
+  learning_rate: 8.0e-05
+  lr_scheduler: constant_with_warmup
+  gradient_accumulation_steps: 1
+  lr_num_cycles: 1
+  lr_power: 1.0
+  lr_warmup_steps: 3000
+  checkpointing_steps: 5000
+  validation_steps: 20000
+  save_model_per_epoch: null
+  validation_before_run: false
+  validation_index:
+  - 204
+  - 912
+  - 1828
+  - 2253
+  - 4467
+  - 5543
+  validation_times: 4
+  validation_batch_size: 1
+  validation_show_box: true
+  validation_show_line: true
+  validation_seed_global: false
+  pipeline_param:
+    guidance_scale: 1.2
+    num_inference_steps: 20
+    eta: 0.0
+    controlnet_conditioning_scale: 1.0
+    guess_mode: false
+    use_zero_map_as_unconditional: false
+    bbox_max_length: null
+
+[2025-07-18 17:20:22,773][root][DEBUG] - start!
+[2025-07-18 17:20:22,773][root][INFO] - ***** Running training *****
+[2025-07-18 17:20:22,773][root][INFO] -   Num examples = 28130
+[2025-07-18 17:20:22,774][root][INFO] -   Num batches each epoch = 704
+[2025-07-18 17:20:22,774][root][INFO] -   Num Epochs = 115
+[2025-07-18 17:20:22,774][root][INFO] -   Instantaneous batch size per device = 10
+[2025-07-18 17:20:22,774][root][INFO] -   Total train batch size (w. parallel, distributed & accumulation) = 40
+[2025-07-18 17:20:22,774][root][INFO] -   Gradient Accumulation steps = 1
+[2025-07-18 17:20:22,774][root][INFO] -   Total optimization steps = 80960
+[2025-07-18 17:20:25,287][root][INFO] - Starting from epoch 113 to 115
+[2025-07-18 17:20:39,875][root][WARNING] - [UNet2DConditionModelMultiview] Forward upsample size to force interpolation output size.
+[2025-07-18 18:00:02,390][root][INFO] - Save your model to: /data/yyang/workspace/X-Scene/work_dirs/x-scene-img_224x400/img_unet_2025-07-18_17-19_224x400

train.log

@@ -0,0 +1,1522 @@
+[2025-07-18 17:19:39,858][torch.distributed.distributed_c10d][INFO] - Added key: store_based_barrier_key:1 to store for rank: 2
+[2025-07-18 17:19:39,859][torch.distributed.distributed_c10d][INFO] - Added key: store_based_barrier_key:1 to store for rank: 3
+[2025-07-18 17:19:39,859][torch.distributed.distributed_c10d][INFO] - Added key: store_based_barrier_key:1 to store for rank: 0
+[2025-07-18 17:19:39,859][torch.distributed.distributed_c10d][INFO] - Rank 3: Completed store-based barrier for key:store_based_barrier_key:1 with 4 nodes.
+[2025-07-18 17:19:39,859][torch.distributed.distributed_c10d][INFO] - Added key: store_based_barrier_key:1 to store for rank: 1
+[2025-07-18 17:19:39,859][torch.distributed.distributed_c10d][INFO] - Rank 0: Completed store-based barrier for key:store_based_barrier_key:1 with 4 nodes.
+[2025-07-18 17:19:39,859][torch.distributed.distributed_c10d][INFO] - Rank 1: Completed store-based barrier for key:store_based_barrier_key:1 with 4 nodes.
+[2025-07-18 17:19:39,868][torch.distributed.distributed_c10d][INFO] - Rank 2: Completed store-based barrier for key:store_based_barrier_key:1 with 4 nodes.
