diff --git a/ABLATION_0225_FreqSelect/.hydra/config.yaml b/ABLATION_0225_FreqSelect/.hydra/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..ab763d2e35a5cba905eaf1fcec37ab5921156d9d
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/.hydra/config.yaml
@@ -0,0 +1,185 @@
+model:
+  encoder:
+    name: dcsplat
+    input_image_shape:
+    - 518
+    - 518
+    head_mode: pcd
+    num_level: 3
+    gs_param_dim: 256
+    align_corners: false
+    use_voxelize: true
+  decoder:
+    name: splatting_cuda
+    background_color:
+    - 0.0
+    - 0.0
+    - 0.0
+    make_scale_invariant: false
+  density_control:
+    name: density_control_module
+    mean_dim: 32
+    gs_param_dim: 256
+    refinement_layer_num: 1
+    num_level: 3
+    grad_mode: absgrad
+    use_mean_features: true
+    refinement_type: voxelize
+    refinement_hidden_dim: 32
+    aggregation_mode: mean
+    num_heads: 1
+    score_mode: frequency
+    latent_dim: 128
+    num_latents: 64
+    num_self_attn_per_block: 2
+    voxel_size: 0.001
+    aux_refine: false
+    refine_error: false
+    use_refine_module: true
+    voxelize_activate: true
+    use_depth: false
+render_loss:
+  mse:
+    weight: 1.0
+  lpips:
+    weight: 0.05
+    apply_after_step: 0
+density_control_loss:
+  error_score:
+    weight: 0.01
+    log_scale: false
+    grad_scale: 10000.0
+    mode: original
+direct_loss:
+  l1:
+    weight: 0.8
+  ssim:
+    weight: 0.2
+wandb:
+  project: DCSplat
+  entity: scene-representation-group
+  name: ABLATION_0225_FreqSelect
+  mode: online
+  tags:
+  - re10k
+  - 256x256
+mode: train
+data_loader:
+  train:
+    num_workers: 16
+    persistent_workers: true
+    batch_size: 16
+    seed: 1234
+  test:
+    num_workers: 4
+    persistent_workers: false
+    batch_size: 1
+    seed: 2345
+  val:
+    num_workers: 1
+    persistent_workers: true
+    batch_size: 1
+    seed: 3456
+optimizer:
+  lr: 0.0002
+  warm_up_steps: 25
+  backbone_lr_multiplier: 0.1
+  backbone_trainable: T+H
+  accumulate: 1
+checkpointing:
+  load: null
+  every_n_train_steps: 1500
+  save_top_k: 2
+  save_weights_only: false
+train:
+  extended_visualization: false
+  print_log_every_n_steps: 10
+  camera_loss: 10.0
+  one_sample_validation: null
+  align_corners: false
+  intrinsic_scaling: false
+  verbose: false
+  beta_dist_param:
+  - 0.5
+  - 4.0
+  use_refine_aux: false
+  train_target_set: true
+  train_gs_num: 1
+  ext_scale_detach: false
+  cam_scale_mode: sum
+  scene_scale_reg_loss: 0.01
+  train_aux: true
+  vggt_cam_loss: true
+  vggt_distil: false
+  context_view_train: false
+test:
+  output_path: test/ablation/re10k
+  align_pose: false
+  pose_align_steps: 100
+  rot_opt_lr: 0.005
+  trans_opt_lr: 0.005
+  compute_scores: true
+  save_image: false
+  save_video: false
+  save_active_mask_image: false
+  save_error_score_image: false
+  save_compare: false
+  pred_intrinsic: false
+  error_threshold: 0.4
+  error_threshold_list:
+  - 0.2
+  - 0.4
+  - 0.6
+  - 0.8
+  - 1.0
+  threshold_mode: ratio
+  nvs_view_N_list:
+  - 3
+  - 6
+  - 16
+  - 32
+  - 64
+seed: 111123
+trainer:
+  max_steps: 3001
+  val_check_interval: 250
+  gradient_clip_val: 0.5
+  num_nodes: 1
+dataset:
+  re10k:
+    make_baseline_1: true
+    relative_pose: true
+    augment: true
+    background_color:
+    - 0.0
+    - 0.0
+    - 0.0
+    overfit_to_scene: null
+    skip_bad_shape: true
+    view_sampler:
+      name: bounded
+      num_target_views: 4
+      num_context_views: 2
+      min_distance_between_context_views: 45
+      max_distance_between_context_views: 90
+      min_distance_to_context_views: 0
+      warm_up_steps: 1000
+      initial_min_distance_between_context_views: 25
+      initial_max_distance_between_context_views: 25
+      same_target_gap: false
+      num_target_set: 3
+    name: re10k
+    roots:
+    - datasets/re10k
+    input_image_shape:
+    - 256
+    - 256
+    original_image_shape:
+    - 360
+    - 640
+    cameras_are_circular: false
+    baseline_min: 0.001
+    baseline_max: 10000000000.0
+    max_fov: 100.0
+    dynamic_context_views: true
+    max_context_views_per_gpu: 24
diff --git a/ABLATION_0225_FreqSelect/.hydra/hydra.yaml b/ABLATION_0225_FreqSelect/.hydra/hydra.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..6bc481196a11d60bf0769cf33bc66e15f3ba56e0
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/.hydra/hydra.yaml
@@ -0,0 +1,165 @@
+hydra:
+  run:
+    dir: outputs/ablation/re10k/${wandb.name}
+  sweep:
+    dir: multirun/${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:
+    - +experiment=re10k_ablation_24v
+    - wandb.mode=online
+    - wandb.name=ABLATION_0225_FreqSelect
+    - model.density_control.score_mode=frequency
+  job:
+    name: main
+    chdir: null
+    override_dirname: +experiment=re10k_ablation_24v,model.density_control.score_mode=frequency,wandb.mode=online,wandb.name=ABLATION_0225_FreqSelect
+    id: ???
+    num: ???
+    config_name: main
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.3.2
+    version_base: '1.3'
+    cwd: /workspace/code/CVPR2026
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/code/CVPR2026/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_FreqSelect
+    choices:
+      experiment: re10k_ablation_24v
+      dataset@dataset.re10k: re10k
+      dataset/view_sampler_dataset_specific_config@dataset.re10k.view_sampler: bounded_re10k
+      dataset/view_sampler@dataset.re10k.view_sampler: bounded
+      model/density_control: density_control_module
+      model/decoder: splatting_cuda
+      model/encoder: dcsplat
+      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
diff --git a/ABLATION_0225_FreqSelect/.hydra/overrides.yaml b/ABLATION_0225_FreqSelect/.hydra/overrides.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..59c636a976739a91ac6c9e9f28874e18a00f3e2d
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/.hydra/overrides.yaml
@@ -0,0 +1,4 @@
+- +experiment=re10k_ablation_24v
+- wandb.mode=online
+- wandb.name=ABLATION_0225_FreqSelect
+- model.density_control.score_mode=frequency
diff --git a/ABLATION_0225_FreqSelect/wandb/debug-internal.log b/ABLATION_0225_FreqSelect/wandb/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..177c83d79642e27d5ccf1769f6566b523253107a
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/wandb/debug-internal.log
@@ -0,0 +1,12 @@
+{"time":"2026-02-24T22:27:39.882209485Z","level":"INFO","msg":"stream: starting","core version":"0.25.0"}
+{"time":"2026-02-24T22:27:40.294571378Z","level":"INFO","msg":"stream: created new stream","id":"y7wvpmyy"}
+{"time":"2026-02-24T22:27:40.2947114Z","level":"INFO","msg":"handler: started","stream_id":"y7wvpmyy"}
+{"time":"2026-02-24T22:27:40.294855053Z","level":"INFO","msg":"stream: started","id":"y7wvpmyy"}
+{"time":"2026-02-24T22:27:40.294904223Z","level":"INFO","msg":"sender: started","stream_id":"y7wvpmyy"}
+{"time":"2026-02-24T22:27:40.294940724Z","level":"INFO","msg":"writer: started","stream_id":"y7wvpmyy"}
+{"time":"2026-02-25T01:00:56.785103175Z","level":"INFO","msg":"api: retrying HTTP error","status":502,"url":"https://api.wandb.ai/files/know/DCSplat/y7wvpmyy/file_stream","body":"\n<html><head>\n<meta http-equiv=\"content-type\" content=\"text/html;charset=utf-8\">\n<title>502 Server Error</title>\n</head>\n<body text=#000000 bgcolor=#ffffff>\n<h1>Error: Server Error</h1>\n<h2>The server encountered a temporary error and could not complete your request.<p>Please try again in 30 seconds.</h2>\n<h2></h2>\n</body></html>\n"}
+{"time":"2026-02-25T01:39:55.965783052Z","level":"INFO","msg":"stream: closing","id":"y7wvpmyy"}
+{"time":"2026-02-25T01:39:56.929029575Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-02-25T01:39:57.1548805Z","level":"INFO","msg":"handler: closed","stream_id":"y7wvpmyy"}
+{"time":"2026-02-25T01:39:57.155103083Z","level":"INFO","msg":"sender: closed","stream_id":"y7wvpmyy"}
+{"time":"2026-02-25T01:39:57.155127144Z","level":"INFO","msg":"stream: closed","id":"y7wvpmyy"}
diff --git a/ABLATION_0225_FreqSelect/wandb/debug.log b/ABLATION_0225_FreqSelect/wandb/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..159b8e67b209c045d745b123179f99e28311277a
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/wandb/debug.log
@@ -0,0 +1,21 @@
+2026-02-24 22:27:39,587 INFO    MainThread:113743 [wandb_setup.py:_flush():81] Current SDK version is 0.25.0
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_setup.py:_flush():81] Configure stats pid to 113743
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug.log
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug-internal.log
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_init.py:init():844] calling init triggers
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
+config: {'model': {'encoder': {'name': 'dcsplat', 'input_image_shape': [518, 518], 'head_mode': 'pcd', 'num_level': 3, 'gs_param_dim': 256, 'align_corners': False, 'use_voxelize': True}, 'decoder': {'name': 'splatting_cuda', 'background_color': [0.0, 0.0, 0.0], 'make_scale_invariant': False}, 'density_control': {'name': 'density_control_module', 'mean_dim': 32, 'gs_param_dim': 256, 'refinement_layer_num': 1, 'num_level': 3, 'grad_mode': 'absgrad', 'use_mean_features': True, 'refinement_type': 'voxelize', 'refinement_hidden_dim': 32, 'aggregation_mode': 'mean', 'num_heads': 1, 'score_mode': 'frequency', 'latent_dim': 128, 'num_latents': 64, 'num_self_attn_per_block': 2, 'voxel_size': 0.001, 'aux_refine': False, 'refine_error': False, 'use_refine_module': True, 'voxelize_activate': True, 'use_depth': False}}, 'render_loss': {'mse': {'weight': 1.0}, 'lpips': {'weight': 0.05, 'apply_after_step': 0}}, 'density_control_loss': {'error_score': {'weight': 0.01, 'log_scale': False, 'grad_scale': 10000.0, 'mode': 'original'}}, 'direct_loss': {'l1': {'weight': 0.8}, 'ssim': {'weight': 0.2}}, 'wandb': {'project': 'DCSplat', 'entity': 'scene-representation-group', 'name': 'ABLATION_0225_FreqSelect', 'mode': 'online', 'tags': ['re10k', '256x256']}, 'mode': 'train', 'data_loader': {'train': {'num_workers': 16, 'persistent_workers': True, 'batch_size': 16, 'seed': 1234}, 'test': {'num_workers': 4, 'persistent_workers': False, 'batch_size': 1, 'seed': 2345}, 'val': {'num_workers': 1, 'persistent_workers': True, 'batch_size': 1, 'seed': 3456}}, 'optimizer': {'lr': 0.0002, 'warm_up_steps': 25, 'backbone_lr_multiplier': 0.1, 'backbone_trainable': 'T+H', 'accumulate': 1}, 'checkpointing': {'load': None, 'every_n_train_steps': 1500, 'save_top_k': 2, 'save_weights_only': False}, 'train': {'extended_visualization': False, 'print_log_every_n_steps': 10, 'camera_loss': 10.0, 'one_sample_validation': None, 'align_corners': False, 'intrinsic_scaling': False, 'verbose': False, 'beta_dist_param': [0.5, 4.0], 'use_refine_aux': False, 'train_target_set': True, 'train_gs_num': 1, 'ext_scale_detach': False, 'cam_scale_mode': 'sum', 'scene_scale_reg_loss': 0.01, 'train_aux': True, 'vggt_cam_loss': True, 'vggt_distil': False, 'context_view_train': False}, 'test': {'output_path': 'test/ablation/re10k', 'align_pose': False, 'pose_align_steps': 100, 'rot_opt_lr': 0.005, 'trans_opt_lr': 0.005, 'compute_scores': True, 'save_image': False, 'save_video': False, 'save_active_mask_image': False, 'save_error_score_image': False, 'save_compare': False, 'pred_intrinsic': False, 'error_threshold': 0.4, 'error_threshold_list': [0.2, 0.4, 0.6, 0.8, 1.0], 'threshold_mode': 'ratio', 'nvs_view_N_list': [3, 6, 16, 32, 64]}, 'seed': 111123, 'trainer': {'max_steps': 3001, 'val_check_interval': 250, 'gradient_clip_val': 0.5, 'num_nodes': 1}, 'dataset': {'re10k': {'make_baseline_1': True, 'relative_pose': True, 'augment': True, 'background_color': [0.0, 0.0, 0.0], 'overfit_to_scene': None, 'skip_bad_shape': True, 'view_sampler': {'name': 'bounded', 'num_target_views': 4, 'num_context_views': 2, 'min_distance_between_context_views': 45, 'max_distance_between_context_views': 90, 'min_distance_to_context_views': 0, 'warm_up_steps': 1000, 'initial_min_distance_between_context_views': 25, 'initial_max_distance_between_context_views': 25, 'same_target_gap': False, 'num_target_set': 3}, 'name': 're10k', 'roots': ['datasets/re10k'], 'input_image_shape': [256, 256], 'original_image_shape': [360, 640], 'cameras_are_circular': False, 'baseline_min': 0.001, 'baseline_max': 10000000000.0, 'max_fov': 100.0, 'dynamic_context_views': True, 'max_context_views_per_gpu': 24}}, '_wandb': {}}
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_init.py:init():892] starting backend
+2026-02-24 22:27:39,873 INFO    MainThread:113743 [wandb_init.py:init():895] sending inform_init request
+2026-02-24 22:27:39,880 INFO    MainThread:113743 [wandb_init.py:init():903] backend started and connected
+2026-02-24 22:27:39,887 INFO    MainThread:113743 [wandb_init.py:init():973] updated telemetry
+2026-02-24 22:27:39,894 INFO    MainThread:113743 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
+2026-02-24 22:27:41,506 INFO    MainThread:113743 [wandb_init.py:init():1042] starting run threads in backend
+2026-02-24 22:27:41,632 INFO    MainThread:113743 [wandb_run.py:_console_start():2524] atexit reg
+2026-02-24 22:27:41,632 INFO    MainThread:113743 [wandb_run.py:_redirect():2373] redirect: wrap_raw
+2026-02-24 22:27:41,632 INFO    MainThread:113743 [wandb_run.py:_redirect():2442] Wrapping output streams.
+2026-02-24 22:27:41,632 INFO    MainThread:113743 [wandb_run.py:_redirect():2465] Redirects installed.
+2026-02-24 22:27:41,635 INFO    MainThread:113743 [wandb_init.py:init():1082] run started, returning control to user process
+2026-02-25 01:39:55,965 INFO    wandb-AsyncioManager-main:113743 [service_client.py:_forward_responses():134] Reached EOF.
+2026-02-25 01:39:55,965 INFO    wandb-AsyncioManager-main:113743 [mailbox.py:close():155] Closing mailbox, abandoning 1 handles.
diff --git a/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/config.yaml b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..d18786cbf1782e3c3cffcaadb215bf1bfb8ea454
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/config.yaml
@@ -0,0 +1,307 @@
+_wandb:
+    value:
+        cli_version: 0.25.0
+        e:
+            1aoh34iwmaamch760bz6silmn5l3ie5b:
+                args:
+                    - +experiment=re10k_ablation_24v
+                    - wandb.mode=online
+                    - wandb.name=ABLATION_0225_FreqSelect
+                    - model.density_control.score_mode=frequency
+                cpu_count: 128
+                cpu_count_logical: 256
+                cudaVersion: "13.1"
+                disk:
+                    /:
+                        total: "1170378588160"
+                        used: "636725506048"
+                email: dna9041@korea.ac.kr
+                executable: /venv/main/bin/python
+                git:
+                    commit: 2512754c6c27ca5150bf17fbcbdde3f192fd53cc
+                    remote: git@github.com:K-nowing/CVPR2026.git
+                gpu: NVIDIA H200
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-2649ab80-a3a6-5a1c-0fa5-12bc11bd75e9
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-e92921d9-c681-246f-af93-637e0dc938ca
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-ffe12ffc-9bb7-82de-5692-1ec0ee2e68d8
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-499e5acd-b6ab-2010-c51b-ee9b5aa65825
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-3b2522d9-1c72-e49b-2c30-96165680b74a
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-a9a280c5-b2f9-dc1e-a8a9-7326a74001ff
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-07d0167b-a6a1-1900-2d27-7c6c11598409
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-8362a999-20d1-c27b-5d18-032d23f859ab
+                host: 27d18dedec6d
+                memory:
+                    total: "1622948257792"
+                os: Linux-6.8.0-90-generic-x86_64-with-glibc2.39
+                program: -m src.main
+                python: CPython 3.12.12
+                root: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_FreqSelect
+                startedAt: "2026-02-24T22:27:39.584882Z"
+                writerId: 1aoh34iwmaamch760bz6silmn5l3ie5b
+        m:
+            - "1": trainer/global_step
+              "6":
+                - 3
+              "7": []
+            - "2": '*'
+              "5": 1
+              "6":
+                - 1
+              "7": []
+        python_version: 3.12.12
+        t:
+            "1":
+                - 1
+                - 41
+                - 49
+                - 50
+                - 106
+            "2":
+                - 1
+                - 41
+                - 49
+                - 50
+                - 106
+            "3":
+                - 7
+                - 13
+                - 15
+                - 16
+                - 66
+            "4": 3.12.12
+            "5": 0.25.0
+            "12": 0.25.0
+            "13": linux-x86_64
+checkpointing:
+    value:
+        every_n_train_steps: 1500
+        load: null
+        save_top_k: 2
+        save_weights_only: false
+data_loader:
+    value:
+        test:
+            batch_size: 1
+            num_workers: 4
+            persistent_workers: false
+            seed: 2345
+        train:
+            batch_size: 16
+            num_workers: 16
+            persistent_workers: true
+            seed: 1234
+        val:
+            batch_size: 1
+            num_workers: 1
+            persistent_workers: true
+            seed: 3456
+dataset:
+    value:
+        re10k:
+            augment: true
+            background_color:
+                - 0
+                - 0
+                - 0
+            baseline_max: 1e+10
+            baseline_min: 0.001
+            cameras_are_circular: false
+            dynamic_context_views: true
+            input_image_shape:
+                - 256
+                - 256
+            make_baseline_1: true
+            max_context_views_per_gpu: 24
+            max_fov: 100
+            name: re10k
+            original_image_shape:
+                - 360
+                - 640
+            overfit_to_scene: null
+            relative_pose: true
+            roots:
+                - datasets/re10k
+            skip_bad_shape: true
+            view_sampler:
+                initial_max_distance_between_context_views: 25
+                initial_min_distance_between_context_views: 25
+                max_distance_between_context_views: 90
+                min_distance_between_context_views: 45
+                min_distance_to_context_views: 0
+                name: bounded
+                num_context_views: 2
+                num_target_set: 3
+                num_target_views: 4
+                same_target_gap: false
+                warm_up_steps: 1000
+density_control_loss:
+    value:
+        error_score:
+            grad_scale: 10000
+            log_scale: false
+            mode: original
+            weight: 0.01
+direct_loss:
+    value:
+        l1:
+            weight: 0.8
+        ssim:
+            weight: 0.2
+mode:
+    value: train
+model:
+    value:
+        decoder:
+            background_color:
+                - 0
+                - 0
+                - 0
+            make_scale_invariant: false
+            name: splatting_cuda
+        density_control:
+            aggregation_mode: mean
+            aux_refine: false
+            grad_mode: absgrad
+            gs_param_dim: 256
+            latent_dim: 128
+            mean_dim: 32
+            name: density_control_module
+            num_heads: 1
+            num_latents: 64
+            num_level: 3
+            num_self_attn_per_block: 2
+            refine_error: false
+            refinement_hidden_dim: 32
+            refinement_layer_num: 1
+            refinement_type: voxelize
+            score_mode: frequency
+            use_depth: false
+            use_mean_features: true
+            use_refine_module: true
+            voxel_size: 0.001
+            voxelize_activate: true
+        encoder:
+            align_corners: false
+            gs_param_dim: 256
+            head_mode: pcd
+            input_image_shape:
+                - 518
+                - 518
+            name: dcsplat
+            num_level: 3
+            use_voxelize: true
+optimizer:
+    value:
+        accumulate: 1
+        backbone_lr_multiplier: 0.1
+        backbone_trainable: T+H
+        lr: 0.0002
+        warm_up_steps: 25
+render_loss:
+    value:
+        lpips:
+            apply_after_step: 0
+            weight: 0.05
+        mse:
+            weight: 1
+seed:
+    value: 111123
+test:
+    value:
+        align_pose: false
+        compute_scores: true
+        error_threshold: 0.4
+        error_threshold_list:
+            - 0.2
+            - 0.4
+            - 0.6
+            - 0.8
+            - 1
+        nvs_view_N_list:
+            - 3
+            - 6
+            - 16
+            - 32
+            - 64
+        output_path: test/ablation/re10k
+        pose_align_steps: 100
+        pred_intrinsic: false
+        rot_opt_lr: 0.005
+        save_active_mask_image: false
+        save_compare: false
+        save_error_score_image: false
+        save_image: false
+        save_video: false
+        threshold_mode: ratio
+        trans_opt_lr: 0.005
+train:
+    value:
+        align_corners: false
+        beta_dist_param:
+            - 0.5
+            - 4
+        cam_scale_mode: sum
+        camera_loss: 10
+        context_view_train: false
+        ext_scale_detach: false
+        extended_visualization: false
+        intrinsic_scaling: false
+        one_sample_validation: null
+        print_log_every_n_steps: 10
+        scene_scale_reg_loss: 0.01
+        train_aux: true
+        train_gs_num: 1
+        train_target_set: true
+        use_refine_aux: false
+        verbose: false
+        vggt_cam_loss: true
+        vggt_distil: false
+trainer:
+    value:
+        gradient_clip_val: 0.5
+        max_steps: 3001
+        num_nodes: 1
+        val_check_interval: 250
+wandb:
+    value:
+        entity: scene-representation-group
+        mode: online
+        name: ABLATION_0225_FreqSelect
+        project: DCSplat
+        tags:
+            - re10k
+            - 256x256
diff --git a/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/output.log b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/output.log
new file mode 100644
index 0000000000000000000000000000000000000000..264738ff41f8d6048f336ca5daabc7ed3594ebaf
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/output.log
@@ -0,0 +1,800 @@
+LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0,1,2,3,4,5,6,7]
+
+  | Name                   | Type                 | Params | Mode
+------------------------------------------------------------------------
+0 | encoder                | OurSplat             | 888 M  | train
+1 | density_control_module | DensityControlModule | 2.6 M  | train
+2 | decoder                | DecoderSplattingCUDA | 0      | train
+3 | render_losses          | ModuleList           | 0      | train
+4 | density_control_losses | ModuleList           | 0      | train
+5 | direct_losses          | ModuleList           | 0      | train
+------------------------------------------------------------------------
+891 M     Trainable params
+0         Non-trainable params
+891 M     Total params
+3,564.326 Total estimated model params size (MB)
+1226      Modules in train mode
+522       Modules in eval mode
+Sanity Checking: |                                                                                                                                                                      | 0/? [00:00<?, ?it/s][2026-02-24 22:27:44,325][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/data_connector.py:425: The 'val_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=31` in the `DataLoader` to improve performance.
+
+[2026-02-24 22:27:44,327][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user.
+  warnings.warn(  # warn only once
+
+Validation epoch start on rank 0
+Sanity Checking DataLoader 0:   0%|                                                                                                                                                     | 0/1 [00:00<?, ?it/s]validation step 0; scene = ['306e2b7785657539'];
+target intrinsic:  tensor(0.8595, device='cuda:0') tensor(0.8597, device='cuda:0')
+pred intrinsic:  tensor(0.8779, device='cuda:0') tensor(0.8773, device='cuda:0')
+[rank0]:W0224 22:27:46.800000 113743 site-packages/torch/utils/cpp_extension.py:2425] TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation.