+[2025-07-18 17:19:39,929][root][INFO] - reset logger for 1
+[2025-07-18 17:19:39,943][root][INFO] - reset logger for 3
+[2025-07-18 17:19:39,952][root][INFO] - reset logger for 2
+[2025-07-18 17:19:48,699][root][INFO] - using data cache from: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+[2025-07-18 17:19:48,932][root][INFO] - [RandomFlip3DwithViews] ratio=0.0, direction=None, reorder=True
+[2025-07-18 17:19:51,589][root][INFO] - using data cache from: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+[2025-07-18 17:20:16,306][root][INFO] - [UNet2DConditionModelMultiview] load pretrained with missing_keys: ['down_blocks.0.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.0.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.0.attentions.0.scene_proj_in.weight', 'down_blocks.0.attentions.0.scene_proj_in.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.0.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.0.attentions.1.scene_proj_in.weight', 'down_blocks.0.attentions.1.scene_proj_in.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.1.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.1.attentions.0.scene_proj_in.weight', 'down_blocks.1.attentions.0.scene_proj_in.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.1.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.1.attentions.1.scene_proj_in.weight', 'down_blocks.1.attentions.1.scene_proj_in.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.norm4.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.norm4.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.norm5.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.norm5.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_4.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_4.bias', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_5.weight', 'down_blocks.2.attentions.0.transformer_blocks.0.connector_5.bias', 'down_blocks.2.attentions.0.scene_proj_in.weight', 'down_blocks.2.attentions.0.scene_proj_in.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.norm4.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.norm4.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.norm5.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.norm5.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_4.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_4.bias', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_5.weight', 'down_blocks.2.attentions.1.transformer_blocks.0.connector_5.bias', 'down_blocks.2.attentions.1.scene_proj_in.weight', 'down_blocks.2.attentions.1.scene_proj_in.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.0.scene_proj_in.weight', 'up_blocks.1.attentions.0.scene_proj_in.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.1.scene_proj_in.weight', 'up_blocks.1.attentions.1.scene_proj_in.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.1.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.1.attentions.2.scene_proj_in.weight', 'up_blocks.1.attentions.2.scene_proj_in.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.0.scene_proj_in.weight', 'up_blocks.2.attentions.0.scene_proj_in.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.1.scene_proj_in.weight', 'up_blocks.2.attentions.1.scene_proj_in.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.2.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.2.attentions.2.scene_proj_in.weight', 'up_blocks.2.attentions.2.scene_proj_in.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.0.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.0.scene_proj_in.weight', 'up_blocks.3.attentions.0.scene_proj_in.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.1.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.1.scene_proj_in.weight', 'up_blocks.3.attentions.1.scene_proj_in.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.norm4.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.norm4.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_q.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_k.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_v.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_out.0.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn4.to_out.0.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.norm5.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.norm5.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_q.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_k.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_v.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_out.0.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.attn5.to_out.0.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_4.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_4.bias', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_5.weight', 'up_blocks.3.attentions.2.transformer_blocks.0.connector_5.bias', 'up_blocks.3.attentions.2.scene_proj_in.weight', 'up_blocks.3.attentions.2.scene_proj_in.bias', 'mid_block.attentions.0.transformer_blocks.0.norm4.weight', 'mid_block.attentions.0.transformer_blocks.0.norm4.bias', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_q.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_k.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_v.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_out.0.weight', 'mid_block.attentions.0.transformer_blocks.0.attn4.to_out.0.bias', 'mid_block.attentions.0.transformer_blocks.0.norm5.weight', 'mid_block.attentions.0.transformer_blocks.0.norm5.bias', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_q.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_k.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_v.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_out.0.weight', 'mid_block.attentions.0.transformer_blocks.0.attn5.to_out.0.bias', 'mid_block.attentions.0.transformer_blocks.0.connector_4.weight', 'mid_block.attentions.0.transformer_blocks.0.connector_4.bias', 'mid_block.attentions.0.transformer_blocks.0.connector_5.weight', 'mid_block.attentions.0.transformer_blocks.0.connector_5.bias', 'mid_block.attentions.0.scene_proj_in.weight', 'mid_block.attentions.0.scene_proj_in.bias']; unexpected_keys: []
+[2025-07-18 17:20:16,309][root][DEBUG] - [BEVControlNetModel] instantiating your own version of controlnet.
+[2025-07-18 17:20:16,310][root][DEBUG] - embedder out dim = 27
+[2025-07-18 17:20:16,321][root][DEBUG] - [BEVControlNetModel] map_embedder: BEVControlNetConditioningEmbedding(
+  (conv_in): Conv2d(4, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+  (blocks): ModuleList(
+    (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (1): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(2, 1))
+    (2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (3): Conv2d(32, 96, kernel_size=(3, 3), stride=(2, 2), padding=(2, 1))
+    (4): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(2, 1))
+    (5): Conv2d(96, 256, kernel_size=(3, 3), stride=(2, 1), padding=(2, 1))
+  )
+  (conv_out): Conv2d(256, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+)
+[2025-07-18 17:20:16,326][root][DEBUG] - [BEVControlNetModel] canvas_embedder: ControlNetConditioningEmbedding(
+  (conv_in): Conv2d(14, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+  (blocks): ModuleList(
+    (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (1): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+    (2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (3): Conv2d(32, 96, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+    (4): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+    (5): Conv2d(96, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
+  )
+  (conv_out): Conv2d(256, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+)
+[2025-07-18 17:20:16,328][root][DEBUG] - embedder out dim = 27
+[2025-07-18 17:20:16,328][root][INFO] - [ContinuousBBoxWithTextEmbedding] bbox embedder has 27 dims.
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,674][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,675][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_4 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_5 to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn1.to_q to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set position_encoder.0.weight to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set position_encoder.0.bias to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set position_encoder.2.weight to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set position_encoder.2.bias to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set conv_in.weight to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set conv_in.bias to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set blocks.0.weight to requires_grad = True
+[2025-07-18 17:20:20,676][root][DEBUG] - [MultiviewRunner] set blocks.0.bias to requires_grad = True
+[2025-07-18 17:20:20,677][root][DEBUG] - [MultiviewRunner] set blocks.1.weight to requires_grad = True
+[2025-07-18 17:20:20,677][root][DEBUG] - [MultiviewRunner] set blocks.1.bias to requires_grad = True
+[2025-07-18 17:20:20,677][root][DEBUG] - [MultiviewRunner] set conv_out.weight to requires_grad = True
+[2025-07-18 17:20:20,677][root][DEBUG] - [MultiviewRunner] set conv_out.bias to requires_grad = True
+[2025-07-18 17:20:21,094][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,095][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,095][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnDownBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,095][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'diffusers.models.unet_2d_blocks.DownBlock2D'>] to gradient_checkpointing
+[2025-07-18 17:20:21,095][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'diffusers.models.unet_2d_blocks.UpBlock2D'>] to gradient_checkpointing
+[2025-07-18 17:20:21,095][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,095][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,095][root][DEBUG] - [UNet2DConditionModelMultiview] set [<class 'xscene.networks.blocks.CrossAttnUpBlock2DT5'>] to gradient_checkpointing
+[2025-07-18 17:20:21,097][root][INFO] - [BaseValidator] Validator use model_param: dict_keys(['vae', 'text_encoder', 'text_encoder_t5', 'tokenizer', 'tokenizer_t5'])
+[2025-07-18 17:20:21,098][root][INFO] - [MultiviewRunner] add 130.82 M params from unet to optimizer.
+[2025-07-18 17:20:21,098][root][INFO] - [MultiviewRunner] have total 525.45 M params from unet and controlnet to optimizer.
+[2025-07-18 17:20:21,098][root][INFO] - [MultiviewRunner] add 2.19 M params from scene_embedder to optimizer.
+[2025-07-18 17:20:22,549][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,549][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,549][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,549][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.0.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,550][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,551][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set down_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,552][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.1.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,553][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.2.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,554][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.1.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set up_blocks.3.attentions.2.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm4 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn4 to fp32
+[2025-07-18 17:20:22,555][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.norm5 to fp32
+[2025-07-18 17:20:22,556][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn5 to fp32