+[rank0]:W0224 22:27:46.800000 113743 site-packages/torch/utils/cpp_extension.py:2425] If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'] to specific architectures.
+[2026-02-24 22:27:46,874][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-24 22:27:46,883][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/utilities/data.py:79: Trying to infer the `batch_size` from an ambiguous collection. The batch size we found is 1. To avoid any miscalculations, use `self.log(..., batch_size=batch_size)`.
+
+Setting up [LPIPS] perceptual loss: trunk [vgg], v[0.1], spatial [off]
+[2026-02-24 22:27:46,884][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-24 22:27:46,884][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+Loading model from: /venv/main/lib/python3.12/site-packages/lpips/weights/v0.1/vgg.pth
+[2026-02-24 22:27:48,528][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/functional.py:554: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /pytorch/aten/src/ATen/native/TensorShape.cpp:4322.)
+  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
+
+Sanity Checking DataLoader 0: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:03<00:00,  0.25it/s][2026-02-24 22:27:48,812][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/psnr', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-24 22:27:48,814][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/lpips', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-24 22:27:48,814][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/ssim', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-24 22:27:48,814][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/gaussian_num_ratio', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-24 22:27:48,814][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('info/global_step', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+Epoch 0: |                                                                                                                                                                              | 0/? [00:00<?, ?it/s]context = [[34, 36, 50, 53, 54, 60, 63, 70, 76, 78, 79, 80, 81, 88, 92, 94, 102, 110, 112, 114, 122, 125, 126, 131]]target = [[126, 96, 109, 55, 99, 116, 43, 60, 113, 85, 103, 90, 130, 62, 76, 123, 35, 102, 125, 128, 98, 67, 129, 79]]
+[2026-02-24 22:27:57,859][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [256, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [256, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-24 22:27:57,942][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                          | 9/? [00:41<00:00,  0.22it/s, v_num=pmyy]train step 10; scene = [['08c26703c4987851']]; loss = 0.847254
+Epoch 0: |                                                                                                                                                         | 10/? [00:45<00:00,  0.22it/s, v_num=pmyy]context = [[98, 107, 112, 119, 122, 123], [21, 22, 27, 36, 38, 46], [63, 66, 67, 77, 84, 88], [56, 62, 65, 68, 78, 81]]target = [[105, 110, 116, 112, 111, 101], [29, 37, 36, 22, 40, 45], [73, 66, 75, 87, 83, 64], [79, 69, 75, 58, 61, 62]]
+Epoch 0: |                                                                                                                                                         | 19/? [01:18<00:00,  0.24it/s, v_num=pmyy]train step 20; scene = [['4012c15c8381568b'], ['af08406c5a9a43a0'], ['9f9f9beffb86fad7'], ['fc8d08df6c875cb0']]; loss = 0.233043
+Epoch 0: |                                                                                                                                                         | 20/? [01:22<00:00,  0.24it/s, v_num=pmyy]context = [[144, 152, 157, 164, 166, 169, 171, 177], [201, 211, 216, 221, 228, 230, 233, 234], [11, 15, 16, 23, 30, 38, 43, 44]]target = [[153, 170, 149, 169, 145, 174, 165, 157], [229, 216, 205, 206, 203, 213, 233, 215], [37, 38, 39, 35, 15, 24, 19, 25]]
+Epoch 0: |                                                                                                                                                         | 24/? [01:37<00:00,  0.25it/s, v_num=pmyy][2026-02-24 22:29:33,113][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+Epoch 0: |                                                                                                                                                         | 29/? [01:55<00:00,  0.25it/s, v_num=pmyy]train step 30; scene = [['00980329a3221f1c'], ['1e7c432d2207b6f2'], ['af2748330e5243d0']]; loss = 0.165788
+Epoch 0: |                                                                                                                                                         | 30/? [01:59<00:00,  0.25it/s, v_num=pmyy]context = [[2, 7, 10, 15, 20, 24, 25, 29, 30, 32, 33, 37, 40, 44, 49, 60, 64, 74, 76, 81, 89, 91, 95, 99]]target = [[79, 84, 35, 43, 89, 44, 63, 58, 48, 13, 65, 7, 96, 27, 51, 20, 60, 15, 85, 59, 22, 66, 17, 62]]
+Epoch 0: |                                                                                                                                                         | 39/? [02:33<00:00,  0.25it/s, v_num=pmyy]train step 40; scene = [['79a9385753d426bc'], ['593538382d2dc847'], ['c9c67636b9d521be']]; loss = 0.133178
+Epoch 0: |                                                                                                                                                         | 40/? [02:37<00:00,  0.25it/s, v_num=pmyy]context = [[200, 208, 223, 225], [6, 8, 12, 31], [104, 107, 111, 129], [69, 80, 86, 95], [11, 15, 18, 37], [52, 60, 70, 78]]target = [[213, 203, 214, 201], [29, 11, 15, 28], [111, 122, 120, 115], [77, 90, 89, 78], [18, 24, 29, 14], [64, 67, 63, 74]]
+Epoch 0: |                                                                                                                                                         | 49/? [03:10<00:00,  0.26it/s, v_num=pmyy]train step 50; scene = [['579a11551b3315d9'], ['c9dd64b7415e788e'], ['6f3fb517d1798d03']]; loss = 0.135591
+Epoch 0: |                                                                                                                                                         | 50/? [03:14<00:00,  0.26it/s, v_num=pmyy]context = [[20, 22, 25, 28, 36, 47], [10, 13, 15, 22, 26, 35], [31, 47, 49, 51, 52, 56], [109, 110, 115, 116, 125, 135]]target = [[40, 24, 23, 30, 37, 33], [19, 32, 25, 24, 26, 28], [35, 52, 50, 55, 45, 36], [131, 123, 130, 110, 115, 134]]
+Epoch 0: |                                                                                                                                                         | 59/? [03:47<00:00,  0.26it/s, v_num=pmyy]train step 60; scene = [['07916b8004a8e336'], ['e51ef9945ae527c4'], ['db84f84b1d775bb8'], ['92ed61f8e16b7e67']]; loss = 0.112286
+Epoch 0: |                                                                                                                                                         | 60/? [03:51<00:00,  0.26it/s, v_num=pmyy]context = [[96, 97, 101, 102, 105, 106, 107, 114, 124, 134, 138, 145], [76, 78, 82, 83, 96, 98, 100, 111, 112, 118, 119, 125]]target = [[144, 124, 137, 102, 107, 130, 119, 129, 118, 123, 133, 126], [98, 108, 96, 100, 113, 123, 77, 78, 112, 109, 80, 102]]
+Epoch 0: |                                                                                                                                                         | 69/? [04:26<00:00,  0.26it/s, v_num=pmyy]train step 70; scene = [['c34efa1505a0cfaa'], ['a3d0cca9fb57fd85'], ['43d0e6dce7bb1e95'], ['d8c2f0a3734cb493']]; loss = 0.086001
+Epoch 0: |                                                                                                                                                         | 70/? [04:28<00:00,  0.26it/s, v_num=pmyy]context = [[204, 208, 211, 218, 232, 234, 237, 239, 240, 241, 247, 253], [6, 7, 10, 12, 35, 36, 37, 38, 40, 45, 48, 55]]target = [[230, 236, 222, 227, 220, 252, 216, 221, 213, 217, 249, 240], [18, 24, 25, 12, 9, 42, 34, 52, 41, 33, 43, 48]]
+Epoch 0: |                                                                                                                                                         | 79/? [05:03<00:00,  0.26it/s, v_num=pmyy]train step 80; scene = [['24d756c820744e31'], ['cd6c21656a85e9b9'], ['f3b24cf238154fc0']]; loss = 0.091763
+Epoch 0: |                                                                                                                                                         | 80/? [05:07<00:00,  0.26it/s, v_num=pmyy]context = [[4, 30], [52, 79], [61, 87], [12, 40], [83, 109], [3, 29], [221, 249], [198, 227], [9, 38], [46, 72], [0, 26], [123, 150]]target = [[27, 26], [58, 57], [70, 77], [26, 20], [87, 98], [26, 14], [223, 224], [221, 200], [23, 33], [67, 64], [4, 25], [140, 148]]
+Epoch 0: |                                                                                                                                                         | 89/? [05:41<00:00,  0.26it/s, v_num=pmyy]train step 90; scene = [['617b4bc98d7e0bb6'], ['666e4a9aba27bb64']]; loss = 0.066406
+Epoch 0: |                                                                                                                                                         | 90/? [05:45<00:00,  0.26it/s, v_num=pmyy]context = [[134, 135, 138, 146, 147, 151, 156, 161, 162, 164, 166, 167, 168, 169, 187, 189, 197, 210, 215, 224, 225, 228, 230, 231]]target = [[143, 149, 191, 182, 151, 226, 165, 140, 208, 171, 179, 223, 168, 136, 194, 207, 227, 144, 187, 185, 145, 218, 139, 170]]
+Epoch 0: |                                                                                                                                                         | 99/? [06:20<00:00,  0.26it/s, v_num=pmyy]train step 100; scene = [['12fee7f1978d52f1'], ['c963bb60939e2d81']]; loss = 0.124442
+Epoch 0: |                                                                                                                                                        | 100/? [06:24<00:00,  0.26it/s, v_num=pmyy]context = [[40, 44, 47, 48, 49, 50, 56, 63, 77, 79, 88, 89], [21, 33, 34, 35, 36, 40, 43, 46, 47, 60, 62, 70]]target = [[58, 81, 76, 64, 68, 72, 51, 87, 77, 65, 88, 45], [31, 47, 25, 35, 55, 22, 48, 65, 29, 40, 63, 67]]
+Epoch 0: |                                                                                                                                                        | 109/? [06:58<00:00,  0.26it/s, v_num=pmyy]train step 110; scene = [['47396d5a5299873e']]; loss = 0.111578
+Epoch 0: |                                                                                                                                                        | 110/? [07:01<00:00,  0.26it/s, v_num=pmyy]context = [[20, 22, 29, 30, 31, 34, 37, 45, 47, 51, 56, 69], [18, 22, 24, 27, 31, 37, 44, 48, 49, 60, 61, 67]]target = [[49, 22, 36, 59, 63, 60, 45, 66, 38, 28, 26, 64], [61, 65, 25, 20, 63, 60, 26, 22, 33, 45, 37, 35]]
+Epoch 0: |                                                                                                                                                        | 119/? [07:36<00:00,  0.26it/s, v_num=pmyy]train step 120; scene = [['9bd7044e7cbf8e60'], ['76e44cf6b5658b26']]; loss = 0.076855
+Epoch 0: |                                                                                                                                                        | 120/? [07:40<00:00,  0.26it/s, v_num=pmyy]context = [[7, 14, 26, 28, 31, 34, 37, 40], [10, 20, 26, 32, 35, 39, 41, 43], [17, 23, 26, 28, 44, 46, 49, 50]]target = [[14, 26, 10, 22, 30, 13, 31, 11], [24, 31, 36, 14, 40, 34, 41, 13], [24, 44, 34, 48, 46, 36, 26, 38]]
+Epoch 0: |                                                                                                                                                        | 129/? [08:14<00:00,  0.26it/s, v_num=pmyy]train step 130; scene = [['a8cef6a851fbea3c'], ['b6699f4d039a5b06'], ['55cf2bbe9e017ea4'], ['6b0dd861e1ab1fec'], ['14db202c335af709'], ['8b6ff6c5153a7794'], ['b75f3820760d835c'], ['f7dbc855fd2a7669'], ['cfb20f8971e6a591'], ['95f2be7bb8303f50'], ['ff422469e034ae11'], ['5a2ad43377e9d18d']]; loss = 0.107566
+Epoch 0: |                                                                                                                                                        | 130/? [08:18<00:00,  0.26it/s, v_num=pmyy]context = [[8, 15, 26, 30, 32, 35, 40, 46, 47, 48, 50, 53, 54, 62, 69, 70, 72, 76, 80, 86, 90, 99, 100, 105]]target = [[26, 12, 49, 100, 89, 24, 10, 81, 37, 63, 52, 17, 39, 70, 16, 56, 40, 55, 43, 34, 72, 28, 48, 45]]
+Epoch 0: |                                                                                                                                                        | 139/? [08:51<00:00,  0.26it/s, v_num=pmyy]train step 140; scene = [['f62a962df5c26a1a'], ['b076420679a04731']]; loss = 0.081166
+Epoch 0: |                                                                                                                                                        | 140/? [08:55<00:00,  0.26it/s, v_num=pmyy]context = [[90, 95, 97, 104, 118, 121], [14, 29, 34, 35, 36, 44], [12, 14, 22, 27, 34, 41], [6, 10, 16, 18, 34, 35]]target = [[99, 104, 93, 113, 110, 96], [29, 37, 22, 21, 39, 41], [19, 25, 39, 27, 29, 13], [26, 28, 9, 7, 13, 34]]
+Epoch 0: |                                                                                                                                                        | 149/? [09:29<00:00,  0.26it/s, v_num=pmyy]train step 150; scene = [['a52d26a78b04aebd']]; loss = 0.066867
+Epoch 0: |                                                                                                                                                        | 150/? [09:33<00:00,  0.26it/s, v_num=pmyy]context = [[115, 132, 134, 145], [16, 32, 39, 44], [57, 63, 78, 88], [9, 11, 23, 36], [15, 26, 28, 44], [1, 8, 22, 33]]target = [[124, 132, 116, 119], [27, 24, 33, 29], [81, 61, 85, 79], [28, 26, 29, 16], [28, 29, 37, 22], [9, 4, 32, 27]]
+Epoch 0: |                                                                                                                                                        | 159/? [10:07<00:00,  0.26it/s, v_num=pmyy]train step 160; scene = [['268fbffc6c479d5b']]; loss = 0.063764
+Epoch 0: |                                                                                                                                                        | 160/? [10:11<00:00,  0.26it/s, v_num=pmyy]context = [[18, 25, 26, 37, 42, 46, 49, 51, 53, 64, 65, 67], [69, 75, 78, 79, 82, 84, 94, 95, 104, 108, 117, 118]]target = [[53, 27, 22, 32, 41, 38, 50, 43, 47, 48, 23, 19], [74, 70, 114, 115, 90, 89, 88, 92, 94, 110, 107, 101]]
+Epoch 0: |                                                                                                                                                        | 169/? [10:45<00:00,  0.26it/s, v_num=pmyy]train step 170; scene = [['719e2e8912e4eed3'], ['a3e51565a737569f']]; loss = 0.124253
+Epoch 0: |                                                                                                                                                        | 170/? [10:49<00:00,  0.26it/s, v_num=pmyy]context = [[14, 18, 20, 21, 24, 27, 33, 40, 45, 47, 48, 51, 52, 60, 64, 70, 75, 77, 80, 85, 90, 98, 102, 111]]target = [[29, 93, 32, 39, 81, 108, 72, 107, 51, 35, 16, 36, 70, 18, 34, 92, 94, 47, 23, 74, 50, 77, 19, 37]]
+Epoch 0: |                                                                                                                                                        | 179/? [11:23<00:00,  0.26it/s, v_num=pmyy]train step 180; scene = [['f44b9aa76a94a0a3']]; loss = 0.074524
+Epoch 0: |                                                                                                                                                        | 180/? [11:27<00:00,  0.26it/s, v_num=pmyy]context = [[0, 6, 17, 22, 26, 28, 33, 41, 50, 55, 56, 57, 71, 76, 79, 81, 84, 85, 86, 87, 89, 95, 96, 97]]target = [[37, 49, 12, 78, 9, 16, 84, 13, 5, 4, 6, 38, 80, 51, 43, 68, 64, 46, 56, 24, 25, 72, 36, 21]]
+Epoch 0: |                                                                                                                                                        | 189/? [12:02<00:00,  0.26it/s, v_num=pmyy]train step 190; scene = [['71bb669d936a5718'], ['a47203cfd5e0a478'], ['4b009f82cf5c7098']]; loss = 0.069627
+Epoch 0: |                                                                                                                                                        | 190/? [12:06<00:00,  0.26it/s, v_num=pmyy]context = [[9, 11, 19, 25, 33, 43, 46, 47, 48, 54, 57, 58, 63, 70, 72, 75, 79, 80, 83, 84, 85, 96, 99, 106]]target = [[30, 53, 82, 31, 55, 12, 20, 72, 104, 70, 24, 52, 21, 32, 102, 71, 35, 11, 46, 10, 15, 74, 33, 26]]
+Epoch 0: |                                                                                                                                                        | 199/? [12:40<00:00,  0.26it/s, v_num=pmyy]train step 200; scene = [['dd5ec950a01c42a0'], ['6d0db0358f7e051e'], ['983fe650a925ec1b']]; loss = 0.083580
+Epoch 0: |                                                                                                                                                        | 200/? [12:44<00:00,  0.26it/s, v_num=pmyy]context = [[8, 10, 14, 15, 22, 23, 27, 30, 36, 38, 39, 47, 64, 65, 67, 78, 79, 80, 83, 86, 93, 96, 98, 105]]target = [[73, 63, 10, 27, 89, 35, 44, 58, 97, 71, 17, 24, 66, 87, 50, 12, 23, 11, 31, 45, 69, 96, 98, 94]]
+[2026-02-24 22:40:40,460][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 209/? [13:24<00:00,  0.26it/s, v_num=pmyy]train step 210; scene = [['9be9b273b3c22c61'], ['4b5883872c9b860c']]; loss = 0.061884
+Epoch 0: |                                                                                                                                                        | 210/? [13:28<00:00,  0.26it/s, v_num=pmyy]context = [[34, 35, 37, 54, 56, 59, 61, 68, 69, 76, 90, 96, 98, 101, 103, 107, 116, 119, 120, 121, 122, 125, 126, 131]]target = [[95, 50, 119, 121, 84, 107, 72, 52, 80, 42, 127, 94, 79, 98, 46, 128, 73, 75, 106, 92, 37, 110, 96, 56]]
+Epoch 0: |                                                                                                                                                        | 219/? [14:03<00:00,  0.26it/s, v_num=pmyy]train step 220; scene = [['a3b6faa8d238d993'], ['df9ba36fbe753843']]; loss = 0.049640
+Epoch 0: |                                                                                                                                                        | 220/? [14:07<00:00,  0.26it/s, v_num=pmyy]context = [[39, 71, 74], [21, 41, 51], [28, 55, 59], [2, 31, 37], [15, 25, 48], [57, 64, 93], [76, 87, 105], [31, 53, 64]]target = [[52, 43, 59], [28, 25, 49], [48, 58, 51], [26, 9, 6], [30, 32, 47], [84, 88, 59], [104, 87, 88], [42, 52, 59]]
+Epoch 0: |                                                                                                                                                        | 229/? [14:42<00:00,  0.26it/s, v_num=pmyy]train step 230; scene = [['ca04de3c55cd1ca0'], ['3d90d586b33daa63'], ['d1772c09b4b6d95f'], ['03d05f69a1cab4f8'], ['60d296908f43a97a'], ['37c400e282bc481e']]; loss = 0.064734
+Epoch 0: |                                                                                                                                                        | 230/? [14:46<00:00,  0.26it/s, v_num=pmyy]context = [[203, 204, 208, 209, 210, 233], [41, 42, 43, 52, 60, 76], [2, 8, 9, 23, 29, 31], [70, 74, 92, 97, 98, 100]]target = [[226, 223, 210, 217, 228, 222], [46, 70, 71, 75, 43, 56], [24, 23, 29, 25, 3, 21], [95, 81, 74, 73, 98, 92]]
+Epoch 0: |                                                                                                                                                        | 239/? [15:21<00:00,  0.26it/s, v_num=pmyy]train step 240; scene = [['9794641b7e015578']]; loss = 0.118654
+Epoch 0: |                                                                                                                                                        | 240/? [15:25<00:00,  0.26it/s, v_num=pmyy]context = [[62, 65, 68, 71, 80, 86, 87, 96, 99, 101, 103, 111], [134, 138, 139, 142, 151, 163, 171, 172, 173, 174, 181, 183]]target = [[87, 89, 96, 93, 103, 71, 65, 77, 63, 98, 102, 105], [165, 177, 159, 147, 138, 152, 171, 141, 181, 146, 161, 176]]
+Epoch 0: |                                                                                                                                                        | 249/? [15:59<00:00,  0.26it/s, v_num=pmyy]train step 250; scene = [['93dff1b985f2c7f9']]; loss = 0.071617
+Epoch 0: |                                                                                                                                                        | 250/? [16:03<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 250; scene = ['49b8f80c849dc341'];
+target intrinsic:  tensor(0.8891, device='cuda:0') tensor(0.8894, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8381, device='cuda:0') tensor(0.8368, device='cuda:0')
+[2026-02-24 22:43:56,273][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 250/? [16:04<00:00,  0.26it/s, v_num=pmyy]context = [[113, 115, 120, 131, 135, 137, 144, 152], [16, 20, 23, 28, 32, 35, 39, 52], [14, 18, 25, 26, 28, 29, 51, 53]]target = [[149, 139, 118, 151, 137, 141, 121, 130], [45, 51, 28, 50, 35, 37, 27, 23], [29, 26, 22, 51, 24, 34, 47, 45]]
+Epoch 0: |                                                                                                                                                        | 259/? [16:38<00:00,  0.26it/s, v_num=pmyy]train step 260; scene = [['b2288bf7003d5d4d']]; loss = 0.068989
+Epoch 0: |                                                                                                                                                        | 260/? [16:41<00:00,  0.26it/s, v_num=pmyy]context = [[206, 208, 212, 216, 221, 236], [58, 60, 68, 69, 70, 93], [15, 19, 26, 34, 41, 44], [15, 18, 23, 34, 42, 44]]target = [[225, 216, 234, 214, 207, 232], [88, 76, 91, 67, 64, 70], [20, 29, 22, 36, 23, 40], [42, 28, 31, 30, 16, 37]]
+Epoch 0: |                                                                                                                                                        | 269/? [17:16<00:00,  0.26it/s, v_num=pmyy]train step 270; scene = [['013ec74a4fde6737'], ['78e816776b064fc4'], ['1b778f72bbee1f27'], ['c71549de92ecb2e4'], ['8e16c8644efeec52'], ['35c5fc80e85db7cd'], ['34c8c62d878eca66'], ['203a5fd3a45ac4a7']]; loss = 0.073272
+Epoch 0: |                                                                                                                                                        | 270/? [17:19<00:00,  0.26it/s, v_num=pmyy]context = [[160, 174, 177, 178, 192, 200], [34, 52, 54, 56, 58, 67], [78, 85, 101, 102, 109, 118], [123, 135, 136, 144, 150, 155]]target = [[169, 195, 175, 174, 194, 189], [55, 53, 57, 48, 40, 66], [90, 108, 110, 84, 109, 83], [131, 130, 127, 124, 139, 133]]
+Epoch 0: |                                                                                                                                                        | 279/? [17:54<00:00,  0.26it/s, v_num=pmyy]train step 280; scene = [['75335793f866b96d'], ['e9d9dc952f5bbd83']]; loss = 0.052505
+Epoch 0: |                                                                                                                                                        | 280/? [17:58<00:00,  0.26it/s, v_num=pmyy]context = [[2, 31], [39, 73], [69, 98], [58, 92], [109, 146], [27, 59], [83, 114], [199, 231], [13, 50], [77, 109], [72, 105], [7, 45]]target = [[20, 26], [71, 57], [92, 77], [61, 72], [125, 129], [44, 51], [110, 101], [211, 202], [27, 17], [98, 104], [80, 89], [36, 37]]