+[2025-07-18 17:20:22,556][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_4 to fp32
+[2025-07-18 17:20:22,556][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.connector_5 to fp32
+[2025-07-18 17:20:22,556][root][DEBUG] - [MultiviewRunner] set mid_block.attentions.0.transformer_blocks.0.attn1.to_q to fp32
+[2025-07-18 17:20:22,557][root][INFO] - [ContinuousBBoxWithTextEmbedding] Initialzing your class_tokens with text_encoder
+[2025-07-18 17:20:22,756][root][DEBUG] - Current config:
+task_id: 224x400
+log_root_prefix: ./work_dirs/x-scene-img_224x400
+projname: img_unet
+try_run: false
+debug: false
+log_root: /data/yyang/workspace/X-Scene/work_dirs/x-scene-img_224x400/img_unet_2025-07-18_17-19_224x400
+init_method: env://
+seed: 42
+fix_seed_within_batch: false
+resume_from_checkpoint: work_dirs/x-scene-img_224x400/img_unet_2025-07-16_10-33_224x400/checkpoint-80000
+resume_only_model: false
+resume_reset_scheduler: false
+validation_only: false
+model:
+  name: img_unet
+  pretrained_model_name_or_path: pretrained/stable-diffusion-v2-1/
+  pretrained_t5_path: pretrained/t5-large/
+  bbox_mode: all-xyz
+  bbox_view_shared: false
+  crossview_attn_type: t5_crossview
+  train_with_same_noise: false
+  train_with_same_t: true
+  runner_module: xscene.runner.multiview_runner.MultiviewRunner
+  pipe_module: xscene.pipeline.pipeline_bev_controlnet.StableDiffusionBEVControlNetPipeline
+  unet_module: xscene.networks.unet_2d_condition_multiview.UNet2DConditionModelMultiview
+  use_fp32_for_unet_trainable: true
+  unet_dir: unet
+  unet:
+    trainable_state: only_new
+    neighboring_view_pair:
+      0:
+      - 5
+      - 1
+      1:
+      - 0
+      - 2
+      2:
+      - 1
+      - 3
+      3:
+      - 2
+      - 4
+      4:
+      - 3
+      - 5
+      5:
+      - 4
+      - 0
+    neighboring_attn_type: add
+    zero_module_type: zero_linear
+    crossview_attn_type: t5_crossview
+    img_size:
+    - 224
+    - 400
+    scene_channels: 320
+    attn1_q_trainable: true
+  scene_embedder_cls: xscene.networks.scene_position_embedder.ScenePositionEmbedding
+  scene_embedder_dir: scene_embedder
+  scene_embedder:
+    embed_dims: 320
+    LID: false
+  model_module: xscene.networks.unet_addon_rawbox.BEVControlNetModel
+  controlnet_dir: controlnet
+  controlnet:
+    camera_in_dim: 189
+    camera_out_dim: 1024
+    map_size:
+    - 4
+    - 200
+    - 200
+    conditioning_embedding_out_channels:
+    - 16
+    - 32
+    - 96
+    - 256
+    uncond_cam_in_dim:
+    - 3
+    - 7
+    use_uncond_map: null
+    drop_cond_ratio: 0.25
+    drop_cam_num: 6
+    drop_cam_with_box: false
+    cam_embedder_param:
+      input_dims: 3
+      num_freqs: 4
+      include_input: true
+      log_sampling: true
+    bbox_embedder_cls: xscene.networks.bbox_embedder.ContinuousBBoxWithTextEmbedding
+    bbox_embedder_param:
+      n_classes: 10
+      class_token_dim: 1024
+      trainable_class_token: false
+      use_text_encoder_init: true
+      embedder_num_freq: 4
+      proj_dims:
+      - 1024
+      - 512
+      - 512
+      - 1024
+      mode: all-xyz
+      minmax_normalize: false
+    with_layout_canvas: true
+    canvas_conditioning_channels: 14
+    canvas_size:
+    - 14
+    - 224
+    - 400
+    with_occ_render_img: false
+    occrender_conditioning_channels: 20
+    render_img_size:
+    - 20
+    - 224
+    - 400
+    occrender_embedding_out_channels:
+    - 16
+    - 32
+    - 64
+    - 96
+    - 256
+dataset:
+  dataset_type: NuScenesDatasetM
+  occ_dataset_type: Occ3D-nuScenes
+  dataset_root: data/nuscenes/
+  triplane_root: data/nuscenes/nuscenes_triplane
+  dataset_process_root: data/nuscenes/nuscenes_mmdet3d-keyframes/
+  dataset_cache_file_tag: 200x200_12Hz_interp
+  dataset_cache_file:
+  - data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+  - data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+  template_clip: A driving scene image at {location}. {description}.
+  template_t5: A driving scene at {location}. {description}. {detailed_description}
+  image_size:
+  - 224
+  - 400
+  map_bound:
+    x:
+    - -40.0
+    - 40.0
+    - 0.4
+    'y':
+    - -40.0
+    - 40.0
+    - 0.4
+    z:
+    - -1.0
+    - 5.4
+    - 0.4
+  tri_size:
+  - 100
+  - 100
+  - 16
+  view_order:
+  - CAM_FRONT_LEFT
+  - CAM_FRONT
+  - CAM_FRONT_RIGHT
+  - CAM_BACK_RIGHT
+  - CAM_BACK
+  - CAM_BACK_LEFT
+  neighboring_view_pair:
+    0:
+    - 5
+    - 1
+    1:
+    - 0
+    - 2
+    2:
+    - 1
+    - 3
+    3:
+    - 2
+    - 4
+    4:
+    - 3
+    - 5
+    5:
+    - 4
+    - 0
+  back_resize:
+  - 896
+  - 1600
+  back_pad:
+  - 0
+  - 4
+  - 0
+  - 0
+  augment2d:
+    resize:
+    - - 0.25
+      - 0.25
+    rotate: null
+  aux_data:
+  - visibility
+  - center_offset
+  - center_ohw
+  - height
+  augment3d:
+    scale:
+    - 1.0
+    - 1.0
+    rotate:
+    - 0.0
+    - 0.0
+    translate: 0
+    flip_ratio: 0.0
+    flip_direction: null
+  object_classes:
+  - barrier
+  - bicycle
+  - bus
+  - car
+  - construction_vehicle
+  - motorcycle
+  - pedestrian
+  - traffic_cone
+  - trailer
+  - truck
+  map_classes:
+  - drivable_area
+  - ped_crossing
+  - walkway
+  - stop_line
+  - carpark_area
+  - road_divider
+  - lane_divider
+  - road_block
+  input_modality:
+    use_lidar: false
+    use_camera: true
+    use_radar: false
+    use_map: false
+    use_external: false
+  train_pipeline:
+  - type: LoadMultiViewImageFromFiles
+    to_float32: true
+  - type: LoadAnnotations3D
+    with_bbox_3d: true
+    with_label_3d: true
+    with_attr_label: false
+  - type: ImageAug3D
+    final_dim:
+    - 224
+    - 400
+    resize_lim:
+    - 0.25
+    - 0.25
+    bot_pct_lim:
+    - 0.0
+    - 0.0
+    rot_lim: null
+    rand_flip: false
+    is_train: false
+  - type: GlobalRotScaleTrans
+    resize_lim:
+    - 1.0
+    - 1.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    trans_lim: 0
+    is_train: true
+  - type: ObjectNameFilterM
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadBEVSegmentationM
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    classes:
+    - drivable_area
+    - ped_crossing
+    - walkway
+    - stop_line
+    - carpark_area
+    - road_divider
+    - lane_divider
+    - road_block
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+    aux_data:
+    - visibility
+    - center_offset
+    - center_ohw
+    - height
+    cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+  - type: LoadBEVHDMap
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    image_size:
+    - 224
+    - 400
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: RandomFlip3DwithViews
+    flip_ratio: 0.0
+    direction: null
+  - type: LoadDescription
+    dataset_root: data/nuscenes/
+    dataset_type: Occ3D-nuScenes
+  - type: ReorderMultiViewImagesM
+    order:
+    - CAM_FRONT_LEFT
+    - CAM_FRONT
+    - CAM_FRONT_RIGHT
+    - CAM_BACK_RIGHT
+    - CAM_BACK
+    - CAM_BACK_LEFT
+    safe: false
+  - type: ImageNormalize
+    mean:
+    - 0.5
+    - 0.5
+    - 0.5
+    std:
+    - 0.5
+    - 0.5
+    - 0.5
+  - type: DefaultFormatBundle3D
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: Collect3D
+    keys:
+    - img
+    - gt_bboxes_3d
+    - gt_labels_3d
+    - gt_masks_bev
+    - gt_aux_bev
+    - bev_hdmap
+    - bev_hdmap_w_box
+    - layout_canvas
+    meta_keys:
+    - camera_intrinsics
+    - lidar2ego
+    - lidar2camera
+    - camera2lidar
+    - lidar2image
+    - img_aug_matrix
+    meta_lis_keys:
+    - timeofday
+    - location
+    - description
+    - detailed_description
+    - filename
+    - token
+  test_pipeline:
+  - type: LoadMultiViewImageFromFiles
+    to_float32: true
+  - type: LoadAnnotations3D
+    with_bbox_3d: true
+    with_label_3d: true
+    with_attr_label: false
+  - type: ImageAug3D
+    final_dim:
+    - 224
+    - 400
+    resize_lim:
+    - 0.25
+    - 0.25
+    bot_pct_lim:
+    - 0.0
+    - 0.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    rand_flip: false
+    is_train: false
+  - type: GlobalRotScaleTrans