+Epoch 0: |                                                                                                                                                        | 289/? [18:30<00:00,  0.26it/s, v_num=pmyy]train step 290; scene = [['51252022ddf74fb9'], ['8dd73309b133b8bf'], ['9e8db62a9b3cbd5e'], ['d41e59ee023e977b'], ['ce1a9465dc08ef4c'], ['e7887dec76685627']]; loss = 0.079274
+Epoch 0: |                                                                                                                                                        | 290/? [18:34<00:00,  0.26it/s, v_num=pmyy]context = [[74, 83, 85, 88, 89, 93, 100, 101, 108, 119, 122, 123], [17, 20, 23, 24, 28, 30, 32, 35, 48, 52, 64, 66]]target = [[110, 79, 95, 92, 111, 109, 99, 89, 77, 98, 81, 102], [28, 29, 52, 36, 54, 64, 49, 23, 40, 58, 31, 27]]
+Epoch 0: |                                                                                                                                                        | 299/? [19:09<00:00,  0.26it/s, v_num=pmyy]train step 300; scene = [['0c5d83212982c0ec'], ['00793a8a3b268d7c'], ['47a9b1e96499a466'], ['a1fb990016d7b3af']]; loss = 0.064584
+Epoch 0: |                                                                                                                                                        | 300/? [19:12<00:00,  0.26it/s, v_num=pmyy]context = [[83, 88, 89, 91, 92, 100, 102, 104, 107, 111, 113, 115, 129, 132, 134, 135, 140, 142, 145, 150, 154, 173, 175, 180]]target = [[115, 140, 130, 169, 114, 163, 175, 177, 104, 152, 151, 89, 154, 123, 132, 85, 122, 119, 155, 91, 117, 171, 158, 157]]
+Epoch 0: |                                                                                                                                                        | 309/? [19:46<00:00,  0.26it/s, v_num=pmyy]train step 310; scene = [['9b73ab94b5c43711'], ['8c845b940aa8244c'], ['b2789c1a5c127a02'], ['3db6c0e172d18826']]; loss = 0.069131
+Epoch 0: |                                                                                                                                                        | 310/? [19:50<00:00,  0.26it/s, v_num=pmyy]context = [[30, 47, 66], [16, 31, 46], [10, 28, 40], [43, 68, 75], [21, 43, 63], [15, 29, 46], [36, 44, 66], [17, 39, 59]]target = [[51, 41, 31], [29, 43, 44], [23, 35, 33], [64, 71, 63], [48, 59, 27], [45, 29, 27], [60, 42, 50], [34, 26, 47]]
+Epoch 0: |                                                                                                                                                        | 319/? [20:25<00:00,  0.26it/s, v_num=pmyy]train step 320; scene = [['591cd9d079cd7842'], ['3dd7802a2c93a865']]; loss = 0.089827
+Epoch 0: |                                                                                                                                                        | 320/? [20:29<00:00,  0.26it/s, v_num=pmyy]context = [[4, 22, 28, 31, 34, 37, 45, 46], [26, 28, 38, 41, 46, 55, 57, 59], [129, 139, 145, 150, 151, 158, 167, 169]]target = [[7, 41, 13, 5, 20, 27, 32, 44], [29, 32, 54, 28, 40, 52, 43, 33], [137, 143, 146, 152, 130, 131, 145, 133]]
+Epoch 0: |                                                                                                                                                        | 329/? [21:03<00:00,  0.26it/s, v_num=pmyy]train step 330; scene = [['30d9f6321281dade'], ['2a08fac923c9e50d']]; loss = 0.062383
+Epoch 0: |                                                                                                                                                        | 330/? [21:07<00:00,  0.26it/s, v_num=pmyy]context = [[4, 6, 12, 14, 24, 41, 42, 44, 45, 48, 49, 53], [102, 106, 112, 113, 118, 122, 125, 130, 134, 140, 149, 151]]target = [[50, 42, 7, 28, 11, 33, 20, 45, 25, 8, 22, 16], [135, 134, 113, 128, 139, 112, 117, 108, 122, 127, 138, 120]]
+Epoch 0: |                                                                                                                                                        | 339/? [21:41<00:00,  0.26it/s, v_num=pmyy]train step 340; scene = [['bd9f2096d355b1b8'], ['07d3325178e7a790'], ['8204d757ce43dda8']]; loss = 0.057876
+Epoch 0: |                                                                                                                                                        | 340/? [21:45<00:00,  0.26it/s, v_num=pmyy]context = [[87, 88, 93, 105, 107, 111, 116, 118, 123, 131, 132, 136], [104, 111, 116, 121, 122, 127, 131, 132, 136, 137, 142, 153]]target = [[114, 108, 107, 130, 95, 129, 118, 104, 116, 96, 109, 99], [127, 106, 145, 129, 114, 109, 143, 150, 111, 133, 146, 122]]
+Epoch 0: |                                                                                                                                                        | 349/? [22:19<00:00,  0.26it/s, v_num=pmyy]train step 350; scene = [['9d0bfbe5b7f98545'], ['06a16655c8e8ad9c']]; loss = 0.098613
+Epoch 0: |                                                                                                                                                        | 350/? [22:23<00:00,  0.26it/s, v_num=pmyy]context = [[33, 52, 53, 56, 58, 64, 73, 75], [100, 102, 115, 117, 130, 134, 139, 142], [223, 225, 240, 242, 255, 256, 263, 267]]target = [[41, 60, 53, 63, 72, 38, 37, 73], [122, 141, 116, 138, 110, 119, 118, 135], [260, 237, 261, 264, 238, 255, 243, 253]]
+Epoch 0: |                                                                                                                                                        | 359/? [22:58<00:00,  0.26it/s, v_num=pmyy]train step 360; scene = [['73b27f4f150327af'], ['169aaaf51ef3849c'], ['068a8406f1a383d8'], ['a9936b77895f33b3']]; loss = 0.070163
+Epoch 0: |                                                                                                                                                        | 360/? [23:02<00:00,  0.26it/s, v_num=pmyy]context = [[0, 14, 18, 23, 25, 32, 42, 45, 46, 49, 53, 55, 59, 60, 61, 74, 78, 80, 84, 85, 89, 90, 92, 97]]target = [[50, 81, 69, 64, 38, 86, 91, 16, 27, 33, 44, 90, 48, 23, 8, 79, 6, 39, 42, 36, 82, 78, 59, 54]]
+Epoch 0: |                                                                                                                                                        | 369/? [23:35<00:00,  0.26it/s, v_num=pmyy]train step 370; scene = [['8673faf0a9d48165'], ['99a0790d72e6c2af'], ['6cbbe9075b0d2138']]; loss = 0.055459
+Epoch 0: |                                                                                                                                                        | 370/? [23:39<00:00,  0.26it/s, v_num=pmyy]context = [[61, 63, 65, 72, 73, 74, 76, 89, 92, 94, 100, 101, 120, 124, 126, 127, 136, 140, 144, 145, 150, 152, 156, 158]]target = [[121, 100, 147, 98, 148, 143, 126, 63, 73, 141, 79, 119, 115, 106, 153, 101, 120, 71, 91, 62, 105, 84, 151, 145]]
+Epoch 0: |                                                                                                                                                        | 379/? [24:14<00:00,  0.26it/s, v_num=pmyy]train step 380; scene = [['656330f47c5df010'], ['6dfb89a98e14ca66']]; loss = 0.060643
+Epoch 0: |                                                                                                                                                        | 380/? [24:17<00:00,  0.26it/s, v_num=pmyy]context = [[210, 215, 223, 225, 231, 234, 235, 249], [28, 32, 35, 53, 54, 56, 57, 68], [149, 150, 154, 155, 174, 177, 181, 183]]target = [[241, 236, 218, 230, 223, 228, 239, 227], [55, 46, 44, 49, 30, 33, 60, 45], [157, 160, 161, 164, 170, 181, 168, 166]]
+Epoch 0: |                                                                                                                                                        | 389/? [24:52<00:00,  0.26it/s, v_num=pmyy]train step 390; scene = [['723f94d150ab09f2'], ['393cdfb7e832d285'], ['14900b71ac66b7bd'], ['452625cd6b071b87'], ['281599bbab3e73dd'], ['0a2b42e240751d33']]; loss = 0.077013
+Epoch 0: |                                                                                                                                                        | 390/? [24:56<00:00,  0.26it/s, v_num=pmyy]context = [[48, 52, 80], [14, 47, 60], [142, 159, 187], [78, 120, 121], [6, 24, 45], [47, 68, 90], [15, 19, 48], [9, 20, 52]]target = [[51, 50, 70], [45, 15, 56], [172, 148, 157], [120, 98, 109], [20, 33, 11], [83, 84, 73], [45, 30, 32], [16, 36, 45]]
+Epoch 0: |                                                                                                                                                        | 399/? [25:29<00:00,  0.26it/s, v_num=pmyy]train step 400; scene = [['4303746d8f23f16b'], ['0fe8246bb7e2fe40'], ['b7d77240852d6a52'], ['6e5505414fd63528'], ['44985936f68c3a36'], ['1550f1b4fff1f2a4'], ['cea3d842c3285c65'], ['b34bb5f53856d34f']]; loss = 0.083431
+Epoch 0: |                                                                                                                                                        | 400/? [25:33<00:00,  0.26it/s, v_num=pmyy]context = [[9, 20, 28, 31, 34, 53], [189, 196, 216, 223, 230, 232], [74, 77, 86, 88, 98, 118], [132, 145, 157, 168, 179, 181]]target = [[51, 32, 14, 50, 45, 34], [212, 226, 201, 205, 215, 222], [93, 113, 107, 94, 83, 79], [161, 139, 177, 151, 156, 172]]
+[2026-02-24 22:53:29,318][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 409/? [26:08<00:00,  0.26it/s, v_num=pmyy]train step 410; scene = [['144e1ec915e46d29'], ['b290b6a0afa1dac7'], ['b3d84dba6581c3d9']]; loss = 0.063015
+Epoch 0: |                                                                                                                                                        | 410/? [26:12<00:00,  0.26it/s, v_num=pmyy]context = [[26, 40, 41, 50, 55, 58], [159, 174, 179, 192, 198, 201], [2, 3, 5, 7, 24, 40], [187, 204, 205, 208, 217, 221]]target = [[38, 32, 44, 52, 42, 27], [198, 191, 196, 162, 163, 169], [33, 17, 23, 37, 15, 39], [214, 211, 195, 220, 210, 192]]
+Epoch 0: |                                                                                                                                                        | 419/? [26:46<00:00,  0.26it/s, v_num=pmyy]train step 420; scene = [['a1dff9c50d92dc9c']]; loss = 0.056775
+Epoch 0: |                                                                                                                                                        | 420/? [26:50<00:00,  0.26it/s, v_num=pmyy]context = [[11, 13, 14, 26, 27, 41, 46, 51], [151, 155, 161, 167, 172, 176, 191, 195], [49, 52, 55, 57, 59, 74, 79, 90]]target = [[25, 44, 17, 33, 37, 24, 38, 40], [193, 191, 194, 162, 158, 176, 163, 171], [87, 85, 67, 58, 86, 68, 83, 78]]
+Epoch 0: |                                                                                                                                                        | 429/? [27:24<00:00,  0.26it/s, v_num=pmyy]train step 430; scene = [['36664e22fd10a141'], ['0474328f4cefd619']]; loss = 0.046890
+Epoch 0: |                                                                                                                                                        | 430/? [27:27<00:00,  0.26it/s, v_num=pmyy]context = [[123, 130, 131, 132, 136, 154, 169, 173], [35, 36, 47, 49, 65, 67, 76, 86], [33, 34, 39, 41, 59, 64, 65, 73]]target = [[145, 147, 167, 130, 146, 170, 138, 137], [59, 38, 49, 42, 70, 76, 39, 40], [56, 44, 36, 35, 38, 45, 48, 58]]
+Epoch 0: |                                                                                                                                                        | 439/? [28:02<00:00,  0.26it/s, v_num=pmyy]train step 440; scene = [['342099a48847f4f6'], ['5ad0327426e3718b'], ['c25b314716aa6b10'], ['c91e2b5399b14430'], ['e1d9ade67e615bd8'], ['46df912c9748215b']]; loss = 0.056666
+Epoch 0: |                                                                                                                                                        | 440/? [28:06<00:00,  0.26it/s, v_num=pmyy]context = [[0, 33], [10, 57], [36, 82], [46, 79], [62, 99], [33, 79], [11, 52], [1, 34], [0, 37], [26, 62], [18, 67], [96, 138]]target = [[22, 18], [37, 23], [44, 54], [69, 50], [68, 69], [69, 56], [49, 41], [27, 30], [31, 17], [42, 32], [59, 54], [108, 115]]
+Epoch 0: |                                                                                                                                                        | 449/? [28:41<00:00,  0.26it/s, v_num=pmyy]train step 450; scene = [['e19c6facac1c9624'], ['5244830b7357365b'], ['b80c2522b1070e2f'], ['6ea0ff32c8ea695c'], ['2f311b2bbbeb5940'], ['3f7992e72a096099']]; loss = 0.058994
+Epoch 0: |                                                                                                                                                        | 450/? [28:45<00:00,  0.26it/s, v_num=pmyy]context = [[41, 42, 44, 49, 57, 66, 68, 73, 75, 76, 83, 91], [4, 5, 15, 24, 32, 35, 37, 38, 45, 47, 49, 56]]target = [[79, 77, 90, 46, 81, 83, 56, 68, 89, 70, 84, 57], [7, 8, 40, 37, 32, 10, 25, 35, 23, 43, 48, 52]]
+Epoch 0: |                                                                                                                                                        | 459/? [29:19<00:00,  0.26it/s, v_num=pmyy]train step 460; scene = [['46fb6702ed1b9967'], ['bdc3f978b0d3aa8f']]; loss = 0.052727
+Epoch 0: |                                                                                                                                                        | 460/? [29:23<00:00,  0.26it/s, v_num=pmyy]context = [[65, 66, 69, 78, 84, 85, 90, 91, 99, 103, 112, 116], [39, 40, 52, 63, 64, 68, 72, 73, 83, 87, 90, 92]]target = [[99, 86, 101, 76, 84, 77, 105, 66, 89, 75, 100, 106], [84, 61, 68, 82, 59, 47, 51, 71, 85, 56, 57, 50]]
+Epoch 0: |                                                                                                                                                        | 469/? [29:56<00:00,  0.26it/s, v_num=pmyy]train step 470; scene = [['2c88995e05a17d17'], ['2b1f47da224557a3'], ['62216d162b71b5b4'], ['61d39a97cb69d99f'], ['42000d5a83b48ee4'], ['cc8480640599f9f3']]; loss = 0.065315
+Epoch 0: |                                                                                                                                                        | 470/? [30:00<00:00,  0.26it/s, v_num=pmyy]context = [[1, 6, 15, 26, 28, 41, 46, 52], [0, 1, 3, 5, 14, 26, 32, 37], [52, 56, 58, 79, 87, 88, 93, 97]]target = [[24, 22, 32, 14, 42, 21, 12, 37], [14, 10, 11, 25, 12, 34, 7, 32], [62, 60, 80, 87, 63, 55, 78, 88]]
+Epoch 0: |                                                                                                                                                        | 479/? [30:34<00:00,  0.26it/s, v_num=pmyy]train step 480; scene = [['2e3bb7fb33e1ed30'], ['7460f503eb18fa6a'], ['bde49071d2088850'], ['e80016be3043dfa4']]; loss = 0.072693
+Epoch 0: |                                                                                                                                                        | 480/? [30:38<00:00,  0.26it/s, v_num=pmyy]context = [[9, 15, 37, 47], [0, 5, 7, 35], [101, 113, 134, 153], [64, 88, 93, 107], [18, 44, 56, 62], [31, 46, 58, 75]]target = [[45, 27, 28, 18], [10, 8, 14, 3], [128, 143, 110, 129], [87, 84, 90, 94], [29, 31, 40, 36], [52, 72, 33, 56]]
+Epoch 0: |                                                                                                                                                        | 489/? [31:12<00:00,  0.26it/s, v_num=pmyy]train step 490; scene = [['83085493f4bc18d2']]; loss = 0.076205
+Epoch 0: |                                                                                                                                                        | 490/? [31:16<00:00,  0.26it/s, v_num=pmyy]context = [[0, 3, 7, 12, 38, 40, 45, 48], [25, 28, 57, 62, 69, 70, 76, 78], [25, 34, 37, 39, 41, 53, 65, 67]]target = [[15, 7, 39, 47, 42, 26, 25, 5], [63, 40, 59, 31, 75, 34, 47, 39], [58, 35, 53, 33, 65, 44, 38, 31]]
+Epoch 0: |                                                                                                                                                        | 499/? [31:50<00:00,  0.26it/s, v_num=pmyy]train step 500; scene = [['1241bcb5732a9502'], ['d33a9e90e1416efb']]; loss = 0.033324
+Epoch 0: |                                                                                                                                                        | 500/? [31:54<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 500; scene = ['73d6f935f31b3fd4'];
+target intrinsic:  tensor(0.8576, device='cuda:0') tensor(0.8579, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8379, device='cuda:0') tensor(0.8341, device='cuda:0')
+[2026-02-24 22:59:47,452][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 500/? [31:56<00:00,  0.26it/s, v_num=pmyy]context = [[1, 21, 23, 29, 39, 48, 49, 52], [17, 18, 20, 28, 34, 39, 45, 51], [99, 103, 106, 110, 116, 136, 138, 140]]target = [[4, 46, 41, 43, 40, 37, 39, 25], [30, 48, 22, 42, 44, 31, 38, 37], [112, 108, 133, 116, 111, 125, 123, 127]]
+Epoch 0: |                                                                                                                                                        | 509/? [32:28<00:00,  0.26it/s, v_num=pmyy]train step 510; scene = [['0eed4548041bea8e'], ['277a96ce456580f4']]; loss = 0.047844
+Epoch 0: |                                                                                                                                                        | 510/? [32:32<00:00,  0.26it/s, v_num=pmyy]context = [[3, 13, 21, 24, 27, 28, 30, 37, 42, 44, 46, 55, 62, 65, 69, 71, 73, 80, 83, 84, 89, 95, 97, 100]]target = [[51, 81, 54, 89, 55, 45, 10, 19, 21, 68, 48, 67, 33, 69, 8, 25, 64, 83, 37, 29, 24, 82, 99, 46]]
+Epoch 0: |                                                                                                                                                        | 519/? [33:07<00:00,  0.26it/s, v_num=pmyy]train step 520; scene = [['625e3aa0ff734714'], ['395802511d26f32e'], ['39343936591c28de']]; loss = 0.077093
+Epoch 0: |                                                                                                                                                        | 520/? [33:11<00:00,  0.26it/s, v_num=pmyy]context = [[13, 16, 17, 22, 25, 30, 35, 39, 53, 56, 62, 68, 78, 87, 90, 92, 93, 95, 99, 100, 103, 106, 107, 110]]target = [[39, 96, 58, 28, 16, 59, 44, 17, 26, 83, 103, 31, 57, 35, 107, 51, 27, 77, 46, 30, 100, 91, 93, 97]]
+Epoch 0: |                                                                                                                                                        | 529/? [33:44<00:00,  0.26it/s, v_num=pmyy]train step 530; scene = [['0c199c575b699444'], ['70d878da47f984e4'], ['15f77c76ea744f99'], ['e54b5eec8cc47776'], ['1969ed97e68d83d9'], ['c7cf9b63dc3e5830'], ['bcef3076b93012b1'], ['ab2680bf91942e23']]; loss = 0.073779
+Epoch 0: |                                                                                                                                                        | 530/? [33:48<00:00,  0.26it/s, v_num=pmyy]context = [[78, 84, 85, 92, 124, 127], [14, 15, 24, 52, 64, 68], [1, 6, 16, 20, 22, 40], [33, 51, 61, 65, 66, 70]]target = [[104, 103, 94, 96, 88, 82], [15, 53, 58, 47, 56, 60], [31, 10, 18, 26, 23, 24], [37, 69, 62, 65, 49, 64]]
+Epoch 0: |                                                                                                                                                        | 539/? [34:23<00:00,  0.26it/s, v_num=pmyy]train step 540; scene = [['a071d9276f6a9272']]; loss = 0.062914
+Epoch 0: |                                                                                                                                                        | 540/? [34:27<00:00,  0.26it/s, v_num=pmyy]context = [[35, 46, 47, 49, 52, 54, 55, 56, 57, 65, 66, 85], [51, 54, 55, 59, 68, 77, 82, 86, 87, 93, 105, 106]]target = [[69, 68, 75, 48, 44, 79, 53, 74, 60, 81, 72, 71], [77, 73, 88, 59, 67, 61, 102, 62, 93, 75, 95, 105]]
+Epoch 0: |                                                                                                                                                        | 549/? [35:01<00:00,  0.26it/s, v_num=pmyy]train step 550; scene = [['836250796ea45b6c']]; loss = 0.086115
+Epoch 0: |                                                                                                                                                        | 550/? [35:05<00:00,  0.26it/s, v_num=pmyy]context = [[30, 34, 39, 45, 47, 79, 80, 83], [16, 21, 22, 24, 33, 42, 51, 60], [115, 128, 129, 138, 141, 150, 156, 163]]target = [[79, 51, 62, 77, 46, 54, 33, 49], [27, 59, 20, 41, 21, 48, 58, 30], [133, 125, 124, 159, 122, 129, 155, 119]]
+Epoch 0: |                                                                                                                                                        | 559/? [35:40<00:00,  0.26it/s, v_num=pmyy]train step 560; scene = [['d70ca840b3c5aec9'], ['65c3f29c43dd1e63'], ['d3917d0a1eda2a1f'], ['5c83dfc8f9ab44fa']]; loss = 0.049655
+Epoch 0: |                                                                                                                                                        | 560/? [35:43<00:00,  0.26it/s, v_num=pmyy]context = [[62, 66, 67, 68, 71, 87, 88, 92, 100, 108, 110, 122], [153, 160, 165, 168, 171, 177, 178, 181, 183, 185, 200, 205]]target = [[95, 104, 83, 68, 115, 81, 74, 106, 119, 79, 90, 72], [204, 165, 196, 185, 197, 198, 195, 187, 155, 188, 154, 201]]
+Epoch 0: |                                                                                                                                                        | 569/? [36:18<00:00,  0.26it/s, v_num=pmyy]train step 570; scene = [['9d8ddcdbe1f7ac42'], ['721df0f45094ca34'], ['fdbfe35f5940d3ad']]; loss = 0.045652
+Epoch 0: |                                                                                                                                                        | 570/? [36:22<00:00,  0.26it/s, v_num=pmyy]context = [[23, 34, 41, 42, 48, 54, 56, 59, 60, 63, 66, 71, 74, 79, 88, 90, 92, 93, 95, 97, 110, 111, 119, 120]]target = [[30, 27, 28, 53, 31, 75, 85, 54, 77, 111, 76, 25, 44, 52, 33, 41, 69, 89, 73, 68, 26, 93, 83, 119]]
+Epoch 0: |                                                                                                                                                        | 579/? [36:56<00:00,  0.26it/s, v_num=pmyy]train step 580; scene = [['88a0267e41b851f0'], ['df71fbb70b19cbc3'], ['1c713c10ecf5a0c9']]; loss = 0.050310
+Epoch 0: |                                                                                                                                                        | 580/? [36:59<00:00,  0.26it/s, v_num=pmyy]context = [[9, 22, 36, 39, 43, 68], [70, 85, 95, 98, 116, 121], [10, 28, 45, 47, 50, 57], [132, 144, 154, 158, 168, 188]]target = [[67, 45, 11, 46, 50, 30], [114, 94, 71, 104, 117, 76], [22, 12, 24, 51, 50, 13], [180, 169, 183, 173, 182, 171]]
+Epoch 0: |                                                                                                                                                        | 589/? [37:34<00:00,  0.26it/s, v_num=pmyy]train step 590; scene = [['3f732b63cdd0729e'], ['9be3165beb073d95'], ['42a6c835ff830674'], ['f928d960cbfae15a'], ['140b10a4f6bb5aa5'], ['cc8e19c8ad1846f4']]; loss = 0.078254
+Epoch 0: |                                                                                                                                                        | 590/? [37:38<00:00,  0.26it/s, v_num=pmyy]context = [[198, 200, 201, 204, 209, 210, 224, 234], [6, 14, 18, 23, 25, 27, 58, 59], [13, 29, 40, 51, 55, 66, 70, 71]]target = [[203, 208, 227, 217, 222, 223, 200, 211], [11, 46, 45, 19, 20, 18, 47, 7], [16, 51, 40, 15, 55, 44, 23, 61]]