+    resize_lim:
+    - 1.0
+    - 1.0
+    rot_lim:
+    - 0.0
+    - 0.0
+    trans_lim: 0
+    is_train: true
+  - type: ObjectNameFilterM
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadBEVSegmentationM
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    classes:
+    - drivable_area
+    - ped_crossing
+    - walkway
+    - stop_line
+    - carpark_area
+    - road_divider
+    - lane_divider
+    - road_block
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+    aux_data:
+    - visibility
+    - center_offset
+    - center_ohw
+    - height
+    cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+  - type: LoadBEVHDMap
+    dataset_root: data/nuscenes/
+    xbound:
+    - -40.0
+    - 40.0
+    - 0.4
+    ybound:
+    - -40.0
+    - 40.0
+    - 0.4
+    image_size:
+    - 224
+    - 400
+    object_classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: LoadDescription
+    dataset_root: data/nuscenes/
+    dataset_type: Occ3D-nuScenes
+  - type: ReorderMultiViewImagesM
+    order:
+    - CAM_FRONT_LEFT
+    - CAM_FRONT
+    - CAM_FRONT_RIGHT
+    - CAM_BACK_RIGHT
+    - CAM_BACK
+    - CAM_BACK_LEFT
+    safe: false
+  - type: ImageNormalize
+    mean:
+    - 0.5
+    - 0.5
+    - 0.5
+    std:
+    - 0.5
+    - 0.5
+    - 0.5
+  - type: DefaultFormatBundle3D
+    classes:
+    - barrier
+    - bicycle
+    - bus
+    - car
+    - construction_vehicle
+    - motorcycle
+    - pedestrian
+    - traffic_cone
+    - trailer
+    - truck
+  - type: Collect3D
+    keys:
+    - img
+    - gt_bboxes_3d
+    - gt_labels_3d
+    - gt_masks_bev
+    - gt_aux_bev
+    - bev_hdmap
+    - bev_hdmap_w_box
+    - layout_canvas
+    meta_keys:
+    - camera_intrinsics
+    - lidar2ego
+    - ego2global
+    - lidar2camera
+    - camera2lidar
+    - lidar2image
+    - img_aug_matrix
+    meta_lis_keys:
+    - timeofday
+    - location
+    - description
+    - detailed_description
+    - filename
+    - token
+    - lidar_token
+    - scene_name
+    - timestamp
+  data:
+    train:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_train.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim: null
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/train_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: RandomFlip3DwithViews
+        flip_ratio: 0.0
+        direction: null
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: false
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+    val:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_val.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - ego2global
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+        - lidar_token
+        - scene_name
+        - timestamp
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: false
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+    test:
+      type: NuScenesDatasetM
+      dataset_root: data/nuscenes/
+      ann_file: data/nuscenes/nuscenes_mmdet3d-keyframes/nuscenes_infos_val.pkl
+      pipeline:
+      - type: LoadMultiViewImageFromFiles
+        to_float32: true
+      - type: LoadAnnotations3D
+        with_bbox_3d: true
+        with_label_3d: true
+        with_attr_label: false
+      - type: ImageAug3D
+        final_dim:
+        - 224
+        - 400
+        resize_lim:
+        - 0.25
+        - 0.25
+        bot_pct_lim:
+        - 0.0
+        - 0.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        rand_flip: false
+        is_train: false
+      - type: GlobalRotScaleTrans
+        resize_lim:
+        - 1.0
+        - 1.0
+        rot_lim:
+        - 0.0
+        - 0.0
+        trans_lim: 0
+        is_train: true
+      - type: ObjectNameFilterM
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadBEVSegmentationM
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        classes:
+        - drivable_area
+        - ped_crossing
+        - walkway
+        - stop_line
+        - carpark_area
+        - road_divider
+        - lane_divider
+        - road_block
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+        aux_data:
+        - visibility
+        - center_offset
+        - center_ohw
+        - height
+        cache_file: data/nuscenes/nuscenes_mmdet3d-keyframes/../nuscenes_map_aux_12Hz_interp/val_200x200_12Hz_interp.h5
+      - type: LoadBEVHDMap
+        dataset_root: data/nuscenes/
+        xbound:
+        - -40.0
+        - 40.0
+        - 0.4
+        ybound:
+        - -40.0
+        - 40.0
+        - 0.4
+        image_size:
+        - 224
+        - 400
+        object_classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: LoadDescription
+        dataset_root: data/nuscenes/
+        dataset_type: Occ3D-nuScenes
+      - type: ReorderMultiViewImagesM
+        order:
+        - CAM_FRONT_LEFT
+        - CAM_FRONT
+        - CAM_FRONT_RIGHT
+        - CAM_BACK_RIGHT
+        - CAM_BACK
+        - CAM_BACK_LEFT
+        safe: false
+      - type: ImageNormalize
+        mean:
+        - 0.5
+        - 0.5
+        - 0.5
+        std:
+        - 0.5
+        - 0.5
+        - 0.5
+      - type: DefaultFormatBundle3D
+        classes:
+        - barrier
+        - bicycle
+        - bus
+        - car
+        - construction_vehicle
+        - motorcycle
+        - pedestrian
+        - traffic_cone
+        - trailer
+        - truck
+      - type: Collect3D
+        keys:
+        - img
+        - gt_bboxes_3d
+        - gt_labels_3d
+        - gt_masks_bev
+        - gt_aux_bev
+        - bev_hdmap
+        - bev_hdmap_w_box
+        - layout_canvas
+        meta_keys:
+        - camera_intrinsics
+        - lidar2ego
+        - ego2global
+        - lidar2camera
+        - camera2lidar
+        - lidar2image
+        - img_aug_matrix
+        meta_lis_keys:
+        - timeofday
+        - location
+        - description
+        - detailed_description
+        - filename
+        - token
+        - lidar_token
+        - scene_name
+        - timestamp
+      object_classes:
+      - barrier
+      - bicycle
+      - bus
+      - car
+      - construction_vehicle
+      - motorcycle
+      - pedestrian
+      - traffic_cone
+      - trailer
+      - truck
+      map_classes:
+      - drivable_area
+      - ped_crossing
+      - walkway
+      - stop_line
+      - carpark_area
+      - road_divider
+      - lane_divider
+      - road_block
+      modality:
+        use_lidar: false
+        use_camera: true
+        use_radar: false
+        use_map: false
+        use_external: false
+      test_mode: true
+      force_all_boxes: true
+      box_type_3d: LiDAR
+      filter_empty_gt: false
+occ_render_path: data/nuscenes/occ_render_map/
+accelerator:
+  gradient_accumulation_steps: 1
+  mixed_precision: fp16
+  report_to: tensorboard
+runner:
+  foreground_loss_weight: 0.0
+  bbox_drop_ratio: 0
+  bbox_add_ratio: 0.1
+  bbox_add_num: 3
+  keyframe_rate: 1
+  num_train_epochs: 115
+  train_batch_size: 10
+  max_train_steps: 80960
+  num_workers: 8
+  prefetch_factor: 4
+  display_per_epoch: 20
+  display_per_n_min: 10
+  max_grad_norm: 1.0
+  set_grads_to_none: true
+  enable_xformers_memory_efficient_attention: true
+  unet_in_fp16: true
+  enable_unet_checkpointing: true
+  enable_controlnet_checkpointing: true
+  noise_offset: 0.0
+  train_with_same_offset: true
+  use_8bit_adam: false
+  adam_beta1: 0.9
+  adam_beta2: 0.999
+  adam_weight_decay: 0.01
+  adam_epsilon: 1.0e-08
+  learning_rate: 8.0e-05
+  lr_scheduler: constant_with_warmup
+  gradient_accumulation_steps: 1
+  lr_num_cycles: 1
+  lr_power: 1.0
+  lr_warmup_steps: 3000
+  checkpointing_steps: 5000
+  validation_steps: 20000
+  save_model_per_epoch: null
+  validation_before_run: false
+  validation_index:
+  - 204
+  - 912
+  - 1828
+  - 2253
+  - 4467
+  - 5543
+  validation_times: 4
+  validation_batch_size: 1
+  validation_show_box: true
+  validation_show_line: true
+  validation_seed_global: false
+  pipeline_param:
+    guidance_scale: 1.2
+    num_inference_steps: 20
+    eta: 0.0
+    controlnet_conditioning_scale: 1.0
+    guess_mode: false
+    use_zero_map_as_unconditional: false
+    bbox_max_length: null
+
+[2025-07-18 17:20:22,777][root][DEBUG] - start!