+Epoch 0: |                                                                                                                                                        | 599/? [38:13<00:00,  0.26it/s, v_num=pmyy]train step 600; scene = [['cb734fdc69e9900e']]; loss = 0.049314
+Epoch 0: |                                                                                                                                                        | 600/? [38:16<00:00,  0.26it/s, v_num=pmyy]context = [[10, 16, 24, 31, 33, 34, 36, 41, 49, 59, 63, 67], [160, 162, 167, 172, 173, 180, 183, 187, 189, 205, 212, 214]]target = [[32, 54, 42, 14, 39, 60, 66, 33, 62, 22, 47, 49], [182, 199, 163, 179, 203, 200, 174, 209, 164, 172, 194, 167]]
+[2026-02-24 23:06:12,751][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 609/? [38:52<00:00,  0.26it/s, v_num=pmyy]train step 610; scene = [['ed9409fa128e193b'], ['a5c03b0c5fb7203e']]; loss = 0.037029
+Epoch 0: |                                                                                                                                                        | 610/? [38:56<00:00,  0.26it/s, v_num=pmyy]context = [[36, 37, 39, 50, 51, 52, 57, 59, 72, 73, 86, 94], [45, 46, 54, 57, 66, 75, 77, 81, 83, 89, 97, 101]]target = [[47, 37, 89, 79, 92, 78, 81, 90, 38, 56, 84, 69], [84, 91, 85, 69, 79, 53, 55, 48, 87, 49, 92, 63]]
+Epoch 0: |                                                                                                                                                        | 619/? [39:30<00:00,  0.26it/s, v_num=pmyy]train step 620; scene = [['7898a828b7203ca4'], ['9c269fce78f0dd27'], ['e1e317857deb7afc'], ['30124191dafb3383'], ['c39f1a9a73797efe'], ['a640a55439a43108']]; loss = 0.046627
+Epoch 0: |                                                                                                                                                        | 620/? [39:34<00:00,  0.26it/s, v_num=pmyy]context = [[2, 15, 16, 21, 26, 29, 47, 56], [6, 12, 15, 37, 45, 47, 50, 68], [9, 18, 23, 26, 39, 48, 50, 60]]target = [[38, 7, 32, 36, 16, 21, 53, 22], [50, 67, 28, 43, 15, 41, 44, 23], [12, 54, 16, 24, 58, 32, 30, 47]]
+Epoch 0: |                                                                                                                                                        | 629/? [40:09<00:00,  0.26it/s, v_num=pmyy]train step 630; scene = [['dd3bbf1f7f832e83'], ['0ff7277275e55096'], ['5f45c360d76a3b12']]; loss = 0.038329
+Epoch 0: |                                                                                                                                                        | 630/? [40:13<00:00,  0.26it/s, v_num=pmyy]context = [[29, 42, 44, 45, 76, 81, 87, 93], [77, 84, 94, 96, 100, 102, 121, 125], [67, 68, 70, 89, 90, 91, 96, 107]]target = [[83, 55, 36, 89, 45, 62, 90, 63], [106, 92, 114, 99, 117, 82, 119, 107], [99, 68, 75, 92, 79, 85, 101, 90]]
+Epoch 0: |                                                                                                                                                        | 639/? [40:48<00:00,  0.26it/s, v_num=pmyy]train step 640; scene = [['867edbda9bb8ef59'], ['1d83764e77e159d8'], ['e318dafa4071cef9'], ['169f09c33ee35289']]; loss = 0.078373
+Epoch 0: |                                                                                                                                                        | 640/? [40:51<00:00,  0.26it/s, v_num=pmyy]context = [[52, 57, 67, 75, 77, 80, 85, 86, 99, 105, 112, 117], [6, 11, 22, 23, 32, 39, 42, 47, 48, 53, 63, 64]]target = [[114, 60, 111, 110, 57, 69, 68, 80, 62, 99, 77, 90], [27, 15, 46, 56, 51, 60, 41, 63, 49, 16, 9, 43]]
+Epoch 0: |                                                                                                                                                        | 649/? [41:25<00:00,  0.26it/s, v_num=pmyy]train step 650; scene = [['23668135f32e0126'], ['daca15248046e480'], ['174ebd189316bd92']]; loss = 0.048676
+Epoch 0: |                                                                                                                                                        | 650/? [41:29<00:00,  0.26it/s, v_num=pmyy]context = [[6, 23, 25, 26, 28, 32, 37, 38, 39, 40, 50, 57], [0, 2, 3, 11, 17, 19, 24, 25, 38, 40, 45, 53]]target = [[15, 47, 23, 31, 12, 30, 20, 55, 33, 11, 22, 10], [51, 9, 13, 34, 52, 30, 26, 45, 38, 27, 11, 14]]
+Epoch 0: |                                                                                                                                                        | 659/? [42:03<00:00,  0.26it/s, v_num=pmyy]train step 660; scene = [['60499200285c9abe']]; loss = 0.036310
+Epoch 0: |                                                                                                                                                        | 660/? [42:07<00:00,  0.26it/s, v_num=pmyy]context = [[57, 65, 76, 86, 93, 99, 100, 122], [31, 34, 42, 47, 60, 80, 81, 83], [13, 16, 19, 30, 33, 49, 50, 55]]target = [[85, 75, 88, 114, 107, 70, 116, 79], [82, 35, 67, 56, 68, 61, 74, 58], [54, 44, 23, 36, 31, 39, 45, 28]]
+Epoch 0: |                                                                                                                                                        | 669/? [42:41<00:00,  0.26it/s, v_num=pmyy]train step 670; scene = [['7665ff641f430aa5']]; loss = 0.039078
+Epoch 0: |                                                                                                                                                        | 670/? [42:45<00:00,  0.26it/s, v_num=pmyy]context = [[50, 57, 61, 69, 82, 87, 98, 101], [26, 33, 44, 47, 53, 56, 62, 66], [11, 33, 39, 48, 56, 67, 75, 78]]target = [[67, 96, 71, 75, 73, 57, 55, 89], [65, 39, 63, 59, 60, 47, 55, 41], [18, 56, 21, 64, 32, 59, 33, 25]]
+Epoch 0: |                                                                                                                                                        | 679/? [43:19<00:00,  0.26it/s, v_num=pmyy]train step 680; scene = [['43c939b11c5fed4a']]; loss = 0.080682
+Epoch 0: |                                                                                                                                                        | 680/? [43:23<00:00,  0.26it/s, v_num=pmyy]context = [[47, 54, 60, 64, 79, 80, 83, 93], [12, 13, 21, 30, 34, 45, 53, 55], [57, 75, 78, 82, 96, 103, 113, 117]]target = [[80, 62, 75, 54, 57, 92, 53, 58], [38, 32, 15, 44, 42, 51, 27, 39], [81, 87, 106, 94, 99, 103, 73, 79]]
+Epoch 0: |                                                                                                                                                        | 689/? [43:56<00:00,  0.26it/s, v_num=pmyy]train step 690; scene = [['1848b8b363d0d2b9'], ['afe6b05d0554a880']]; loss = 0.042606
+Epoch 0: |                                                                                                                                                        | 690/? [44:00<00:00,  0.26it/s, v_num=pmyy]context = [[12, 18, 20, 27, 36, 37, 38, 43, 48, 49, 52, 58, 59, 62, 67, 81, 83, 86, 93, 95, 101, 104, 108, 109]]target = [[34, 24, 43, 31, 87, 30, 51, 54, 52, 94, 86, 21, 44, 97, 61, 95, 38, 60, 49, 73, 41, 19, 65, 67]]
+Epoch 0: |                                                                                                                                                        | 699/? [44:34<00:00,  0.26it/s, v_num=pmyy]train step 700; scene = [['674ef9fb9cf20f9f'], ['8624ee0839cb6e4c'], ['caed302f388b799f']]; loss = 0.046811
+Epoch 0: |                                                                                                                                                        | 700/? [44:37<00:00,  0.26it/s, v_num=pmyy]context = [[31, 32, 41, 46, 53, 54, 55, 57, 65, 68, 73, 74, 80, 85, 100, 105, 108, 109, 113, 114, 116, 118, 126, 128]]target = [[117, 52, 85, 57, 37, 45, 78, 100, 125, 35, 113, 66, 105, 103, 61, 83, 88, 40, 116, 60, 79, 32, 102, 107]]
+Epoch 0: |                                                                                                                                                        | 709/? [45:12<00:00,  0.26it/s, v_num=pmyy]train step 710; scene = [['db6cd90de8fee2ff'], ['7a20ba81fb778529'], ['970350268b239272']]; loss = 0.046765
+Epoch 0: |                                                                                                                                                        | 710/? [45:16<00:00,  0.26it/s, v_num=pmyy]context = [[3, 12, 16, 31, 41, 45, 49, 53, 54, 55, 67, 69], [208, 211, 217, 221, 222, 227, 230, 231, 239, 247, 252, 267]]target = [[38, 34, 68, 25, 44, 63, 48, 7, 42, 6, 64, 28], [228, 253, 259, 266, 244, 238, 249, 234, 232, 241, 220, 213]]
+Epoch 0: |                                                                                                                                                        | 719/? [45:50<00:00,  0.26it/s, v_num=pmyy]train step 720; scene = [['f63d2df8871ce70c'], ['0fdeda15097ed4a4']]; loss = 0.041310
+Epoch 0: |                                                                                                                                                        | 720/? [45:54<00:00,  0.26it/s, v_num=pmyy]context = [[55, 57, 71, 72, 75, 85, 90, 92, 98, 99, 102, 108, 112, 116, 125, 126, 130, 131, 135, 141, 144, 145, 146, 152]]target = [[61, 145, 86, 66, 62, 119, 100, 139, 105, 125, 58, 101, 140, 132, 118, 128, 65, 141, 151, 78, 104, 107, 150, 138]]
+Epoch 0: |                                                                                                                                                        | 729/? [46:29<00:00,  0.26it/s, v_num=pmyy]train step 730; scene = [['232abb354c423e81'], ['d34926c73ae1277e']]; loss = 0.034553
+Epoch 0: |                                                                                                                                                        | 730/? [46:32<00:00,  0.26it/s, v_num=pmyy]context = [[45, 50, 54, 58, 66, 70, 72, 78, 85, 88, 94, 105, 109, 110, 111, 117, 120, 122, 126, 127, 128, 133, 135, 142]]target = [[95, 96, 50, 114, 135, 124, 104, 100, 49, 119, 139, 62, 92, 123, 58, 46, 57, 112, 116, 90, 54, 101, 85, 81]]
+Epoch 0: |                                                                                                                                                        | 739/? [47:06<00:00,  0.26it/s, v_num=pmyy]train step 740; scene = [['19f7966006ad778d'], ['dde0212418df7ca9'], ['ad75e36b74f6b033'], ['ea97e5ae55e56208'], ['9d29b0289133ab4e'], ['282938f90821bdef']]; loss = 0.095513
+Epoch 0: |                                                                                                                                                        | 740/? [47:10<00:00,  0.26it/s, v_num=pmyy]context = [[8, 13, 19, 22, 30, 31, 36, 45, 46, 48, 54, 57, 58, 59, 60, 61, 65, 73, 77, 83, 86, 88, 93, 105]]target = [[46, 101, 90, 61, 31, 23, 37, 95, 18, 67, 32, 100, 93, 35, 89, 45, 10, 70, 60, 97, 85, 81, 66, 79]]
+Epoch 0: |                                                                                                                                                        | 749/? [47:44<00:00,  0.26it/s, v_num=pmyy]train step 750; scene = [['f85921f42c5d98d7'], ['a95dacbd3ea3db36']]; loss = 0.044777
+Epoch 0: |                                                                                                                                                        | 750/? [47:48<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 750; scene = ['91fda69e1cda4602'];
+target intrinsic:  tensor(0.8937, device='cuda:0') tensor(0.8939, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9484, device='cuda:0') tensor(0.9495, device='cuda:0')
+[2026-02-24 23:15:41,458][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 750/? [47:50<00:00,  0.26it/s, v_num=pmyy]context = [[1, 38, 46, 49, 52, 59], [213, 218, 227, 233, 252, 257], [99, 105, 109, 124, 125, 153], [8, 17, 20, 37, 62, 65]]target = [[54, 19, 39, 57, 7, 4], [252, 219, 223, 216, 248, 237], [148, 136, 102, 103, 149, 132], [12, 52, 13, 44, 63, 39]]
+Epoch 0: |                                                                                                                                                        | 759/? [48:23<00:00,  0.26it/s, v_num=pmyy]train step 760; scene = [['75617c97bff1e873'], ['ff02f88545dfa566']]; loss = 0.037410
+Epoch 0: |                                                                                                                                                        | 760/? [48:27<00:00,  0.26it/s, v_num=pmyy]context = [[71, 73, 111], [2, 14, 56], [29, 82, 88], [120, 144, 168], [181, 214, 250], [0, 10, 71], [12, 53, 54], [14, 27, 87]]target = [[101, 74, 84], [43, 51, 38], [82, 55, 72], [156, 139, 147], [198, 214, 207], [22, 5, 40], [36, 13, 39], [68, 35, 70]]
+Epoch 0: |                                                                                                                                                        | 769/? [49:02<00:00,  0.26it/s, v_num=pmyy]train step 770; scene = [['62b0d4ee613af70f'], ['f7926eb1096de201'], ['c63b37ec347f0d0e'], ['b43d9f7c70f5caa0']]; loss = 0.072210
+Epoch 0: |                                                                                                                                                        | 770/? [49:06<00:00,  0.26it/s, v_num=pmyy]context = [[13, 29, 30, 31, 32, 34, 41, 42, 46, 52, 63, 66], [146, 149, 157, 166, 177, 180, 181, 193, 197, 204, 211, 213]]target = [[35, 55, 52, 59, 24, 51, 63, 18, 42, 44, 33, 61], [183, 158, 171, 175, 163, 196, 177, 181, 210, 167, 187, 174]]
+Epoch 0: |                                                                                                                                                        | 779/? [49:41<00:00,  0.26it/s, v_num=pmyy]train step 780; scene = [['b41f4db8b8a42a71']]; loss = 0.065992
+Epoch 0: |                                                                                                                                                        | 780/? [49:45<00:00,  0.26it/s, v_num=pmyy]context = [[43, 50, 53, 55, 57, 58, 59, 60, 68, 81, 94, 96, 102, 106, 108, 110, 112, 121, 125, 126, 130, 131, 136, 140]]target = [[74, 122, 93, 98, 70, 84, 49, 136, 77, 117, 135, 138, 123, 89, 119, 45, 129, 105, 50, 58, 63, 103, 82, 121]]
+Epoch 0: |                                                                                                                                                        | 789/? [50:19<00:00,  0.26it/s, v_num=pmyy]train step 790; scene = [['d79666d294813d8e']]; loss = 0.156959
+Epoch 0: |                                                                                                                                                        | 790/? [50:23<00:00,  0.26it/s, v_num=pmyy]context = [[59, 79, 88, 93, 100, 107], [3, 16, 19, 31, 46, 51], [41, 49, 54, 56, 72, 83], [226, 229, 235, 244, 255, 272]]target = [[103, 98, 81, 75, 74, 72], [48, 44, 20, 31, 15, 46], [64, 82, 70, 50, 47, 60], [251, 249, 271, 250, 259, 260]]
+Epoch 0: |                                                                                                                                                        | 799/? [50:57<00:00,  0.26it/s, v_num=pmyy]train step 800; scene = [['cb797cd30542e55c']]; loss = 0.049692
+Epoch 0: |                                                                                                                                                        | 800/? [51:01<00:00,  0.26it/s, v_num=pmyy]context = [[10, 19, 20, 39, 42, 55], [1, 6, 11, 20, 30, 42], [0, 3, 4, 19, 42, 45], [21, 35, 52, 82, 85, 88]]target = [[38, 39, 26, 21, 37, 40], [11, 40, 18, 27, 3, 24], [43, 33, 8, 24, 5, 19], [83, 71, 66, 32, 63, 47]]
+[2026-02-24 23:18:57,458][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 809/? [51:36<00:00,  0.26it/s, v_num=pmyy]train step 810; scene = [['9bd08fc9288bef8b']]; loss = 0.054979
+Epoch 0: |                                                                                                                                                        | 810/? [51:40<00:00,  0.26it/s, v_num=pmyy]context = [[5, 6, 14, 17, 19, 23, 30, 37, 40, 49, 53, 55], [3, 6, 15, 17, 27, 34, 38, 40, 42, 46, 56, 65]]target = [[16, 8, 48, 23, 44, 49, 10, 7, 18, 50, 22, 15], [34, 18, 9, 42, 31, 38, 45, 26, 43, 6, 49, 44]]
+Epoch 0: |                                                                                                                                                        | 819/? [52:14<00:00,  0.26it/s, v_num=pmyy]train step 820; scene = [['49952f737be91dd2']]; loss = 0.079824
+Epoch 0: |                                                                                                                                                        | 820/? [52:18<00:00,  0.26it/s, v_num=pmyy]context = [[22, 23, 26, 28, 31, 43, 51, 53, 57, 65, 67, 68, 71, 76, 77, 78, 80, 84, 88, 90, 107, 112, 116, 119]]target = [[35, 45, 31, 72, 24, 71, 89, 54, 25, 61, 44, 97, 56, 47, 82, 87, 50, 110, 108, 75, 103, 67, 86, 117]]
+Epoch 0: |                                                                                                                                                        | 829/? [52:51<00:00,  0.26it/s, v_num=pmyy]train step 830; scene = [['2f9c9d1b56eb7f75'], ['1c392cc98b3a7642']]; loss = 0.060911
+Epoch 0: |                                                                                                                                                        | 830/? [52:55<00:00,  0.26it/s, v_num=pmyy]context = [[42, 67, 90, 119], [18, 54, 72, 90], [43, 67, 73, 102], [2, 22, 40, 43], [23, 40, 52, 67], [19, 37, 75, 78]]target = [[86, 112, 87, 94], [59, 40, 20, 47], [58, 97, 66, 77], [34, 19, 32, 17], [62, 37, 35, 54], [54, 22, 31, 50]]
+Epoch 0: |                                                                                                                                                        | 839/? [53:29<00:00,  0.26it/s, v_num=pmyy]train step 840; scene = [['c51c7bc0c8151abb'], ['a0d16e79ab441c4f']]; loss = 0.148927
+Epoch 0: |                                                                                                                                                        | 840/? [53:32<00:00,  0.26it/s, v_num=pmyy]context = [[1, 18, 22, 27, 29, 35, 38, 49], [3, 12, 30, 36, 43, 47, 54, 55], [17, 36, 55, 65, 72, 77, 86, 88]]target = [[5, 36, 48, 11, 2, 39, 35, 37], [6, 54, 43, 18, 30, 15, 10, 11], [52, 59, 56, 77, 25, 68, 22, 27]]
+Epoch 0: |                                                                                                                                                        | 849/? [54:07<00:00,  0.26it/s, v_num=pmyy]train step 850; scene = [['818df63ddd1cf294'], ['3002f9cbe7f00e6c'], ['8ec42c5dfea6823b'], ['64ae75e57c6aa0a4']]; loss = 0.112991
+Epoch 0: |                                                                                                                                                        | 850/? [54:11<00:00,  0.26it/s, v_num=pmyy]context = [[31, 75, 86], [20, 22, 73], [7, 55, 63], [3, 62, 68], [10, 43, 54], [52, 92, 111], [0, 34, 42], [51, 90, 102]]target = [[40, 57, 36], [39, 67, 34], [12, 9, 52], [29, 12, 39], [49, 13, 17], [106, 93, 85], [39, 20, 9], [79, 60, 53]]
+Epoch 0: |                                                                                                                                                        | 859/? [54:45<00:00,  0.26it/s, v_num=pmyy]train step 860; scene = [['eb0aa1a4fb58c50c']]; loss = 0.037343
+Epoch 0: |                                                                                                                                                        | 860/? [54:49<00:00,  0.26it/s, v_num=pmyy]context = [[147, 155, 165, 189, 194, 195], [9, 15, 44, 47, 49, 82], [5, 22, 23, 51, 61, 64], [97, 98, 140, 149, 160, 170]]target = [[194, 192, 155, 151, 174, 158], [22, 27, 54, 18, 81, 31], [25, 27, 29, 23, 6, 59], [137, 162, 98, 111, 152, 164]]
+Epoch 0: |                                                                                                                                                        | 869/? [55:23<00:00,  0.26it/s, v_num=pmyy]train step 870; scene = [['c357fbd8aca05570']]; loss = 0.044715
+Epoch 0: |                                                                                                                                                        | 870/? [55:27<00:00,  0.26it/s, v_num=pmyy]context = [[44, 45, 82, 100], [70, 93, 96, 147], [144, 176, 184, 205], [0, 51, 65, 67], [105, 106, 130, 152], [5, 6, 55, 60]]target = [[81, 66, 67, 97], [116, 118, 137, 91], [200, 193, 197, 176], [11, 13, 22, 32], [120, 117, 127, 130], [17, 14, 58, 27]]
+Epoch 0: |                                                                                                                                                        | 879/? [56:02<00:00,  0.26it/s, v_num=pmyy]train step 880; scene = [['c672fa3960b73528'], ['a60e4127f167ac93'], ['a9739ec3a34012af']]; loss = 0.050499
+Epoch 0: |                                                                                                                                                        | 880/? [56:06<00:00,  0.26it/s, v_num=pmyy]context = [[15, 85, 91], [0, 68, 74], [73, 85, 121], [4, 53, 80], [15, 36, 63], [42, 81, 92], [0, 45, 49], [65, 97, 132]]target = [[49, 21, 35], [47, 44, 69], [119, 89, 106], [57, 76, 6], [28, 32, 55], [62, 43, 49], [18, 44, 24], [90, 104, 71]]
+Epoch 0: |                                                                                                                                                        | 889/? [56:39<00:00,  0.26it/s, v_num=pmyy]train step 890; scene = [['40a3f4f9389dd20c'], ['b14ec6f019932d8d'], ['4b9ed7532c875dab'], ['10c36bd5ef5f5a6b'], ['bc9a64096787007d'], ['d58a26d24f2776b2'], ['46f2228076e6f3f7'], ['399668567ff33ad7']]; loss = 0.065888
+Epoch 0: |                                                                                                                                                        | 890/? [56:42<00:00,  0.26it/s, v_num=pmyy]context = [[32, 34, 35, 37, 39, 44, 52, 59, 73, 84, 87, 93], [20, 21, 22, 34, 43, 47, 52, 57, 64, 69, 72, 96]]target = [[87, 80, 41, 73, 35, 37, 86, 66, 72, 39, 50, 83], [28, 22, 50, 33, 26, 79, 75, 73, 29, 77, 60, 37]]
+Epoch 0: |                                                                                                                                                        | 899/? [57:17<00:00,  0.26it/s, v_num=pmyy]train step 900; scene = [['711ade236bebd618']]; loss = 0.055804
+Epoch 0: |                                                                                                                                                        | 900/? [57:21<00:00,  0.26it/s, v_num=pmyy]context = [[6, 7, 14, 15, 17, 19, 39, 43, 44, 45, 47, 48, 49, 53, 57, 61, 67, 69, 72, 73, 78, 90, 92, 103]]target = [[18, 19, 38, 66, 14, 96, 50, 55, 41, 57, 45, 83, 12, 100, 24, 53, 85, 40, 54, 58, 20, 44, 86, 95]]
+Epoch 0: |                                                                                                                                                        | 909/? [57:55<00:00,  0.26it/s, v_num=pmyy]train step 910; scene = [['6b495ce9634d2fbb']]; loss = 0.075657
+Epoch 0: |                                                                                                                                                        | 910/? [57:59<00:00,  0.26it/s, v_num=pmyy]context = [[59, 61, 62, 63, 73, 74, 77, 90, 100, 105, 111, 112], [8, 15, 18, 21, 26, 35, 37, 40, 41, 56, 68, 70]]target = [[76, 98, 64, 84, 69, 87, 100, 99, 67, 108, 77, 78], [22, 24, 41, 15, 45, 44, 34, 28, 13, 39, 11, 55]]
+Epoch 0: |                                                                                                                                                        | 919/? [58:33<00:00,  0.26it/s, v_num=pmyy]train step 920; scene = [['013264a550df794f'], ['4203a06d618eeb97']]; loss = 0.078105
+Epoch 0: |                                                                                                                                                        | 920/? [58:37<00:00,  0.26it/s, v_num=pmyy]context = [[14, 16, 18, 28, 30, 37, 45, 49, 50, 61, 63, 65], [97, 98, 114, 115, 121, 134, 137, 140, 143, 144, 169, 179]]target = [[38, 62, 50, 58, 42, 55, 57, 46, 43, 49, 56, 40], [158, 178, 130, 108, 145, 120, 168, 164, 156, 177, 125, 111]]
+Epoch 0: |                                                                                                                                                        | 929/? [59:11<00:00,  0.26it/s, v_num=pmyy]train step 930; scene = [['5747f1d12ad10026'], ['48c9bb29482ddf76'], ['a45bddb856f554a1']]; loss = 0.038018
+Epoch 0: |                                                                                                                                                        | 930/? [59:14<00:00,  0.26it/s, v_num=pmyy]context = [[0, 2, 12, 23, 26, 31, 35, 37, 41, 42, 44, 49], [164, 174, 179, 182, 184, 187, 189, 201, 202, 205, 207, 216]]target = [[33, 7, 10, 26, 44, 36, 39, 30, 43, 11, 38, 27], [214, 173, 176, 191, 182, 190, 175, 210, 215, 167, 195, 198]]
+Epoch 0: |                                                                                                                                                        | 939/? [59:49<00:00,  0.26it/s, v_num=pmyy]train step 940; scene = [['db02cd4ba6a027da'], ['8f5e074629cedd06']]; loss = 0.036899
+Epoch 0: |                                                                                                                                                        | 940/? [59:52<00:00,  0.26it/s, v_num=pmyy]context = [[127, 141, 154, 157, 160, 163, 172, 173, 189, 190, 192, 196], [6, 23, 36, 41, 46, 51, 57, 61, 65, 66, 83, 84]]target = [[157, 179, 153, 148, 161, 137, 194, 177, 128, 162, 166, 174], [38, 41, 32, 28, 8, 37, 36, 9, 77, 18, 80, 26]]
+Epoch 0: |                                                                                                                                                      | 949/? [1:00:27<00:00,  0.26it/s, v_num=pmyy]train step 950; scene = [['5b98e84e8e7ffef0'], ['d7ca47da5fac7140'], ['eff98653337775a8'], ['ba00608cd351deb0']]; loss = 0.036079
+Epoch 0: |                                                                                                                                                      | 950/? [1:00:31<00:00,  0.26it/s, v_num=pmyy]context = [[28, 33, 43, 46, 47, 51, 54, 67, 68, 72, 74, 75, 88, 90, 94, 98, 101, 102, 103, 105, 107, 113, 118, 125]]target = [[56, 59, 68, 66, 119, 65, 80, 100, 101, 39, 97, 94, 110, 62, 109, 40, 91, 42, 44, 78, 60, 73, 108, 50]]
+Epoch 0: |                                                                                                                                                      | 959/? [1:01:05<00:00,  0.26it/s, v_num=pmyy]train step 960; scene = [['f6d65c637ff68de3'], ['9b4d466924c40d8b'], ['b6c9aa729ebc703e']]; loss = 0.081706
+Epoch 0: |                                                                                                                                                      | 960/? [1:01:09<00:00,  0.26it/s, v_num=pmyy]context = [[2, 10, 11, 14, 19, 21, 23, 33, 39, 45, 49, 52, 53, 55, 58, 61, 71, 74, 80, 83, 84, 90, 94, 99]]target = [[41, 39, 14, 8, 52, 95, 19, 76, 24, 68, 75, 69, 22, 3, 47, 98, 17, 38, 89, 35, 21, 57, 45, 43]]
+Epoch 0: |                                                                                                                                                      | 969/? [1:01:44<00:00,  0.26it/s, v_num=pmyy]train step 970; scene = [['4b6cbf1f4c87d918']]; loss = 0.040110
+Epoch 0: |                                                                                                                                                      | 970/? [1:01:48<00:00,  0.26it/s, v_num=pmyy]context = [[20, 23, 30, 51, 53, 56, 60, 70], [145, 151, 184, 191, 192, 195, 215, 219], [36, 61, 63, 93, 106, 109, 111, 114]]target = [[52, 61, 27, 43, 44, 68, 67, 65], [210, 148, 165, 147, 166, 153, 170, 176], [108, 92, 76, 81, 71, 104, 41, 54]]
+Epoch 0: |                                                                                                                                                      | 979/? [1:02:22<00:00,  0.26it/s, v_num=pmyy]train step 980; scene = [['0492d6125268e9ae']]; loss = 0.040824
+Epoch 0: |                                                                                                                                                      | 980/? [1:02:26<00:00,  0.26it/s, v_num=pmyy]context = [[2, 4, 66, 70], [30, 96, 104, 105], [2, 13, 43, 65], [6, 42, 79, 91], [48, 85, 95, 96], [43, 62, 80, 101]]target = [[16, 66, 47, 53], [66, 42, 33, 84], [51, 25, 31, 53], [39, 90, 74, 16], [89, 80, 56, 68], [49, 45, 72, 48]]
+Epoch 0: |                                                                                                                                                      | 989/? [1:03:00<00:00,  0.26it/s, v_num=pmyy]train step 990; scene = [['b0413d361b4e8abb']]; loss = 0.030760
+Epoch 0: |                                                                                                                                                      | 990/? [1:03:04<00:00,  0.26it/s, v_num=pmyy]context = [[14, 19, 24, 29, 30, 38, 39, 40, 53, 57, 58, 65], [0, 1, 21, 33, 35, 38, 39, 43, 49, 58, 59, 64]]target = [[38, 37, 40, 27, 39, 45, 15, 52, 55, 22, 23, 31], [12, 42, 48, 30, 10, 58, 52, 59, 8, 46, 19, 44]]
+Epoch 0: |                                                                                                                                                      | 999/? [1:03:38<00:00,  0.26it/s, v_num=pmyy]train step 1000; scene = [['2bf91de5cd028c93'], ['2abe932fd9d76528'], ['8fabc39ad677dece'], ['188398a54205f797']]; loss = 0.056078
+Epoch 0: |                                                                                                                                                     | 1000/? [1:03:42<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1000; scene = ['647f2049bf4cb3f3'];
+target intrinsic:  tensor(0.8998, device='cuda:0') tensor(0.9001, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8769, device='cuda:0') tensor(0.8779, device='cuda:0')
+[2026-02-24 23:31:34,760][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1000/? [1:03:43<00:00,  0.26it/s, v_num=pmyy]context = [[176, 185, 186, 188, 208, 224, 228, 246], [0, 6, 9, 11, 37, 43, 46, 54], [0, 16, 23, 40, 41, 44, 55, 68]]target = [[205, 177, 207, 223, 188, 184, 231, 179], [7, 27, 9, 11, 39, 15, 12, 28], [38, 2, 54, 3, 25, 7, 37, 46]]
+[2026-02-24 23:31:38,465][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1009/? [1:04:19<00:00,  0.26it/s, v_num=pmyy]train step 1010; scene = [['3b42fa1245f6b00b'], ['c472581eb9a351cb'], ['222255ddcacd20cf'], ['b953af75bafccce8'], ['2c862f06019afc7e'], ['4178cb9ef3e0fd6d'], ['7f9480301fa3e38b'], ['8ccc9f1ead9cad5b']]; loss = 0.053823
+Epoch 0: |                                                                                                                                                     | 1010/? [1:04:22<00:00,  0.26it/s, v_num=pmyy]context = [[0, 73], [12, 74], [41, 117], [34, 98], [3, 58], [117, 195], [117, 198], [33, 101], [0, 71], [45, 102], [36, 110], [82, 153]]target = [[44, 27], [42, 20], [116, 46], [88, 73], [5, 49], [190, 129], [127, 171], [75, 70], [61, 20], [97, 100], [51, 43], [151, 139]]
+Epoch 0: |                                                                                                                                                     | 1019/? [1:04:56<00:00,  0.26it/s, v_num=pmyy]train step 1020; scene = [['3a95efa051965ee0']]; loss = 0.071461
+Epoch 0: |                                                                                                                                                     | 1020/? [1:05:00<00:00,  0.26it/s, v_num=pmyy]context = [[8, 11, 13, 14, 15, 20, 21, 31, 34, 43, 46, 50, 55, 61, 71, 75, 76, 77, 81, 83, 85, 87, 93, 105]]target = [[83, 40, 76, 65, 29, 87, 84, 91, 31, 73, 50, 22, 38, 57, 99, 20, 19, 24, 26, 82, 36, 34, 44, 15]]
+Epoch 0: |                                                                                                                                                     | 1029/? [1:05:34<00:00,  0.26it/s, v_num=pmyy]train step 1030; scene = [['6bd6d8270eee0b81'], ['2ac1cf1adda42447'], ['a017de2545c81e26'], ['4d5593ce15e317ed'], ['437188c182ae6a10'], ['86440d86f60644b7'], ['27508406fddac6dd'], ['7feae1e0e2c1701b'], ['a86dd7257f21bc73'], ['a025554e94654c2f'], ['20ee785f1bc6340d'], ['db6cd90de8fee2ff']]; loss = 0.099014
+Epoch 0: |                                                                                                                                                     | 1030/? [1:05:38<00:00,  0.26it/s, v_num=pmyy]context = [[84, 89, 90, 91, 93, 111, 112, 113, 116, 117, 121, 126, 128, 134, 138, 139, 145, 153, 159, 161, 169, 173, 178, 181]]target = [[147, 138, 152, 92, 157, 97, 115, 154, 98, 113, 142, 137, 177, 165, 158, 146, 120, 139, 136, 122, 104, 85, 128, 144]]
+Epoch 0: |                                                                                                                                                     | 1039/? [1:06:12<00:00,  0.26it/s, v_num=pmyy]train step 1040; scene = [['80fd51689742e978'], ['5eb07ee6a9e4fec2']]; loss = 0.049193
+Epoch 0: |                                                                                                                                                     | 1040/? [1:06:16<00:00,  0.26it/s, v_num=pmyy]context = [[98, 99, 101, 110, 114, 116, 123, 126, 131, 137, 139, 150, 153, 155, 156, 158, 165, 177, 178, 184, 185, 187, 192, 195]]target = [[134, 141, 100, 117, 190, 107, 162, 158, 179, 192, 121, 151, 157, 139, 118, 175, 156, 183, 127, 167, 103, 172, 154, 116]]
+Epoch 0: |                                                                                                                                                     | 1049/? [1:06:50<00:00,  0.26it/s, v_num=pmyy]train step 1050; scene = [['77dd1e7595b8f9a1']]; loss = 0.055747
+Epoch 0: |                                                                                                                                                     | 1050/? [1:06:53<00:00,  0.26it/s, v_num=pmyy]context = [[114, 115, 118, 119, 124, 126, 130, 147, 149, 151, 158, 166, 169, 170, 173, 174, 177, 180, 191, 192, 203, 206, 208, 211]]target = [[209, 128, 160, 132, 164, 207, 201, 145, 117, 163, 120, 202, 151, 119, 124, 162, 189, 133, 184, 198, 153, 116, 188, 193]]
+Epoch 0: |                                                                                                                                                     | 1059/? [1:07:28<00:00,  0.26it/s, v_num=pmyy]train step 1060; scene = [['51b6e63c903275b4'], ['2a9baa599c00613f'], ['1d7dc2d489714b09'], ['9f92d7e79d4a35bb']]; loss = 0.042210
+Epoch 0: |                                                                                                                                                     | 1060/? [1:07:32<00:00,  0.26it/s, v_num=pmyy]context = [[15, 17, 24, 33, 36, 40, 69, 74, 77, 82, 83, 84], [1, 5, 9, 16, 25, 29, 32, 33, 44, 47, 49, 51]]target = [[41, 42, 50, 25, 44, 31, 43, 24, 81, 18, 26, 71], [45, 3, 37, 11, 42, 29, 49, 41, 17, 22, 39, 23]]
+Epoch 0: |                                                                                                                                                     | 1069/? [1:08:06<00:00,  0.26it/s, v_num=pmyy]train step 1070; scene = [['a464ce7c8383dbbb']]; loss = 0.080393
+Epoch 0: |                                                                                                                                                     | 1070/? [1:08:10<00:00,  0.26it/s, v_num=pmyy]context = [[6, 10, 28, 32, 57, 71, 89, 92], [34, 52, 65, 75, 88, 89, 114, 117], [119, 123, 131, 133, 137, 160, 180, 190]]target = [[90, 59, 62, 24, 52, 29, 47, 35], [72, 39, 106, 85, 58, 74, 77, 40], [188, 149, 156, 166, 161, 120, 125, 158]]
+Epoch 0: |                                                                                                                                                     | 1079/? [1:08:45<00:00,  0.26it/s, v_num=pmyy]train step 1080; scene = [['b382af5f342061fa'], ['f6a0556897b15d6b']]; loss = 0.036790
+Epoch 0: |                                                                                                                                                     | 1080/? [1:08:48<00:00,  0.26it/s, v_num=pmyy]context = [[127, 150, 162, 189], [25, 41, 98, 102], [6, 17, 20, 68], [24, 48, 78, 100], [154, 155, 175, 235], [41, 55, 86, 109]]target = [[180, 163, 148, 162], [45, 57, 100, 82], [16, 24, 44, 47], [50, 42, 47, 92], [206, 210, 196, 211], [80, 89, 82, 62]]
+Epoch 0: |                                                                                                                                                     | 1089/? [1:09:22<00:00,  0.26it/s, v_num=pmyy]train step 1090; scene = [['5beb85aaf29d1242']]; loss = 0.060822
+Epoch 0: |                                                                                                                                                     | 1090/? [1:09:26<00:00,  0.26it/s, v_num=pmyy]context = [[27, 37, 53, 65, 72, 76, 79, 83, 86, 97, 104, 109], [10, 17, 18, 32, 38, 45, 46, 52, 67, 68, 72, 79]]target = [[103, 63, 65, 74, 40, 85, 44, 107, 57, 58, 53, 29], [18, 61, 49, 74, 69, 21, 73, 59, 30, 67, 11, 43]]
+Epoch 0: |                                                                                                                                                     | 1099/? [1:10:00<00:00,  0.26it/s, v_num=pmyy]train step 1100; scene = [['db811a2460c4f9b5'], ['7db8c4965bba509a'], ['a7a79393e5bb8108'], ['2d77b1dd90856337']]; loss = 0.070515
+Epoch 0: |                                                                                                                                                     | 1100/? [1:10:04<00:00,  0.26it/s, v_num=pmyy]context = [[1, 10, 15, 20, 27, 35, 38, 44, 51, 52, 53, 70], [40, 44, 45, 46, 52, 54, 56, 63, 88, 95, 96, 98]]target = [[18, 63, 10, 69, 55, 26, 40, 21, 59, 51, 42, 48], [84, 92, 51, 54, 77, 50, 59, 47, 86, 43, 89, 87]]
+Epoch 0: |                                                                                                                                                     | 1109/? [1:10:38<00:00,  0.26it/s, v_num=pmyy]train step 1110; scene = [['a815d5a5f2ec9562'], ['3be8c5ae6c95c9b4']]; loss = 0.036679
+Epoch 0: |                                                                                                                                                     | 1110/? [1:10:42<00:00,  0.26it/s, v_num=pmyy]context = [[0, 2, 11, 15, 25, 28, 29, 32, 33, 35, 40, 44, 47, 49, 55, 56, 66, 72, 74, 84, 86, 88, 96, 97]]target = [[15, 85, 74, 83, 43, 47, 21, 40, 39, 66, 3, 26, 29, 9, 38, 23, 94, 36, 62, 19, 12, 44, 76, 31]]
+Epoch 0: |                                                                                                                                                     | 1119/? [1:11:16<00:00,  0.26it/s, v_num=pmyy]train step 1120; scene = [['3ab1ce6779776017']]; loss = 0.106461
+Epoch 0: |                                                                                                                                                     | 1120/? [1:11:20<00:00,  0.26it/s, v_num=pmyy]context = [[9, 16, 17, 21, 23, 30, 31, 43, 55, 57, 59, 61], [11, 25, 51, 66, 72, 76, 82, 87, 91, 94, 97, 99]]target = [[56, 22, 30, 35, 59, 36, 38, 23, 34, 11, 26, 51], [19, 73, 37, 71, 50, 43, 17, 80, 25, 54, 28, 24]]
+Epoch 0: |                                                                                                                                                     | 1129/? [1:11:54<00:00,  0.26it/s, v_num=pmyy]train step 1130; scene = [['fd0ebd5afbfd1acf'], ['edf6636dfd51ba3c'], ['7b2c118f021e6902'], ['b3356e816a130b87'], ['78cbac3ff58f2e41'], ['94c654bd3e031bcb'], ['b281bf93286a0573'], ['d9dce3382830aea6'], ['070a524bacb9aa38'], ['cae0139a521aa052'], ['5645a008715acf0a'], ['23174a6cd65a0731']]; loss = 0.087992
+Epoch 0: |                                                                                                                                                     | 1130/? [1:11:57<00:00,  0.26it/s, v_num=pmyy]context = [[8, 50, 81, 84], [2, 42, 44, 54], [25, 26, 79, 87], [25, 30, 88, 92], [45, 100, 129, 135], [17, 68, 77, 83]]target = [[62, 80, 76, 55], [11, 29, 39, 13], [42, 27, 75, 58], [31, 63, 29, 48], [72, 58, 93, 106], [66, 77, 60, 59]]
+Epoch 0: |                                                                                                                                                     | 1139/? [1:12:31<00:00,  0.26it/s, v_num=pmyy]train step 1140; scene = [['090a038e9d844b4e'], ['d28a2455cc34badb']]; loss = 0.034319
+Epoch 0: |                                                                                                                                                     | 1140/? [1:12:35<00:00,  0.26it/s, v_num=pmyy]context = [[19, 20, 28, 32, 34, 36, 38, 40, 52, 54, 63, 65, 68, 77, 78, 87, 94, 98, 100, 105, 107, 113, 114, 116]]target = [[32, 23, 97, 28, 51, 21, 114, 76, 46, 84, 74, 31, 98, 108, 77, 82, 72, 49, 103, 47, 80, 39, 59, 70]]
+Epoch 0: |                                                                                                                                                     | 1149/? [1:13:09<00:00,  0.26it/s, v_num=pmyy]train step 1150; scene = [['82347d56a4a55c27'], ['ee03755cce11b682']]; loss = 0.067113
+Epoch 0: |                                                                                                                                                     | 1150/? [1:13:13<00:00,  0.26it/s, v_num=pmyy]context = [[13, 16, 50, 52, 55, 62], [55, 56, 79, 89, 111, 117], [35, 43, 73, 74, 82, 85], [30, 56, 58, 80, 102, 108]]target = [[39, 59, 30, 34, 42, 27], [91, 98, 71, 70, 84, 68], [42, 44, 39, 68, 69, 50], [42, 68, 72, 55, 87, 51]]
+Epoch 0: |                                                                                                                                                     | 1159/? [1:13:48<00:00,  0.26it/s, v_num=pmyy]train step 1160; scene = [['255998558abc7172']]; loss = 0.076604
+Epoch 0: |                                                                                                                                                     | 1160/? [1:13:52<00:00,  0.26it/s, v_num=pmyy]context = [[13, 16, 18, 19, 21, 25, 28, 29, 36, 38, 43, 47, 49, 50, 53, 56, 67, 68, 69, 71, 72, 93, 108, 110]]target = [[47, 34, 77, 80, 28, 94, 97, 19, 29, 26, 68, 40, 53, 20, 60, 30, 92, 70, 83, 48, 106, 15, 63, 41]]
+Epoch 0: |                                                                                                                                                     | 1169/? [1:14:26<00:00,  0.26it/s, v_num=pmyy]train step 1170; scene = [['771aa992eae9a574'], ['4f9716bb3dc7feec'], ['fbed2318ae410b31'], ['8e1b4054949b6a46']]; loss = 0.046667
+Epoch 0: |                                                                                                                                                     | 1170/? [1:14:30<00:00,  0.26it/s, v_num=pmyy]context = [[207, 214, 223, 228, 242, 250, 268, 269], [35, 58, 71, 96, 97, 109, 110, 114], [0, 27, 28, 46, 50, 60, 68, 75]]target = [[216, 210, 265, 253, 237, 223, 241, 234], [105, 106, 100, 96, 73, 51, 109, 93], [2, 54, 60, 70, 28, 22, 51, 62]]
+Epoch 0: |                                                                                                                                                     | 1179/? [1:15:05<00:00,  0.26it/s, v_num=pmyy]train step 1180; scene = [['18c880b4b5ef683e']]; loss = 0.072818
+Epoch 0: |                                                                                                                                                     | 1180/? [1:15:08<00:00,  0.26it/s, v_num=pmyy]context = [[48, 60, 70, 80, 90, 109, 110, 114], [3, 8, 11, 23, 28, 34, 62, 82], [14, 17, 19, 24, 30, 34, 59, 67]]target = [[85, 101, 84, 113, 106, 74, 87, 98], [81, 68, 48, 46, 67, 69, 56, 22], [31, 33, 45, 20, 17, 61, 37, 44]]
+Epoch 0: |                                                                                                                                                     | 1189/? [1:15:42<00:00,  0.26it/s, v_num=pmyy]train step 1190; scene = [['019fdd708d7163bd'], ['7046980b0d3d3c63'], ['f1af5d4039ce3a2c']]; loss = 0.043256
+Epoch 0: |                                                                                                                                                     | 1190/? [1:15:46<00:00,  0.26it/s, v_num=pmyy]context = [[0, 7, 8, 11, 15, 20, 21, 28, 30, 33, 38, 42, 47, 54, 59, 63, 66, 71, 73, 79, 81, 88, 90, 97]]target = [[83, 9, 22, 57, 21, 76, 37, 10, 7, 13, 43, 47, 19, 6, 51, 71, 95, 78, 24, 92, 30, 28, 55, 84]]
+Epoch 0: |                                                                                                                                                     | 1199/? [1:16:21<00:00,  0.26it/s, v_num=pmyy]train step 1200; scene = [['5070ea042c65de0d'], ['19f950b6900d6176'], ['02768fad99be3290'], ['ed6c9a5913622c3d'], ['0a4cbe699be68e5c'], ['0892e76375b283ba'], ['efd001b0d5127d61'], ['9adb745c741c85e0'], ['e4774e728791dc20'], ['23710dc8de8b4a49'], ['347d7c1f3516f732'], ['97283dc038203c65']]; loss = 0.096859
+Epoch 0: |                                                                                                                                                     | 1200/? [1:16:25<00:00,  0.26it/s, v_num=pmyy]context = [[44, 48, 49, 54, 63, 66, 67, 70, 76, 77, 78, 84, 90, 100, 101, 105, 106, 111, 113, 116, 120, 123, 125, 141]]target = [[63, 111, 132, 61, 108, 104, 110, 81, 59, 89, 131, 126, 73, 100, 134, 74, 121, 84, 80, 51, 140, 133, 55, 83]]
+[2026-02-24 23:44:21,241][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1209/? [1:17:02<00:00,  0.26it/s, v_num=pmyy]train step 1210; scene = [['7ba6259f378c70f8']]; loss = 0.043018
+Epoch 0: |                                                                                                                                                     | 1210/? [1:17:06<00:00,  0.26it/s, v_num=pmyy]context = [[77, 82, 86, 95, 101, 107, 111, 112, 113, 117, 120, 130], [9, 29, 33, 44, 49, 53, 54, 58, 67, 68, 81, 96]]target = [[99, 107, 90, 119, 117, 113, 92, 120, 121, 123, 86, 94], [32, 87, 79, 70, 51, 71, 33, 67, 53, 82, 41, 56]]
+Epoch 0: |                                                                                                                                                     | 1219/? [1:17:40<00:00,  0.26it/s, v_num=pmyy]train step 1220; scene = [['0049c83bad21bbdf'], ['a73f4bf6ce8b2a00'], ['f7bef32e09ab061e'], ['175e1611df44964f']]; loss = 0.035434
+Epoch 0: |                                                                                                                                                     | 1220/? [1:17:44<00:00,  0.26it/s, v_num=pmyy]context = [[44, 47, 55, 57, 65, 80, 85, 87, 88, 94, 112, 116], [25, 37, 49, 59, 61, 70, 82, 84, 86, 94, 97, 100]]target = [[96, 81, 62, 98, 79, 76, 52, 95, 59, 65, 56, 72], [34, 26, 32, 50, 44, 97, 67, 53, 75, 66, 59, 40]]
+Epoch 0: |                                                                                                                                                     | 1229/? [1:18:18<00:00,  0.26it/s, v_num=pmyy]train step 1230; scene = [['342ff7b7111b53c1'], ['9ebe2434b1d68246'], ['b84cce034b55e1e4']]; loss = 0.042477
+Epoch 0: |                                                                                                                                                     | 1230/? [1:18:22<00:00,  0.26it/s, v_num=pmyy]context = [[5, 8, 11, 16, 19, 23, 25, 34, 35, 36, 43, 49, 51, 57, 60, 68, 77, 78, 86, 88, 97, 98, 101, 102]]target = [[84, 48, 64, 18, 41, 19, 60, 25, 52, 49, 51, 50, 75, 38, 42, 81, 92, 32, 69, 67, 26, 100, 65, 61]]