+[2025-07-18 17:20:22,778][root][INFO] - ***** Running training *****
+[2025-07-18 17:20:22,778][root][INFO] -   Num examples = 28130
+[2025-07-18 17:20:22,778][root][INFO] -   Num batches each epoch = 704
+[2025-07-18 17:20:22,778][root][INFO] -   Num Epochs = 115
+[2025-07-18 17:20:22,778][root][INFO] -   Instantaneous batch size per device = 10
+[2025-07-18 17:20:22,778][root][INFO] -   Total train batch size (w. parallel, distributed & accumulation) = 40
+[2025-07-18 17:20:22,778][root][INFO] -   Gradient Accumulation steps = 1
+[2025-07-18 17:20:22,778][root][INFO] -   Total optimization steps = 80960
+[2025-07-18 17:20:22,779][accelerate.accelerator][INFO] - Loading states from work_dirs/x-scene-img_224x400/img_unet_2025-07-16_10-33_224x400/checkpoint-80000
+[2025-07-18 17:20:24,597][accelerate.checkpointing][INFO] - All model weights loaded successfully
+[2025-07-18 17:20:25,736][accelerate.checkpointing][INFO] - All optimizer states loaded successfully
+[2025-07-18 17:20:25,737][accelerate.checkpointing][INFO] - All scheduler states loaded successfully
+[2025-07-18 17:20:25,738][accelerate.checkpointing][INFO] - GradScaler state loaded successfully
+[2025-07-18 17:20:25,738][accelerate.checkpointing][INFO] - All random states loaded successfully
+[2025-07-18 17:20:25,741][accelerate.accelerator][INFO] - Loading in 0 custom states
+[2025-07-18 17:20:25,742][root][INFO] - Starting from epoch 113 to 115
+[2025-07-18 17:20:39,039][root][WARNING] - [UNet2DConditionModelMultiview] Forward upsample size to force interpolation output size.
+[2025-07-18 18:00:02,390][root][INFO] - Save your model to: /data/yyang/workspace/X-Scene/work_dirs/x-scene-img_224x400/img_unet_2025-07-18_17-19_224x400

unet/config.json

@@ -0,0 +1,103 @@
+{
+  "_class_name": "UNet2DConditionModelMultiview",
+  "_diffusers_version": "0.17.1",
+  "_name_or_path": "pretrained/stable-diffusion-v2-1/",
+  "act_fn": "silu",
+  "addition_embed_type": null,
+  "addition_embed_type_num_heads": 64,
+  "attention_head_dim": [
+    5,
+    10,
+    20,
+    20
+  ],
+  "attn1_q_trainable": true,
+  "block_out_channels": [
+    320,
+    640,
+    1280,
+    1280
+  ],
+  "center_input_sample": false,
+  "class_embed_type": null,
+  "class_embeddings_concat": false,
+  "conv_in_kernel": 3,
+  "conv_out_kernel": 3,
+  "cross_attention_dim": 1024,
+  "cross_attention_norm": null,
+  "crossview_attn_type": "t5_crossview",
+  "down_block_types": [
+    "CrossAttnDownBlock2D",
+    "CrossAttnDownBlock2D",
+    "CrossAttnDownBlock2D",
+    "DownBlock2D"
+  ],
+  "downsample_padding": 1,
+  "dual_cross_attention": false,
+  "encoder_hid_dim": null,
+  "encoder_hid_dim_type": null,
+  "flip_sin_to_cos": true,
+  "freq_shift": 0,
+  "img_size": [
+    224,
+    400
+  ],
+  "in_channels": 4,
+  "layers_per_block": 2,
+  "mid_block_only_cross_attention": null,
+  "mid_block_scale_factor": 1,
+  "mid_block_type": "UNetMidBlock2DCrossAttn",
+  "neighboring_attn_type": "add",
+  "neighboring_view_pair": {
+    "0": [
+      5,
+      1
+    ],
+    "1": [
+      0,
+      2
+    ],
+    "2": [
+      1,
+      3
+    ],
+    "3": [
+      2,
+      4
+    ],
+    "4": [
+      3,
+      5
+    ],
+    "5": [
+      4,
+      0
+    ]
+  },
+  "norm_eps": 1e-05,
+  "norm_num_groups": 32,
+  "num_class_embeds": null,
+  "only_cross_attention": false,
+  "out_channels": 4,
+  "projection_class_embeddings_input_dim": null,
+  "resnet_out_scale_factor": 1.0,
+  "resnet_skip_time_act": false,
+  "resnet_time_scale_shift": "default",
+  "sample_size": 64,
+  "scene_channels": 320,
+  "time_cond_proj_dim": null,
+  "time_embedding_act_fn": null,
+  "time_embedding_dim": null,
+  "time_embedding_type": "positional",
+  "timestep_post_act": null,
+  "trainable_state": "only_new",
+  "up_block_types": [
+    "UpBlock2D",
+    "CrossAttnUpBlock2D",
+    "CrossAttnUpBlock2D",
+    "CrossAttnUpBlock2D"
+  ],
+  "upcast_attention": false,
+  "use_linear_projection": true,
+  "zero_module_type": "zero_linear"
+}

unet/diffusion_pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e568bdb2cc4c15366ac413bc8ae7147bbfef18e7e15b634bc4d2064207280dc
+size 2328074859