+Epoch 0: |                                                                                                                                                     | 1239/? [1:18:57<00:00,  0.26it/s, v_num=pmyy]train step 1240; scene = [['d7ac888a45c3c904'], ['60d988ddbc2d04f1'], ['617340307747e227'], ['969c91a2507e2d81'], ['af197296b340b564'], ['d488b5f08aadff0c']]; loss = 0.068918
+Epoch 0: |                                                                                                                                                     | 1240/? [1:19:01<00:00,  0.26it/s, v_num=pmyy]context = [[9, 16, 18, 25, 28, 30, 41, 43, 46, 51, 53, 55, 56, 57, 59, 68, 71, 72, 74, 77, 87, 95, 105, 106]]target = [[72, 53, 26, 17, 11, 65, 42, 97, 60, 75, 93, 10, 45, 15, 102, 38, 101, 54, 23, 98, 13, 68, 16, 46]]
+Epoch 0: |                                                                                                                                                     | 1249/? [1:19:34<00:00,  0.26it/s, v_num=pmyy]train step 1250; scene = [['d51d569c27d0b2b5'], ['e33b2a9076f25c4d'], ['2f330103819454c6'], ['944e92ff3fea78eb'], ['5a6387d05cd51e02'], ['2f269b68e14e256a'], ['55f1b82ae9c5571f'], ['1490c145692a1899'], ['b1b24b049d5a5da4'], ['834e851000651b8f'], ['4bc6b34a301aac73'], ['97caecafbfa7f1f6']]; loss = 0.096965
+Epoch 0: |                                                                                                                                                     | 1250/? [1:19:38<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1250; scene = ['70b0a33083333dc9'];
+target intrinsic:  tensor(0.8872, device='cuda:0') tensor(0.8874, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8930, device='cuda:0') tensor(0.8948, device='cuda:0')
+[2026-02-24 23:47:31,382][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1250/? [1:19:39<00:00,  0.26it/s, v_num=pmyy]context = [[0, 2, 3, 7, 9, 14, 16, 19, 27, 30, 32, 34, 39, 43, 56, 62, 65, 73, 75, 80, 82, 85, 91, 97]]target = [[7, 68, 16, 60, 94, 67, 34, 41, 53, 11, 1, 12, 30, 6, 61, 76, 22, 32, 25, 81, 88, 33, 9, 45]]
+Epoch 0: |                                                                                                                                                     | 1259/? [1:20:13<00:00,  0.26it/s, v_num=pmyy]train step 1260; scene = [['1a1f0a618bad5d30'], ['3046f14dad8f19f2'], ['8b4aecd7318912e3']]; loss = 0.042916
+Epoch 0: |                                                                                                                                                     | 1260/? [1:20:17<00:00,  0.26it/s, v_num=pmyy]context = [[44, 48, 50, 51, 55, 60, 62, 66, 72, 78, 79, 85, 91, 98, 99, 100, 102, 112, 117, 121, 123, 125, 129, 141]]target = [[73, 109, 111, 54, 63, 69, 65, 124, 82, 119, 125, 110, 117, 56, 123, 86, 99, 126, 134, 53, 84, 112, 74, 100]]
+Epoch 0: |                                                                                                                                                     | 1269/? [1:20:52<00:00,  0.26it/s, v_num=pmyy]train step 1270; scene = [['a0438ace9b7619cf'], ['336b3140d9d8bebd']]; loss = 0.060107
+Epoch 0: |                                                                                                                                                     | 1270/? [1:20:56<00:00,  0.26it/s, v_num=pmyy]context = [[20, 22, 25, 26, 40, 44, 45, 51, 55, 63, 67, 70, 72, 76, 80, 81, 82, 84, 88, 96, 104, 109, 110, 117]]target = [[84, 115, 116, 28, 110, 27, 111, 85, 100, 39, 61, 63, 68, 56, 101, 113, 108, 94, 73, 62, 67, 98, 47, 55]]
+Epoch 0: |                                                                                                                                                     | 1279/? [1:21:29<00:00,  0.26it/s, v_num=pmyy]train step 1280; scene = [['118f75d62c9a1f46'], ['3bcbf64a56113736'], ['6ca39802dcef328e'], ['0d754618c77e44f6']]; loss = 0.055051
+Epoch 0: |                                                                                                                                                     | 1280/? [1:21:33<00:00,  0.26it/s, v_num=pmyy]context = [[5, 19, 21, 22, 26, 37, 47, 50, 53, 75, 85, 87], [4, 19, 23, 26, 29, 36, 37, 63, 69, 72, 75, 80]]target = [[65, 33, 14, 32, 52, 25, 54, 34, 15, 6, 60, 30], [11, 5, 9, 35, 70, 15, 51, 38, 20, 42, 68, 22]]
+Epoch 0: |                                                                                                                                                     | 1289/? [1:22:07<00:00,  0.26it/s, v_num=pmyy]train step 1290; scene = [['4738c57794aa47f8']]; loss = 0.051468
+Epoch 0: |                                                                                                                                                     | 1290/? [1:22:10<00:00,  0.26it/s, v_num=pmyy]context = [[1, 4, 14, 17, 27, 30, 32, 37, 44, 52, 64, 80], [0, 8, 12, 14, 24, 28, 29, 37, 49, 54, 56, 67]]target = [[64, 28, 44, 35, 23, 68, 42, 61, 3, 77, 52, 14], [1, 11, 56, 59, 13, 63, 55, 61, 38, 9, 18, 37]]
+Epoch 0: |                                                                                                                                                     | 1299/? [1:22:45<00:00,  0.26it/s, v_num=pmyy]train step 1300; scene = [['f18e524ab2d288c3'], ['9dff9a317da1d49d'], ['11a87d6d490e024d']]; loss = 0.039099
+Epoch 0: |                                                                                                                                                     | 1300/? [1:22:49<00:00,  0.26it/s, v_num=pmyy]context = [[5, 6, 8, 12, 30, 45, 58, 64, 65, 75, 79, 85], [22, 26, 27, 28, 39, 45, 47, 55, 56, 57, 66, 72]]target = [[34, 57, 32, 7, 11, 58, 73, 68, 83, 14, 61, 55], [42, 34, 36, 47, 33, 67, 31, 30, 50, 27, 24, 58]]
+Epoch 0: |                                                                                                                                                     | 1309/? [1:23:23<00:00,  0.26it/s, v_num=pmyy]train step 1310; scene = [['589e362118b32d25'], ['d13522abd38eddfe'], ['f1657c6128d2b332'], ['1cb3ecb30e3e9d0e'], ['827c5f3b3886553d'], ['4d27fb96530fe02b']]; loss = 0.096130
+Epoch 0: |                                                                                                                                                     | 1310/? [1:23:26<00:00,  0.26it/s, v_num=pmyy]context = [[20, 21, 27, 31, 36, 40, 48, 51, 55, 62, 69, 72, 74, 75, 83, 89, 96, 97, 99, 104, 105, 106, 114, 117]]target = [[57, 113, 99, 80, 42, 30, 24, 74, 34, 52, 63, 59, 100, 101, 76, 36, 105, 66, 55, 27, 84, 22, 79, 86]]
+Epoch 0: |                                                                                                                                                     | 1319/? [1:24:01<00:00,  0.26it/s, v_num=pmyy]train step 1320; scene = [['c21766bd51fed5bc'], ['47a2a8b326b9f40e'], ['d0c0d78936b6e5ce'], ['e0ee3878561a5fed']]; loss = 0.060979
+Epoch 0: |                                                                                                                                                     | 1320/? [1:24:04<00:00,  0.26it/s, v_num=pmyy]context = [[194, 242], [34, 99], [51, 136], [156, 207], [36, 100], [0, 48], [10, 99], [22, 79], [18, 99], [2, 74], [39, 128], [201, 272]]target = [[225, 201], [63, 36], [58, 72], [193, 181], [39, 77], [13, 1], [29, 68], [36, 58], [45, 66], [58, 18], [100, 50], [234, 262]]
+Epoch 0: |                                                                                                                                                     | 1329/? [1:24:39<00:00,  0.26it/s, v_num=pmyy]train step 1330; scene = [['07fdb102ee3677f5']]; loss = 0.037399
+Epoch 0: |                                                                                                                                                     | 1330/? [1:24:42<00:00,  0.26it/s, v_num=pmyy]context = [[20, 23, 28, 29, 37, 43, 49, 53, 56, 57, 61, 65, 69, 70, 80, 92, 98, 102, 103, 104, 106, 108, 116, 117]]target = [[37, 57, 73, 79, 30, 102, 42, 115, 86, 33, 28, 25, 114, 53, 76, 107, 36, 106, 52, 49, 59, 66, 61, 101]]
+Epoch 0: |                                                                                                                                                     | 1339/? [1:25:17<00:00,  0.26it/s, v_num=pmyy]train step 1340; scene = [['2ff4f3b2475e0e8c'], ['db062134f9dec5f1']]; loss = 0.034154
+Epoch 0: |                                                                                                                                                     | 1340/? [1:25:21<00:00,  0.26it/s, v_num=pmyy]context = [[20, 22, 50, 84], [0, 36, 48, 51], [10, 38, 52, 66], [0, 24, 25, 53], [0, 4, 19, 45], [54, 57, 73, 112]]target = [[42, 28, 78, 23], [3, 12, 46, 44], [57, 60, 15, 41], [47, 3, 31, 2], [42, 13, 23, 28], [102, 69, 98, 70]]
+Epoch 0: |                                                                                                                                                     | 1349/? [1:25:55<00:00,  0.26it/s, v_num=pmyy]train step 1350; scene = [['87e1164e050e9686'], ['3803b2c8fd539a88']]; loss = 0.057174
+Epoch 0: |                                                                                                                                                     | 1350/? [1:25:59<00:00,  0.26it/s, v_num=pmyy]context = [[52, 53, 55, 63, 64, 66, 74, 75, 77, 79, 85, 89, 90, 99, 121, 122, 123, 126, 133, 134, 135, 136, 147, 149]]target = [[99, 80, 109, 136, 92, 134, 112, 76, 135, 68, 116, 138, 54, 145, 144, 143, 61, 72, 102, 77, 90, 85, 139, 58]]
+Epoch 0: |                                                                                                                                                     | 1359/? [1:26:33<00:00,  0.26it/s, v_num=pmyy]train step 1360; scene = [['7db4ed902d003a63']]; loss = 0.067330
+Epoch 0: |                                                                                                                                                     | 1360/? [1:26:37<00:00,  0.26it/s, v_num=pmyy]context = [[13, 26, 58, 83], [117, 127, 165, 171], [12, 45, 50, 66], [21, 23, 67, 101], [26, 34, 95, 97], [42, 73, 81, 98]]target = [[65, 36, 43, 27], [124, 136, 141, 132], [33, 37, 62, 29], [82, 90, 80, 97], [45, 44, 32, 86], [55, 82, 58, 89]]
+Epoch 0: |                                                                                                                                                     | 1369/? [1:27:12<00:00,  0.26it/s, v_num=pmyy]train step 1370; scene = [['36df585860d0ad88']]; loss = 0.035190
+Epoch 0: |                                                                                                                                                     | 1370/? [1:27:16<00:00,  0.26it/s, v_num=pmyy]context = [[41, 42, 45, 55, 59, 62, 67, 68, 69, 73, 79, 87, 90, 94, 104, 114, 117, 120, 121, 125, 128, 130, 135, 138]]target = [[79, 103, 135, 133, 104, 59, 92, 71, 62, 66, 61, 109, 68, 124, 74, 54, 44, 46, 57, 96, 108, 58, 85, 83]]
+Epoch 0: |                                                                                                                                                     | 1379/? [1:27:49<00:00,  0.26it/s, v_num=pmyy]train step 1380; scene = [['256ae648672281d1'], ['7043e8afce176c8c']]; loss = 0.059948
+Epoch 0: |                                                                                                                                                     | 1380/? [1:27:53<00:00,  0.26it/s, v_num=pmyy]context = [[31, 33, 48, 63, 106, 115], [8, 28, 29, 44, 67, 97], [65, 72, 100, 110, 115, 124], [8, 9, 63, 70, 75, 79]]target = [[57, 76, 83, 113, 107, 38], [51, 64, 72, 39, 18, 77], [89, 108, 95, 114, 84, 116], [20, 59, 14, 56, 15, 38]]
+Epoch 0: |                                                                                                                                                     | 1389/? [1:28:27<00:00,  0.26it/s, v_num=pmyy]train step 1390; scene = [['ca65a604c0f00319'], ['937ea87bb5a2047f'], ['a2b430d7bec915d8'], ['c8c116c28ca108b0']]; loss = 0.035484
+Epoch 0: |                                                                                                                                                     | 1390/? [1:28:31<00:00,  0.26it/s, v_num=pmyy]context = [[0, 3, 5, 26, 31, 34, 38, 40, 44, 53, 62, 70], [1, 4, 21, 29, 36, 37, 39, 44, 52, 54, 59, 64]]target = [[36, 5, 37, 41, 54, 19, 44, 21, 40, 34, 59, 32], [59, 51, 53, 46, 27, 9, 21, 52, 24, 54, 32, 48]]
+Epoch 0: |                                                                                                                                                     | 1399/? [1:29:05<00:00,  0.26it/s, v_num=pmyy]train step 1400; scene = [['73fef5139753e974'], ['4b3e117d4f50b167'], ['0e916e63743f841b']]; loss = 0.060205
+Epoch 0: |                                                                                                                                                     | 1400/? [1:29:08<00:00,  0.26it/s, v_num=pmyy]context = [[23, 26, 27, 29, 31, 32, 41, 47, 49, 50, 55, 56, 58, 71, 77, 78, 87, 91, 97, 100, 101, 102, 113, 120]]target = [[55, 52, 103, 28, 68, 66, 49, 93, 76, 109, 105, 112, 100, 101, 31, 70, 64, 116, 77, 56, 46, 48, 32, 85]]
+[2026-02-24 23:57:04,915][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1409/? [1:29:47<00:00,  0.26it/s, v_num=pmyy]train step 1410; scene = [['d9ce9620cf31e776'], ['36bf8b485c1284dd']]; loss = 0.037828
+Epoch 0: |                                                                                                                                                     | 1410/? [1:29:51<00:00,  0.26it/s, v_num=pmyy]context = [[44, 47, 51, 55, 57, 59, 61, 62, 68, 72, 78, 80, 82, 88, 91, 97, 109, 112, 116, 118, 135, 138, 140, 141]]target = [[54, 91, 112, 61, 115, 80, 93, 45, 105, 51, 131, 102, 55, 60, 57, 136, 129, 99, 125, 113, 84, 89, 58, 65]]
+Epoch 0: |                                                                                                                                                     | 1419/? [1:30:25<00:00,  0.26it/s, v_num=pmyy]train step 1420; scene = [['df4e1717577f82c1'], ['50b38521b86dd7b6'], ['e4fc6fa4bbd6efa1'], ['54d3668b6c7ed4b8']]; loss = 0.045056
+Epoch 0: |                                                                                                                                                     | 1420/? [1:30:29<00:00,  0.26it/s, v_num=pmyy]context = [[38, 43, 53, 54, 68, 87, 97, 101], [52, 53, 74, 84, 86, 96, 132, 133], [157, 173, 184, 190, 200, 203, 210, 211]]target = [[77, 71, 57, 46, 97, 67, 93, 96], [83, 127, 75, 122, 56, 78, 130, 107], [177, 183, 188, 174, 191, 203, 200, 170]]
+Epoch 0: |                                                                                                                                                     | 1429/? [1:31:03<00:00,  0.26it/s, v_num=pmyy]train step 1430; scene = [['e70c21625fceaebd'], ['4cfdc3ed984caecf'], ['dfc00f3016d34131'], ['f5338d00f9c021a3'], ['89fbc4149a5d7348'], ['d2d3474ebd2be7e8'], ['0b42b8cf14bd1e13'], ['358a2d1387de3df8'], ['abe463f94db2c398'], ['d35b7c2c52c45a54'], ['12f629409a280733'], ['6b09a022387bd762']]; loss = 0.102157
+Epoch 0: |                                                                                                                                                     | 1430/? [1:31:07<00:00,  0.26it/s, v_num=pmyy]context = [[146, 159, 179, 187, 191, 200], [3, 11, 37, 39, 40, 48], [3, 16, 25, 27, 47, 49], [31, 36, 39, 84, 87, 94]]target = [[154, 176, 156, 190, 172, 150], [13, 36, 28, 46, 33, 32], [33, 7, 23, 32, 20, 35], [53, 82, 44, 85, 38, 54]]
+Epoch 0: |                                                                                                                                                     | 1439/? [1:31:42<00:00,  0.26it/s, v_num=pmyy]train step 1440; scene = [['83f69d126eb1528d'], ['e92e4346a6224492']]; loss = 0.037447
+Epoch 0: |                                                                                                                                                     | 1440/? [1:31:46<00:00,  0.26it/s, v_num=pmyy]context = [[1, 19, 33, 39, 51, 53, 57, 61], [4, 18, 19, 43, 44, 47, 49, 75], [40, 51, 69, 84, 88, 93, 99, 104]]target = [[3, 30, 34, 24, 46, 8, 13, 4], [19, 24, 28, 56, 55, 10, 66, 21], [54, 88, 79, 100, 81, 84, 83, 72]]
+Epoch 0: |                                                                                                                                                     | 1449/? [1:32:20<00:00,  0.26it/s, v_num=pmyy]train step 1450; scene = [['c05fd148d2da6d26'], ['b49cc4ec7c6b0050'], ['0d158225b3c47682']]; loss = 0.062648
+Epoch 0: |                                                                                                                                                     | 1450/? [1:32:24<00:00,  0.26it/s, v_num=pmyy]context = [[2, 15, 27, 38, 41, 50, 51, 57, 58, 61, 62, 73], [16, 17, 19, 26, 41, 44, 80, 89, 90, 92, 100, 101]]target = [[37, 21, 47, 67, 70, 31, 10, 57, 34, 30, 50, 52], [84, 77, 99, 81, 19, 59, 24, 78, 64, 25, 30, 90]]
+Epoch 0: |                                                                                                                                                     | 1459/? [1:32:58<00:00,  0.26it/s, v_num=pmyy]train step 1460; scene = [['8d583bfb265295b9']]; loss = 0.043909
+Epoch 0: |                                                                                                                                                     | 1460/? [1:33:01<00:00,  0.26it/s, v_num=pmyy]context = [[143, 146, 157, 160, 179, 202], [1, 13, 19, 36, 45, 46], [175, 189, 227, 237, 252, 257], [22, 29, 47, 51, 54, 67]]target = [[156, 146, 191, 160, 196, 187], [28, 40, 8, 23, 33, 44], [234, 248, 199, 178, 220, 226], [38, 63, 50, 35, 66, 33]]
+Epoch 0: |                                                                                                                                                     | 1469/? [1:33:36<00:00,  0.26it/s, v_num=pmyy]train step 1470; scene = [['6ebab888069161eb']]; loss = 0.039288
+Epoch 0: |                                                                                                                                                     | 1470/? [1:33:40<00:00,  0.26it/s, v_num=pmyy]context = [[28, 29, 32, 37, 43, 46, 50, 65, 67, 71, 76, 77], [132, 133, 136, 137, 140, 141, 145, 147, 169, 171, 200, 204]]target = [[31, 53, 62, 64, 65, 67, 34, 72, 39, 58, 57, 68], [174, 172, 164, 145, 162, 199, 202, 142, 163, 159, 176, 200]]
+Epoch 0: |                                                                                                                                                     | 1479/? [1:34:15<00:00,  0.26it/s, v_num=pmyy]train step 1480; scene = [['964d888d8d08f2aa'], ['58901334e2d813d9'], ['ea4146e3386ff1ac']]; loss = 0.079983
+Epoch 0: |                                                                                                                                                     | 1480/? [1:34:18<00:00,  0.26it/s, v_num=pmyy]context = [[50, 57, 67, 77, 79, 101], [66, 67, 75, 83, 96, 117], [53, 65, 70, 81, 130, 138], [19, 45, 53, 57, 65, 66]]target = [[76, 85, 70, 53, 90, 65], [87, 116, 103, 85, 111, 97], [80, 116, 87, 117, 54, 112], [35, 47, 59, 51, 53, 28]]
+Epoch 0: |                                                                                                                                                     | 1489/? [1:34:52<00:00,  0.26it/s, v_num=pmyy]train step 1490; scene = [['f73db02fdfe72073'], ['a4eeae8de8e2e98e']]; loss = 0.033035
+Epoch 0: |                                                                                                                                                     | 1490/? [1:34:55<00:00,  0.26it/s, v_num=pmyy]context = [[10, 19, 23, 29, 35, 40, 41, 42, 43, 53, 54, 57, 63, 65, 66, 69, 72, 73, 77, 84, 91, 94, 95, 107]]target = [[97, 79, 11, 35, 65, 87, 80, 51, 32, 13, 83, 72, 85, 34, 105, 92, 59, 25, 81, 40, 36, 82, 88, 15]]
+Epoch 0: |                                                                                                                                                     | 1499/? [1:35:30<00:00,  0.26it/s, v_num=pmyy]train step 1500; scene = [['fa1ddac84aafd9b7']]; loss = 0.170823
+Epoch 0: |                                                                                                                                                     | 1500/? [1:35:33<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1500; scene = ['45592a7f307bccd0'];
+target intrinsic:  tensor(0.8508, device='cuda:0') tensor(0.8510, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8519, device='cuda:0') tensor(0.8518, device='cuda:0')
+[2026-02-25 00:03:37,630][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1500/? [1:35:46<00:00,  0.26it/s, v_num=pmyy]context = [[5, 10, 19, 23, 27, 33, 40, 57, 60, 63, 64, 68], [97, 102, 107, 111, 113, 120, 137, 148, 157, 163, 168, 169]]target = [[54, 31, 25, 44, 17, 36, 32, 65, 16, 11, 23, 38], [99, 138, 103, 159, 164, 104, 123, 135, 120, 124, 130, 148]]
+Epoch 0: |                                                                                                                                                     | 1509/? [1:36:19<00:00,  0.26it/s, v_num=pmyy]train step 1510; scene = [['289a0ef9678c7c70'], ['f4fd1fee2a8f69ee'], ['a88b44d03528e2f2']]; loss = 0.039536
+Epoch 0: |                                                                                                                                                     | 1510/? [1:36:22<00:00,  0.26it/s, v_num=pmyy]context = [[1, 2, 8, 18, 19, 27, 33, 35, 48, 49, 63, 77], [0, 9, 12, 15, 20, 22, 23, 30, 34, 58, 66, 68]]target = [[12, 51, 13, 69, 65, 38, 6, 30, 37, 18, 53, 2], [8, 14, 60, 49, 63, 46, 47, 3, 35, 13, 21, 44]]
+Epoch 0: |                                                                                                                                                     | 1519/? [1:36:55<00:00,  0.26it/s, v_num=pmyy]train step 1520; scene = [['85cff5568e6f52e7'], ['9de970f9a14770a9'], ['a08eb87db37b694a'], ['0a8fda80930b52ae']]; loss = 0.040779
+Epoch 0: |                                                                                                                                                     | 1520/? [1:36:59<00:00,  0.26it/s, v_num=pmyy]context = [[6, 32, 33, 39, 62, 69, 77, 96], [32, 33, 42, 61, 83, 93, 99, 107], [13, 27, 41, 70, 78, 84, 86, 93]]target = [[28, 30, 62, 7, 57, 86, 77, 36], [101, 43, 41, 68, 46, 36, 47, 105], [27, 16, 29, 71, 40, 45, 74, 65]]
+Epoch 0: |                                                                                                                                                     | 1529/? [1:37:33<00:00,  0.26it/s, v_num=pmyy]train step 1530; scene = [['43cc195276cf0c56']]; loss = 0.103984
+Epoch 0: |                                                                                                                                                     | 1530/? [1:37:37<00:00,  0.26it/s, v_num=pmyy]context = [[86, 89, 91, 102, 109, 136, 173, 175], [1, 12, 32, 33, 36, 37, 43, 47], [15, 24, 40, 41, 44, 70, 73, 102]]target = [[138, 158, 117, 118, 170, 120, 165, 133], [37, 44, 31, 32, 24, 9, 40, 38], [16, 27, 88, 46, 69, 52, 73, 44]]
+Epoch 0: |                                                                                                                                                     | 1539/? [1:38:11<00:00,  0.26it/s, v_num=pmyy]train step 1540; scene = [['4b7f3f58b0838d38']]; loss = 0.039140
+Epoch 0: |                                                                                                                                                     | 1540/? [1:38:15<00:00,  0.26it/s, v_num=pmyy]context = [[17, 21, 24, 40, 42, 44, 49, 50, 60, 62, 63, 64, 65, 72, 73, 76, 77, 85, 89, 93, 96, 104, 108, 114]]target = [[104, 84, 41, 96, 107, 103, 94, 44, 111, 97, 83, 54, 100, 57, 27, 38, 113, 29, 18, 99, 76, 101, 55, 36]]
+Epoch 0: |                                                                                                                                                     | 1549/? [1:38:50<00:00,  0.26it/s, v_num=pmyy]train step 1550; scene = [['d6cc1a3af543a7f8'], ['f3dd12e8d9dd4c20'], ['b74b90e3a87f285c'], ['c51f5219c3d09e33'], ['3f6666062c86b73a'], ['3b7bd7e723f069b2'], ['b3992ad0aff60272'], ['30d52bc66d89221d']]; loss = 0.057441
+Epoch 0: |                                                                                                                                                     | 1550/? [1:38:54<00:00,  0.26it/s, v_num=pmyy]context = [[9, 79], [29, 75], [29, 95], [124, 170], [42, 102], [92, 159], [94, 157], [6, 75], [188, 255], [5, 76], [133, 205], [2, 90]]target = [[55, 20], [47, 52], [85, 42], [136, 147], [97, 56], [158, 99], [145, 143], [17, 9], [199, 198], [34, 26], [184, 165], [39, 28]]
+Epoch 0: |                                                                                                                                                     | 1559/? [1:39:29<00:00,  0.26it/s, v_num=pmyy]train step 1560; scene = [['f6a87eade96cceb1']]; loss = 0.054147
+Epoch 0: |                                                                                                                                                     | 1560/? [1:39:33<00:00,  0.26it/s, v_num=pmyy]context = [[9, 17, 23, 26, 33, 40, 46, 62, 63, 68, 69, 71], [117, 126, 137, 140, 146, 163, 169, 173, 178, 186, 201, 205]]target = [[25, 36, 53, 31, 59, 20, 68, 24, 66, 13, 33, 67], [128, 179, 169, 126, 188, 187, 122, 127, 144, 165, 150, 119]]
+Epoch 0: |                                                                                                                                                     | 1569/? [1:40:06<00:00,  0.26it/s, v_num=pmyy]train step 1570; scene = [['ab1c5358d3bb05db']]; loss = 0.034436
+Epoch 0: |                                                                                                                                                     | 1570/? [1:40:10<00:00,  0.26it/s, v_num=pmyy]context = [[6, 8, 14, 38, 44, 69, 83, 89], [31, 47, 49, 50, 68, 77, 89, 91], [32, 46, 50, 61, 66, 84, 89, 100]]target = [[69, 79, 13, 31, 70, 24, 8, 27], [53, 42, 44, 68, 90, 65, 87, 49], [85, 91, 86, 84, 67, 83, 72, 92]]
+Epoch 0: |                                                                                                                                                     | 1579/? [1:40:44<00:00,  0.26it/s, v_num=pmyy]train step 1580; scene = [['572357b6b69cb9ec'], ['326dd7b41ce515ac']]; loss = 0.043368
+Epoch 0: |                                                                                                                                                     | 1580/? [1:40:48<00:00,  0.26it/s, v_num=pmyy]context = [[26, 32, 46, 57, 64, 105], [32, 38, 66, 84, 85, 95], [75, 87, 98, 106, 108, 130], [7, 9, 12, 20, 48, 60]]target = [[83, 29, 52, 46, 45, 39], [41, 73, 88, 76, 54, 82], [78, 99, 126, 113, 86, 105], [29, 45, 35, 34, 32, 16]]
+Epoch 0: |                                                                                                                                                     | 1589/? [1:41:23<00:00,  0.26it/s, v_num=pmyy]train step 1590; scene = [['4558408811e71246'], ['4a3ff4b7939ec268']]; loss = 0.064918
+Epoch 0: |                                                                                                                                                     | 1590/? [1:41:27<00:00,  0.26it/s, v_num=pmyy]context = [[187, 198, 222, 224, 227, 240], [196, 213, 228, 234, 257, 263], [15, 24, 48, 51, 62, 85], [64, 73, 82, 83, 109, 117]]target = [[231, 221, 188, 222, 220, 232], [257, 228, 261, 254, 244, 210], [45, 49, 23, 62, 65, 84], [74, 94, 93, 85, 66, 104]]
+Epoch 0: |                                                                                                                                                     | 1599/? [1:42:01<00:00,  0.26it/s, v_num=pmyy]train step 1600; scene = [['1849a41079039a3a']]; loss = 0.047532
+Epoch 0: |                                                                                                                                                     | 1600/? [1:42:05<00:00,  0.26it/s, v_num=pmyy]context = [[7, 10, 13, 17, 18, 30, 32, 34, 35, 38, 42, 49, 64, 66, 71, 74, 77, 84, 89, 91, 93, 95, 103, 104]]target = [[52, 78, 50, 26, 46, 43, 79, 25, 89, 70, 58, 18, 103, 98, 16, 74, 20, 13, 85, 96, 71, 76, 91, 34]]
+[2026-02-25 00:10:01,185][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1609/? [1:42:46<00:00,  0.26it/s, v_num=pmyy]train step 1610; scene = [['6405acc04d687493'], ['a802441c288b598f'], ['e2a52c2d94f21d4d']]; loss = 0.059974
+Epoch 0: |                                                                                                                                                     | 1610/? [1:42:49<00:00,  0.26it/s, v_num=pmyy]context = [[110, 111, 116, 122, 133, 134, 135, 143, 145, 148, 155, 167, 173, 177, 178, 180, 181, 183, 192, 193, 197, 199, 203, 207]]target = [[137, 115, 169, 197, 176, 116, 146, 159, 144, 119, 129, 161, 181, 160, 151, 196, 131, 204, 201, 157, 138, 123, 180, 117]]
+Epoch 0: |                                                                                                                                                     | 1619/? [1:43:23<00:00,  0.26it/s, v_num=pmyy]train step 1620; scene = [['a270004198eddd9b'], ['cc306ca65c5dbc24'], ['1a5022549d2718ae'], ['cdde30f3352829cb']]; loss = 0.054471
+Epoch 0: |                                                                                                                                                     | 1620/? [1:43:28<00:00,  0.26it/s, v_num=pmyy]context = [[72, 79, 103, 136], [6, 15, 31, 60], [1, 48, 52, 80], [95, 135, 145, 162], [136, 166, 176, 206], [0, 7, 45, 61]]target = [[90, 120, 96, 123], [15, 25, 11, 32], [5, 49, 65, 22], [116, 131, 128, 139], [144, 141, 183, 190], [34, 48, 35, 12]]
+Epoch 0: |                                                                                                                                                     | 1629/? [1:44:02<00:00,  0.26it/s, v_num=pmyy]train step 1630; scene = [['22c6139e04c88c0c']]; loss = 0.037232
+Epoch 0: |                                                                                                                                                     | 1630/? [1:44:06<00:00,  0.26it/s, v_num=pmyy]context = [[5, 7, 10, 22, 36, 52], [7, 11, 21, 27, 46, 72], [0, 50, 64, 79, 88, 90], [4, 24, 28, 39, 44, 49]]target = [[41, 40, 13, 50, 28, 25], [39, 20, 47, 50, 25, 55], [75, 69, 56, 70, 5, 13], [27, 35, 24, 13, 14, 40]]
+Epoch 0: |                                                                                                                                                     | 1639/? [1:44:41<00:00,  0.26it/s, v_num=pmyy]train step 1640; scene = [['62ce0e4e9490317d']]; loss = 0.027289
+Epoch 0: |                                                                                                                                                     | 1640/? [1:44:45<00:00,  0.26it/s, v_num=pmyy]context = [[4, 6, 8, 15, 23, 24, 26, 32, 35, 39, 50, 55], [0, 10, 22, 26, 27, 29, 35, 42, 60, 66, 73, 77]]target = [[41, 43, 30, 44, 33, 16, 39, 51, 12, 54, 53, 29], [4, 16, 21, 43, 47, 46, 44, 53, 45, 24, 51, 50]]
+Epoch 0: |                                                                                                                                                     | 1649/? [1:45:20<00:00,  0.26it/s, v_num=pmyy]train step 1650; scene = [['bad5f833677cd32f'], ['2cf78f93fd569aa1']]; loss = 0.058665
+Epoch 0: |                                                                                                                                                     | 1650/? [1:45:24<00:00,  0.26it/s, v_num=pmyy]context = [[9, 71], [34, 102], [46, 110], [10, 61], [137, 186], [1, 52], [128, 195], [10, 100], [198, 258], [16, 67], [12, 66], [5, 66]]target = [[26, 42], [83, 46], [67, 48], [14, 57], [154, 155], [49, 29], [155, 140], [41, 56], [236, 235], [32, 61], [50, 32], [35, 61]]
+Epoch 0: |                                                                                                                                                     | 1659/? [1:45:59<00:00,  0.26it/s, v_num=pmyy]train step 1660; scene = [['1f3f484a027f93d9'], ['12e85fb6e140ee85'], ['197edde42c1eaac3'], ['59058fb21817aa6b']]; loss = 0.050054
+Epoch 0: |                                                                                                                                                     | 1660/? [1:46:03<00:00,  0.26it/s, v_num=pmyy]context = [[24, 30, 39, 55, 81, 105], [182, 183, 195, 207, 215, 233], [25, 65, 77, 78, 93, 105], [20, 40, 46, 77, 82, 103]]target = [[48, 70, 100, 51, 59, 97], [197, 190, 194, 208, 198, 218], [44, 91, 86, 70, 101, 30], [65, 21, 102, 25, 30, 54]]
+Epoch 0: |                                                                                                                                                     | 1669/? [1:46:37<00:00,  0.26it/s, v_num=pmyy]train step 1670; scene = [['25aff15f6c54558b']]; loss = 0.038386
+Epoch 0: |                                                                                                                                                     | 1670/? [1:46:41<00:00,  0.26it/s, v_num=pmyy]context = [[154, 158, 162, 163, 165, 191, 194, 196, 201, 203, 207, 211], [137, 138, 151, 153, 161, 165, 166, 167, 185, 186, 197, 203]]target = [[203, 173, 193, 192, 207, 165, 187, 157, 174, 190, 182, 176], [165, 166, 189, 176, 201, 140, 139, 163, 184, 169, 186, 157]]
+Epoch 0: |                                                                                                                                                     | 1679/? [1:47:16<00:00,  0.26it/s, v_num=pmyy]train step 1680; scene = [['c513a3c2f59aa548'], ['493538f3442cb9fd']]; loss = 0.043610
+Epoch 0: |                                                                                                                                                     | 1680/? [1:47:19<00:00,  0.26it/s, v_num=pmyy]context = [[8, 9, 12, 14, 16, 17, 18, 20, 23, 25, 29, 36, 38, 46, 52, 63, 64, 78, 82, 90, 95, 97, 102, 105]]target = [[19, 24, 49, 13, 77, 40, 51, 54, 85, 10, 74, 88, 50, 94, 78, 71, 68, 87, 44, 95, 63, 81, 82, 96]]
+Epoch 0: |                                                                                                                                                     | 1689/? [1:47:54<00:00,  0.26it/s, v_num=pmyy]train step 1690; scene = [['da5de01cf7e41541']]; loss = 0.070067
+Epoch 0: |                                                                                                                                                     | 1690/? [1:47:58<00:00,  0.26it/s, v_num=pmyy]context = [[166, 168, 170, 180, 181, 183, 184, 191, 219, 234, 239, 240], [52, 57, 62, 79, 87, 99, 104, 107, 111, 115, 118, 124]]target = [[226, 177, 191, 171, 214, 212, 213, 234, 229, 186, 238, 219], [60, 111, 110, 109, 72, 56, 96, 101, 61, 66, 118, 69]]
+Epoch 0: |                                                                                                                                                     | 1699/? [1:48:32<00:00,  0.26it/s, v_num=pmyy]train step 1700; scene = [['3e763e3c28e87eed']]; loss = 0.020947
+Epoch 0: |                                                                                                                                                     | 1700/? [1:48:36<00:00,  0.26it/s, v_num=pmyy]context = [[45, 50, 54, 60, 92, 105, 108, 110], [0, 8, 30, 31, 41, 42, 43, 45], [22, 28, 31, 33, 41, 59, 72, 99]]target = [[71, 52, 86, 64, 72, 69, 73, 59], [14, 36, 2, 1, 35, 4, 16, 20], [74, 61, 93, 94, 31, 42, 26, 28]]
+Epoch 0: |                                                                                                                                                     | 1709/? [1:49:09<00:00,  0.26it/s, v_num=pmyy]train step 1710; scene = [['3727bb4b44708f89']]; loss = 0.087378
+Epoch 0: |                                                                                                                                                     | 1710/? [1:49:13<00:00,  0.26it/s, v_num=pmyy]context = [[59, 62, 71, 76, 105, 110, 112, 114], [1, 2, 12, 23, 30, 34, 44, 48], [75, 78, 86, 87, 90, 114, 117, 122]]target = [[104, 100, 89, 101, 94, 102, 96, 84], [4, 24, 32, 21, 39, 26, 25, 20], [89, 94, 76, 90, 85, 96, 83, 77]]
+Epoch 0: |                                                                                                                                                     | 1719/? [1:49:47<00:00,  0.26it/s, v_num=pmyy]train step 1720; scene = [['d938c73738634cbf'], ['d7f3b4e12f3f3af6'], ['1a17332c1d690519']]; loss = 0.029354
+Epoch 0: |                                                                                                                                                     | 1720/? [1:49:51<00:00,  0.26it/s, v_num=pmyy]context = [[26, 28, 33, 44, 45, 49, 55, 57, 58, 59, 67, 72, 75, 79, 82, 94, 98, 100, 113, 115, 117, 118, 122, 123]]target = [[107, 48, 44, 86, 57, 93, 73, 51, 58, 88, 78, 59, 33, 99, 100, 87, 62, 38, 36, 84, 119, 122, 91, 74]]
+Epoch 0: |                                                                                                                                                     | 1729/? [1:50:26<00:00,  0.26it/s, v_num=pmyy]train step 1730; scene = [['2d9ea631e6423141'], ['b0c6597c77c51a8c']]; loss = 0.047816
+Epoch 0: |                                                                                                                                                     | 1730/? [1:50:30<00:00,  0.26it/s, v_num=pmyy]context = [[87, 105, 125, 127, 131, 138, 143, 150], [1, 14, 30, 33, 45, 48, 57, 66], [8, 18, 21, 26, 59, 92, 93, 94]]target = [[89, 146, 124, 104, 103, 96, 110, 143], [29, 33, 7, 12, 20, 27, 53, 51], [57, 81, 93, 74, 29, 30, 33, 60]]
+Epoch 0: |                                                                                                                                                     | 1739/? [1:51:05<00:00,  0.26it/s, v_num=pmyy]train step 1740; scene = [['56863fb499e2ff9a'], ['f5a3eca31fcdabf8'], ['946597adeb926d39'], ['23eecb3bfb301179'], ['4d999bce04f7516a'], ['b1f32a7b0a8e25ce'], ['72ac5af57c88795c'], ['cc45417d1c5dab8d']]; loss = 0.070275
+Epoch 0: |                                                                                                                                                     | 1740/? [1:51:09<00:00,  0.26it/s, v_num=pmyy]context = [[1, 4, 8, 12, 19, 26, 28, 29, 39, 42, 47, 50, 55, 66, 67, 68, 70, 76, 83, 85, 91, 92, 97, 98]]target = [[72, 40, 83, 3, 21, 13, 14, 61, 46, 53, 36, 57, 6, 10, 12, 37, 7, 22, 19, 78, 56, 25, 54, 80]]
+Epoch 0: |                                                                                                                                                     | 1749/? [1:51:44<00:00,  0.26it/s, v_num=pmyy]train step 1750; scene = [['7351b1a8a7405871'], ['edacb8db81943446']]; loss = 0.049949
+Epoch 0: |                                                                                                                                                     | 1750/? [1:51:48<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1750; scene = ['3b273cb40c55db95'];
+target intrinsic:  tensor(1.0504, device='cuda:0') tensor(1.0506, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(1.0464, device='cuda:0') tensor(1.0464, device='cuda:0')
+[2026-02-25 00:19:40,967][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1750/? [1:51:49<00:00,  0.26it/s, v_num=pmyy]context = [[6, 12, 18, 21, 26, 27, 30, 33, 34, 35, 50, 55], [2, 5, 11, 19, 32, 36, 39, 44, 45, 46, 47, 56]]target = [[43, 26, 45, 30, 52, 34, 39, 12, 21, 11, 37, 19], [8, 19, 54, 25, 33, 4, 10, 55, 46, 34, 52, 3]]
+Epoch 0: |                                                                                                                                                     | 1759/? [1:52:22<00:00,  0.26it/s, v_num=pmyy]train step 1760; scene = [['616d045c5963ecb2'], ['4bbb9fadc993969a'], ['c8b833961ffd8ac5'], ['321addf5693a66aa']]; loss = 0.039439
+Epoch 0: |                                                                                                                                                     | 1760/? [1:52:26<00:00,  0.26it/s, v_num=pmyy]context = [[111, 116, 130, 132, 142, 146, 147, 158, 160, 164, 176, 181], [19, 20, 21, 24, 28, 31, 47, 58, 61, 70, 71, 77]]target = [[125, 122, 167, 145, 144, 152, 130, 121, 153, 165, 113, 139], [70, 73, 42, 59, 66, 23, 39, 72, 22, 60, 26, 25]]
+Epoch 0: |                                                                                                                                                     | 1769/? [1:52:59<00:00,  0.26it/s, v_num=pmyy]train step 1770; scene = [['8eb7ab71f603c9d4'], ['cc966a9b2af2232f']]; loss = 0.055111
+Epoch 0: |                                                                                                                                                     | 1770/? [1:53:03<00:00,  0.26it/s, v_num=pmyy]context = [[102, 104, 107, 108, 112, 122, 123, 126, 129, 133, 136, 142, 146, 147, 150, 161, 162, 163, 164, 184, 189, 191, 193, 199]]target = [[120, 136, 118, 149, 133, 122, 193, 137, 187, 174, 159, 156, 194, 170, 121, 189, 103, 178, 197, 196, 129, 195, 115, 181]]
+Epoch 0: |                                                                                                                                                     | 1779/? [1:53:38<00:00,  0.26it/s, v_num=pmyy]train step 1780; scene = [['bb28a1c09df8a484']]; loss = 0.035646
+Epoch 0: |                                                                                                                                                     | 1780/? [1:53:42<00:00,  0.26it/s, v_num=pmyy]context = [[11, 70], [25, 75], [78, 152], [51, 119], [24, 81], [136, 193], [20, 96], [30, 109], [7, 87], [58, 129], [66, 145], [27, 74]]target = [[17, 37], [64, 73], [135, 104], [69, 118], [42, 58], [145, 168], [30, 83], [99, 43], [42, 16], [61, 111], [134, 98], [55, 71]]
+Epoch 0: |                                                                                                                                                     | 1789/? [1:54:16<00:00,  0.26it/s, v_num=pmyy]train step 1790; scene = [['48c2b63b452555af'], ['1aa5fdc9ff855ee2']]; loss = 0.056721
+Epoch 0: |                                                                                                                                                     | 1790/? [1:54:20<00:00,  0.26it/s, v_num=pmyy]context = [[32, 36, 37, 42, 43, 52, 63, 67, 68, 73, 74, 90], [27, 40, 53, 55, 57, 63, 68, 79, 80, 81, 100, 102]]target = [[70, 67, 89, 83, 52, 35, 74, 73, 84, 44, 54, 33], [44, 60, 41, 98, 100, 51, 46, 99, 47, 92, 37, 31]]
+Epoch 0: |                                                                                                                                                     | 1799/? [1:54:55<00:00,  0.26it/s, v_num=pmyy]train step 1800; scene = [['b41ece74dd5aa87d'], ['e620345f5468d2e3']]; loss = 0.024444
+Epoch 0: |                                                                                                                                                     | 1800/? [1:54:58<00:00,  0.26it/s, v_num=pmyy]context = [[2, 47], [107, 163], [22, 107], [67, 146], [82, 130], [87, 151], [59, 133], [191, 242], [121, 203], [113, 167], [106, 159], [27, 94]]target = [[37, 42], [161, 109], [92, 85], [142, 139], [116, 104], [106, 134], [116, 87], [201, 206], [142, 129], [159, 155], [116, 134], [57, 36]]
+[2026-02-25 00:22:54,741][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1809/? [1:55:31<00:00,  0.26it/s, v_num=pmyy]train step 1810; scene = [['9e1b58ec292d6690'], ['63f6e847b2e2a2e7'], ['aa570297256d99b6'], ['02195bc2cdded89e'], ['8839e65c6f6bb394'], ['c6bd8d3b3367444f']]; loss = 0.064478
+Epoch 0: |                                                                                                                                                     | 1810/? [1:55:35<00:00,  0.26it/s, v_num=pmyy]context = [[3, 14, 18, 21, 23, 31, 32, 34, 36, 63, 70, 78], [24, 32, 33, 47, 62, 69, 70, 73, 80, 85, 86, 88]]target = [[5, 7, 41, 28, 48, 67, 12, 68, 47, 26, 10, 31], [25, 68, 40, 79, 82, 58, 26, 28, 36, 48, 66, 29]]
+Epoch 0: |                                                                                                                                                     | 1819/? [1:56:09<00:00,  0.26it/s, v_num=pmyy]train step 1820; scene = [['8ad3918394be8d98']]; loss = 0.053691
+Epoch 0: |                                                                                                                                                     | 1820/? [1:56:13<00:00,  0.26it/s, v_num=pmyy]context = [[0, 4, 7, 9, 13, 15, 17, 35, 37, 61, 69, 72], [5, 23, 24, 25, 26, 31, 50, 56, 60, 68, 74, 76]]target = [[45, 49, 51, 46, 6, 14, 52, 22, 65, 69, 10, 37], [9, 37, 72, 6, 28, 14, 64, 47, 15, 16, 55, 33]]
+Epoch 0: |                                                                                                                                                     | 1829/? [1:56:48<00:00,  0.26it/s, v_num=pmyy]train step 1830; scene = [['af4565fb713ed79f']]; loss = 0.031482
+Epoch 0: |                                                                                                                                                     | 1830/? [1:56:51<00:00,  0.26it/s, v_num=pmyy]context = [[133, 138, 168, 169, 171, 180, 190, 192, 196, 203, 204, 213], [81, 82, 84, 87, 92, 93, 95, 97, 105, 110, 118, 131]]target = [[147, 196, 136, 153, 201, 185, 205, 161, 169, 149, 210, 189], [114, 103, 99, 119, 107, 110, 88, 124, 86, 122, 126, 120]]
+Epoch 0: |                                                                                                                                                     | 1839/? [1:57:27<00:00,  0.26it/s, v_num=pmyy]train step 1840; scene = [['c690f7b93b2b75c5'], ['f0dbc908f65693ae'], ['ee81b7d704303a2c'], ['a2a3a75cafb630f5']]; loss = 0.074975
+Epoch 0: |                                                                                                                                                     | 1840/? [1:57:30<00:00,  0.26it/s, v_num=pmyy]context = [[95, 106, 107, 115, 116, 118, 120, 126, 128, 130, 133, 139, 148, 149, 150, 153, 159, 163, 173, 178, 182, 183, 184, 192]]target = [[167, 153, 132, 96, 110, 126, 111, 129, 162, 144, 185, 176, 106, 97, 156, 108, 177, 172, 148, 157, 133, 137, 149, 141]]
+Epoch 0: |                                                                                                                                                     | 1849/? [1:58:05<00:00,  0.26it/s, v_num=pmyy]train step 1850; scene = [['a1fa4bf737381a53'], ['2d7243929156069e'], ['201287bd653fa906'], ['2d7243929156069e'], ['d25d6a9b1356b134'], ['4a0f95a3db913b56']]; loss = 0.059825
+Epoch 0: |                                                                                                                                                     | 1850/? [1:58:08<00:00,  0.26it/s, v_num=pmyy]context = [[30, 32, 34, 52, 72, 85, 90, 108], [9, 25, 30, 33, 41, 46, 48, 84], [51, 72, 76, 82, 83, 99, 109, 115]]target = [[46, 41, 107, 87, 40, 64, 57, 53], [35, 56, 83, 38, 62, 21, 53, 37], [109, 111, 114, 53, 86, 67, 91, 104]]
+Epoch 0: |                                                                                                                                                     | 1859/? [1:58:41<00:00,  0.26it/s, v_num=pmyy]train step 1860; scene = [['082779aebfbf4a46']]; loss = 0.039268
+Epoch 0: |                                                                                                                                                     | 1860/? [1:58:45<00:00,  0.26it/s, v_num=pmyy]context = [[3, 17, 39, 49], [112, 121, 139, 157], [128, 136, 147, 178], [2, 10, 33, 87], [25, 50, 81, 98], [28, 48, 69, 113]]target = [[8, 40, 43, 26], [113, 124, 154, 133], [156, 164, 137, 168], [62, 24, 86, 11], [68, 70, 78, 73], [73, 100, 109, 29]]
+Epoch 0: |                                                                                                                                                     | 1869/? [1:59:18<00:00,  0.26it/s, v_num=pmyy]train step 1870; scene = [['01c143c81a4d2145'], ['9b0b82db99ff360a'], ['80f5da17b4119bbf'], ['8c5dce5d79b3d2aa'], ['d3a0a89d951a6101'], ['da1f9f2859b59142'], ['20b791440b7ec0b5'], ['00703cbf7531ef11'], ['4aef5d4b9287e08a'], ['27ed8e7077af6540'], ['04e7f97215df7078'], ['08735c801ab7efb5']]; loss = 0.084776
+Epoch 0: |                                                                                                                                                     | 1870/? [1:59:22<00:00,  0.26it/s, v_num=pmyy]context = [[41, 44, 47, 48, 50, 51, 56, 67, 72, 81, 82, 83, 84, 85, 89, 90, 97, 105, 107, 110, 115, 125, 135, 138]]target = [[79, 108, 90, 99, 85, 65, 59, 69, 73, 126, 131, 100, 114, 52, 87, 47, 123, 54, 43, 137, 94, 88, 70, 75]]
+Epoch 0: |                                                                                                                                                     | 1879/? [1:59:57<00:00,  0.26it/s, v_num=pmyy]train step 1880; scene = [['42b2dabdc5ea4d93']]; loss = 0.029754
+Epoch 0: |                                                                                                                                                     | 1880/? [2:00:00<00:00,  0.26it/s, v_num=pmyy]context = [[17, 29, 31, 34, 36, 39, 48, 59, 64, 66, 67, 79], [42, 47, 51, 56, 57, 61, 64, 70, 74, 83, 92, 93]]target = [[53, 20, 25, 74, 39, 30, 21, 59, 50, 66, 40, 44], [43, 84, 45, 72, 82, 85, 56, 73, 92, 76, 75, 46]]
+Epoch 0: |                                                                                                                                                     | 1889/? [2:00:36<00:00,  0.26it/s, v_num=pmyy]train step 1890; scene = [['ed4855892fd5fa4a'], ['9683be82b6c1e851']]; loss = 0.051107
+Epoch 0: |                                                                                                                                                     | 1890/? [2:00:39<00:00,  0.26it/s, v_num=pmyy]context = [[22, 37, 39, 49, 50, 52, 54, 56, 59, 61, 72, 74, 84, 88, 93, 94, 97, 102, 104, 105, 108, 109, 112, 119]]target = [[59, 53, 102, 34, 27, 44, 100, 45, 32, 65, 36, 42, 87, 82, 85, 26, 80, 60, 90, 113, 86, 78, 40, 68]]
+Epoch 0: |                                                                                                                                                     | 1899/? [2:01:14<00:00,  0.26it/s, v_num=pmyy]train step 1900; scene = [['1844e80dafb62927'], ['b43f32da9a7800ee'], ['11094b4b71a9ab00'], ['0f5daf32c2e8fefd'], ['f8e7bd9e403fc04a'], ['b11afc47f2feb21c'], ['5994966dd0897ead'], ['6d2aae2a7dd35e14'], ['e545aecc18cfa501'], ['82d896f5142ee6dd'], ['140b10a4f6bb5aa5'], ['2ec8edfc07c8841f']]; loss = 0.079810
+Epoch 0: |                                                                                                                                                     | 1900/? [2:01:18<00:00,  0.26it/s, v_num=pmyy]context = [[1, 20, 26, 32, 34, 53], [83, 87, 95, 106, 125, 139], [0, 13, 29, 39, 46, 77], [90, 122, 126, 130, 159, 167]]target = [[36, 12, 39, 17, 4, 7], [92, 86, 107, 112, 135, 137], [32, 69, 36, 28, 50, 38], [132, 161, 105, 165, 142, 102]]
+Epoch 0: |                                                                                                                                                     | 1909/? [2:01:52<00:00,  0.26it/s, v_num=pmyy]train step 1910; scene = [['50f7971dda42084f'], ['159a472b24f1c395'], ['04a191933c5b05ec']]; loss = 0.073106
+Epoch 0: |                                                                                                                                                     | 1910/? [2:01:56<00:00,  0.26it/s, v_num=pmyy]context = [[74, 78, 86, 94, 96, 101, 104, 107, 108, 109, 115, 117, 119, 131, 133, 135, 137, 144, 151, 152, 155, 158, 161, 171]]target = [[108, 142, 163, 121, 87, 111, 149, 96, 120, 145, 97, 122, 147, 101, 123, 95, 94, 132, 86, 89, 99, 140, 104, 136]]
+Epoch 0: |                                                                                                                                                     | 1919/? [2:02:31<00:00,  0.26it/s, v_num=pmyy]train step 1920; scene = [['bf4bbf858718317d'], ['39ed8d2efe760b94'], ['d9644f4985e51a1f']]; loss = 0.048118
+Epoch 0: |                                                                                                                                                     | 1920/? [2:02:34<00:00,  0.26it/s, v_num=pmyy]context = [[9, 79], [15, 63], [28, 73], [101, 167], [5, 70], [53, 102], [2, 75], [86, 135], [1, 58], [5, 84], [1, 76], [135, 208]]target = [[43, 20], [62, 45], [70, 64], [114, 144], [60, 52], [93, 101], [13, 68], [120, 89], [30, 23], [43, 16], [57, 14], [194, 205]]
+Epoch 0: |                                                                                                                                                     | 1929/? [2:03:07<00:00,  0.26it/s, v_num=pmyy]train step 1930; scene = [['6efa62598fececd0'], ['b25a0f4ffca51d79'], ['90dbfac63e6b89be'], ['08a366317a388734']]; loss = 0.064742
+Epoch 0: |                                                                                                                                                     | 1930/? [2:03:10<00:00,  0.26it/s, v_num=pmyy]context = [[3, 19, 23, 28, 32, 40, 48, 50, 53, 60, 63, 64, 66, 70, 74, 78, 80, 82, 87, 93, 96, 97, 98, 100]]target = [[63, 24, 82, 55, 68, 93, 95, 7, 53, 65, 76, 35, 83, 11, 22, 34, 52, 16, 61, 92, 26, 47, 32, 46]]
+Epoch 0: |                                                                                                                                                     | 1939/? [2:03:45<00:00,  0.26it/s, v_num=pmyy]train step 1940; scene = [['1236365ec263ad76']]; loss = 0.034800
+Epoch 0: |                                                                                                                                                     | 1940/? [2:03:49<00:00,  0.26it/s, v_num=pmyy]context = [[10, 11, 16, 23, 26, 33, 34, 43, 44, 48, 54, 58, 70, 73, 75, 78, 82, 86, 89, 91, 102, 105, 106, 107]]target = [[88, 68, 24, 39, 22, 100, 19, 44, 57, 94, 41, 37, 53, 69, 96, 63, 81, 104, 50, 64, 101, 15, 40, 26]]
+Epoch 0: |                                                                                                                                                     | 1949/? [2:04:24<00:00,  0.26it/s, v_num=pmyy]train step 1950; scene = [['2bc8b64aafc5870c'], ['1202c32d91ad3ee3'], ['300571576edc008c']]; loss = 0.099111
+Epoch 0: |                                                                                                                                                     | 1950/? [2:04:28<00:00,  0.26it/s, v_num=pmyy]context = [[11, 14, 25, 42, 47, 49, 58, 66, 68, 69, 79, 81], [39, 55, 56, 60, 63, 70, 74, 86, 93, 96, 99, 104]]target = [[25, 55, 33, 13, 26, 62, 22, 52, 32, 67, 30, 40], [95, 42, 43, 41, 59, 96, 102, 88, 48, 66, 100, 40]]
+Epoch 0: |                                                                                                                                                     | 1959/? [2:05:02<00:00,  0.26it/s, v_num=pmyy]train step 1960; scene = [['3ded3e3c1fe76ee3']]; loss = 0.042400
+Epoch 0: |                                                                                                                                                     | 1960/? [2:05:06<00:00,  0.26it/s, v_num=pmyy]context = [[109, 112, 117, 120, 130, 137, 142, 144, 145, 150, 151, 152, 155, 159, 160, 162, 166, 171, 175, 179, 194, 195, 196, 206]]target = [[168, 171, 125, 169, 131, 150, 188, 139, 145, 152, 126, 184, 205, 142, 119, 196, 148, 155, 185, 197, 154, 203, 143, 147]]
+Epoch 0: |                                                                                                                                                     | 1969/? [2:05:40<00:00,  0.26it/s, v_num=pmyy]train step 1970; scene = [['3b21f48e23e4917f']]; loss = 0.037027
+Epoch 0: |                                                                                                                                                     | 1970/? [2:05:44<00:00,  0.26it/s, v_num=pmyy]context = [[13, 15, 19, 21, 30, 33, 39, 49, 54, 57, 72, 85], [18, 26, 35, 36, 39, 46, 50, 60, 70, 71, 74, 80]]target = [[39, 30, 38, 35, 26, 78, 75, 56, 18, 44, 77, 71], [22, 24, 70, 65, 49, 19, 62, 59, 71, 73, 51, 39]]
+Epoch 0: |                                                                                                                                                     | 1979/? [2:06:18<00:00,  0.26it/s, v_num=pmyy]train step 1980; scene = [['723db63d24c84d1d']]; loss = 0.042857
+Epoch 0: |                                                                                                                                                     | 1980/? [2:06:22<00:00,  0.26it/s, v_num=pmyy]context = [[114, 115, 120, 122, 123, 125, 126, 140, 141, 145, 147, 161, 170, 179, 180, 181, 193, 195, 196, 197, 199, 203, 207, 211]]target = [[201, 148, 171, 190, 135, 143, 189, 141, 205, 151, 187, 206, 183, 144, 200, 147, 181, 184, 120, 207, 152, 128, 170, 204]]
+Epoch 0: |                                                                                                                                                     | 1989/? [2:06:57<00:00,  0.26it/s, v_num=pmyy]train step 1990; scene = [['12eb36bba5c89eeb'], ['40c5311e3b3accef'], ['06d2876e8c40a3b6'], ['2da17464ef895b63'], ['21951a6ae1c4b225'], ['ab78b3eb64029b73'], ['f6ef16edbf87f358'], ['66e4f3268dafe823']]; loss = 0.081296
+Epoch 0: |                                                                                                                                                     | 1990/? [2:07:00<00:00,  0.26it/s, v_num=pmyy]context = [[6, 9, 11, 12, 24, 34, 39, 44, 45, 53, 55, 59, 61, 66, 70, 71, 76, 78, 79, 82, 85, 87, 92, 103]]target = [[32, 63, 22, 65, 39, 36, 57, 91, 64, 13, 66, 60, 14, 71, 26, 41, 23, 101, 29, 93, 92, 90, 98, 73]]
+Epoch 0: |                                                                                                                                                     | 1999/? [2:07:35<00:00,  0.26it/s, v_num=pmyy]train step 2000; scene = [['1ac1373478877088']]; loss = 0.048109
+Epoch 0: |                                                                                                                                                     | 2000/? [2:07:39<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 2000; scene = ['be75142d4652fe3e'];
+target intrinsic:  tensor(0.9402, device='cuda:0') tensor(0.9404, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8845, device='cuda:0') tensor(0.8853, device='cuda:0')
+[2026-02-25 00:35:31,652][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2000/? [2:07:40<00:00,  0.26it/s, v_num=pmyy]context = [[16, 17, 18, 27, 35, 40, 41, 47, 48, 51, 52, 53, 54, 56, 79, 80, 87, 89, 92, 96, 100, 101, 106, 113]]target = [[102, 53, 81, 57, 21, 101, 52, 64, 43, 58, 59, 94, 46, 108, 32, 79, 36, 104, 26, 106, 31, 29, 60, 41]]
+[2026-02-25 00:35:35,995][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2009/? [2:08:18<00:00,  0.26it/s, v_num=pmyy]train step 2010; scene = [['e2f2a27bfce53270']]; loss = 0.034677
+Epoch 0: |                                                                                                                                                     | 2010/? [2:08:22<00:00,  0.26it/s, v_num=pmyy]context = [[32, 34, 44, 61, 73, 79], [66, 91, 112, 119, 124, 128], [52, 79, 88, 91, 97, 100], [65, 83, 97, 105, 110, 128]]target = [[68, 72, 41, 73, 74, 40], [74, 67, 110, 72, 105, 98], [93, 66, 74, 77, 83, 65], [81, 72, 117, 80, 83, 114]]
+Epoch 0: |                                                                                                                                                     | 2019/? [2:08:55<00:00,  0.26it/s, v_num=pmyy]train step 2020; scene = [['71f80e4a8d79031c'], ['aec60a6dd8090690'], ['475ee0875ccb635a']]; loss = 0.046994
+Epoch 0: |                                                                                                                                                     | 2020/? [2:08:59<00:00,  0.26it/s, v_num=pmyy]context = [[1, 4, 12, 16, 21, 23, 29, 40, 49, 54, 58, 60, 62, 74, 75, 79, 81, 82, 85, 91, 92, 94, 97, 98]]target = [[52, 69, 55, 22, 50, 24, 41, 46, 80, 58, 32, 92, 54, 81, 38, 23, 3, 29, 97, 40, 25, 39, 78, 73]]
+Epoch 0: |                                                                                                                                                     | 2029/? [2:09:34<00:00,  0.26it/s, v_num=pmyy]train step 2030; scene = [['60b6914bafa3c3f2']]; loss = 0.063650
+Epoch 0: |                                                                                                                                                     | 2030/? [2:09:38<00:00,  0.26it/s, v_num=pmyy]context = [[45, 87, 105], [54, 118, 143], [79, 120, 143], [186, 243, 262], [14, 67, 94], [27, 68, 76], [13, 28, 58], [1, 69, 80]]target = [[68, 84, 80], [134, 122, 96], [131, 97, 136], [244, 198, 209], [34, 77, 22], [70, 43, 37], [43, 32, 33], [15, 54, 19]]
+Epoch 0: |                                                                                                                                                     | 2039/? [2:10:12<00:00,  0.26it/s, v_num=pmyy]train step 2040; scene = [['8e94de9c1bc0732c'], ['0bb8a7807f7095fd']]; loss = 0.028131
+Epoch 0: |                                                                                                                                                     | 2040/? [2:10:15<00:00,  0.26it/s, v_num=pmyy]context = [[25, 79, 84, 91], [12, 32, 41, 66], [10, 17, 31, 64], [69, 71, 109, 114], [31, 49, 67, 84], [14, 27, 46, 84]]target = [[48, 80, 51, 88], [16, 41, 21, 33], [63, 61, 13, 38], [73, 107, 79, 90], [41, 35, 60, 56], [29, 80, 26, 25]]
+Epoch 0: |                                                                                                                                                     | 2049/? [2:10:50<00:00,  0.26it/s, v_num=pmyy]train step 2050; scene = [['c4e12df63403eadf'], ['8359c9726f078a38'], ['e0b75e74fdeffde9']]; loss = 0.065066
+Epoch 0: |                                                                                                                                                     | 2050/? [2:10:54<00:00,  0.26it/s, v_num=pmyy]context = [[152, 157, 160, 167, 191, 197], [71, 81, 105, 126, 127, 139], [10, 12, 47, 54, 66, 83], [85, 88, 106, 117, 127, 136]]target = [[167, 195, 187, 184, 157, 176], [120, 107, 92, 124, 115, 110], [14, 13, 80, 40, 24, 47], [93, 96, 97, 132, 102, 131]]
+Epoch 0: |                                                                                                                                                     | 2059/? [2:11:27<00:00,  0.26it/s, v_num=pmyy]train step 2060; scene = [['22062ed897320134'], ['08e4f6a5098b0d3a'], ['d3057752d15cc3ed']]; loss = 0.093615
+Epoch 0: |                                                                                                                                                     | 2060/? [2:11:31<00:00,  0.26it/s, v_num=pmyy]context = [[164, 195, 232], [18, 42, 81], [5, 55, 93], [175, 199, 220], [6, 42, 65], [2, 7, 60], [149, 190, 195], [33, 42, 82]]target = [[209, 178, 225], [53, 24, 75], [32, 35, 24], [201, 183, 204], [50, 59, 38], [5, 37, 42], [187, 161, 189], [44, 78, 62]]
+Epoch 0: |                                                                                                                                                     | 2069/? [2:12:06<00:00,  0.26it/s, v_num=pmyy]train step 2070; scene = [['85ef0eb4a42e3425'], ['703430ad773c95bc'], ['4fd9d45647d536e5']]; loss = 0.129975
+Epoch 0: |                                                                                                                                                     | 2070/? [2:12:10<00:00,  0.26it/s, v_num=pmyy]context = [[78, 84, 92, 97, 108, 109, 115, 124, 138, 150, 158, 159], [42, 43, 53, 57, 69, 71, 77, 83, 84, 86, 88, 92]]target = [[127, 113, 118, 102, 153, 144, 132, 139, 125, 129, 84, 79], [57, 81, 79, 87, 72, 61, 48, 88, 54, 84, 44, 65]]
+Epoch 0: |                                                                                                                                                     | 2079/? [2:12:44<00:00,  0.26it/s, v_num=pmyy]train step 2080; scene = [['968b857dda7e955a'], ['8e713ad26d00feac']]; loss = 0.038065
+Epoch 0: |                                                                                                                                                     | 2080/? [2:12:48<00:00,  0.26it/s, v_num=pmyy]context = [[33, 57, 88, 112], [20, 29, 43, 76], [20, 56, 96, 110], [3, 13, 50, 53], [44, 50, 64, 108], [65, 92, 120, 152]]target = [[101, 99, 69, 88], [37, 25, 36, 30], [99, 72, 91, 75], [14, 32, 43, 39], [105, 93, 56, 106], [69, 98, 70, 122]]
+Epoch 0: |                                                                                                                                                     | 2089/? [2:13:22<00:00,  0.26it/s, v_num=pmyy]train step 2090; scene = [['0565bd311bf73bbb'], ['eceeb7a49f302da9'], ['2073f379b98e47e8'], ['c5536f755d325407'], ['06ed257e33ae67f5'], ['b12f64d6002ec745']]; loss = 0.061526
+Epoch 0: |                                                                                                                                                     | 2090/? [2:13:26<00:00,  0.26it/s, v_num=pmyy]context = [[14, 20, 22, 23, 24, 26, 36, 41, 74, 81, 83, 87], [2, 10, 11, 16, 25, 26, 34, 35, 41, 49, 62, 63]]target = [[16, 23, 39, 54, 53, 59, 48, 70, 20, 63, 50, 36], [60, 32, 15, 7, 5, 51, 40, 52, 10, 56, 22, 41]]
+Epoch 0: |                                                                                                                                                     | 2099/? [2:14:00<00:00,  0.26it/s, v_num=pmyy]train step 2100; scene = [['3101309753e2f063'], ['d55fbb3dcd24f08e']]; loss = 0.035022
+Epoch 0: |                                                                                                                                                     | 2100/? [2:14:04<00:00,  0.26it/s, v_num=pmyy]context = [[145, 173, 174, 182, 187, 197, 204, 211], [65, 79, 86, 88, 93, 112, 118, 122], [98, 107, 122, 145, 156, 159, 162, 163]]target = [[208, 167, 187, 146, 173, 180, 166, 172], [108, 72, 116, 104, 92, 115, 83, 68], [160, 128, 158, 103, 134, 125, 101, 109]]
+Epoch 0: |                                                                                                                                                     | 2109/? [2:14:38<00:00,  0.26it/s, v_num=pmyy]train step 2110; scene = [['4fd151a48542df52'], ['c53e0b350f04f159'], ['ec1dcf652eae675d'], ['f001500643d191d6']]; loss = 0.110492
+Epoch 0: |                                                                                                                                                     | 2110/? [2:14:42<00:00,  0.26it/s, v_num=pmyy]context = [[42, 45, 48, 60, 81, 84, 88, 90, 102, 104, 121, 130], [22, 29, 34, 38, 39, 44, 49, 50, 53, 56, 59, 74]]target = [[108, 62, 127, 115, 43, 114, 52, 51, 100, 126, 61, 87], [64, 41, 45, 69, 25, 27, 65, 43, 68, 71, 38, 35]]
+Epoch 0: |                                                                                                                                                     | 2119/? [2:15:17<00:00,  0.26it/s, v_num=pmyy]train step 2120; scene = [['4f515c197a061a67'], ['0c7171edef36d44d'], ['ec5f9801aa83c8aa']]; loss = 0.058942
+Epoch 0: |                                                                                                                                                     | 2120/? [2:15:20<00:00,  0.26it/s, v_num=pmyy]context = [[124, 126, 135, 171], [32, 69, 83, 84], [30, 85, 88, 105], [27, 33, 42, 117], [2, 33, 53, 56], [24, 47, 86, 102]]target = [[165, 167, 132, 145], [80, 75, 52, 61], [73, 92, 100, 67], [87, 74, 112, 34], [45, 41, 49, 35], [80, 73, 95, 30]]
+Epoch 0: |                                                                                                                                                     | 2129/? [2:15:54<00:00,  0.26it/s, v_num=pmyy]train step 2130; scene = [['2dd8ae2b71457753'], ['3670e4d9d26e7534']]; loss = 0.044133
+Epoch 0: |                                                                                                                                                     | 2130/? [2:15:57<00:00,  0.26it/s, v_num=pmyy]context = [[24, 25, 27, 28, 32, 34, 36, 39, 48, 50, 55, 61, 62, 63, 64, 67, 77, 87, 102, 106, 110, 116, 117, 121]]target = [[45, 42, 59, 107, 40, 101, 31, 75, 70, 54, 81, 53, 83, 73, 64, 100, 51, 37, 47, 117, 60, 43, 48, 90]]
+Epoch 0: |                                                                                                                                                     | 2139/? [2:16:32<00:00,  0.26it/s, v_num=pmyy]train step 2140; scene = [['158ecf8d21e5af57'], ['4dfd4268f80ef274'], ['04c4eef824be2a53']]; loss = 0.042263
+Epoch 0: |                                                                                                                                                     | 2140/? [2:16:36<00:00,  0.26it/s, v_num=pmyy]context = [[131, 152, 157, 167, 188, 191], [139, 146, 149, 153, 171, 223], [126, 136, 156, 187, 188, 202], [32, 37, 41, 56, 82, 87]]target = [[189, 157, 139, 148, 184, 180], [151, 183, 209, 179, 185, 140], [159, 166, 155, 170, 182, 131], [78, 79, 34, 85, 50, 59]]
+Epoch 0: |                                                                                                                                                     | 2149/? [2:17:11<00:00,  0.26it/s, v_num=pmyy]train step 2150; scene = [['db62795c284bf764'], ['916b86f95631b480']]; loss = 0.052812
+Epoch 0: |                                                                                                                                                     | 2150/? [2:17:15<00:00,  0.26it/s, v_num=pmyy]context = [[4, 6, 16, 19, 21, 30, 35, 38, 41, 48, 51, 52, 54, 62, 63, 67, 69, 77, 80, 81, 86, 93, 95, 101]]target = [[56, 16, 36, 50, 25, 34, 65, 87, 40, 5, 42, 54, 18, 32, 68, 76, 98, 52, 37, 85, 41, 79, 13, 91]]
+Epoch 0: |                                                                                                                                                     | 2159/? [2:17:49<00:00,  0.26it/s, v_num=pmyy]train step 2160; scene = [['72077f7ff39fc73e'], ['9ae93c878c7dbe8a'], ['4bac79c3a17ed149'], ['208bd0af6bc8937b'], ['84e6b90f0c2567d8'], ['f4ba6d204cb14df7'], ['758770f2884e9a79'], ['36df585860d0ad88'], ['69a6e3951e138ca8'], ['5d986af113fbac56'], ['d46599d6e4a2b451'], ['ee020a8773034321']]; loss = 0.061699
+Epoch 0: |                                                                                                                                                     | 2160/? [2:17:53<00:00,  0.26it/s, v_num=pmyy]context = [[6, 8, 10, 50, 66, 87], [50, 69, 72, 95, 105, 111], [126, 165, 170, 172, 176, 211], [5, 9, 19, 42, 47, 50]]target = [[9, 68, 72, 19, 24, 67], [80, 85, 64, 70, 104, 91], [184, 172, 136, 189, 179, 175], [48, 14, 45, 22, 40, 16]]
+Epoch 0: |                                                                                                                                                     | 2169/? [2:18:28<00:00,  0.26it/s, v_num=pmyy]train step 2170; scene = [['e1d9afaa8899ee32'], ['1060ca933281b55c'], ['fbf253fca4a29e87'], ['5ff87250a0eb913b'], ['d02e6b104723d39a'], ['42b208082fce3bc2'], ['ee6e5709a57be759'], ['a191c34eb75bbaec'], ['0cbafecfcb0f7727'], ['e6461cee8a9474d5'], ['fda8bac8ddac590f'], ['465fa8314b741006']]; loss = 0.103594
+Epoch 0: |                                                                                                                                                     | 2170/? [2:18:31<00:00,  0.26it/s, v_num=pmyy]context = [[5, 20, 23, 37, 49, 52, 59, 76], [11, 28, 31, 35, 41, 43, 44, 57], [5, 15, 30, 36, 47, 59, 61, 62]]target = [[53, 50, 40, 18, 39, 33, 23, 17], [27, 38, 40, 28, 55, 45, 26, 25], [56, 9, 61, 23, 59, 30, 45, 35]]
+Epoch 0: |                                                                                                                                                     | 2179/? [2:19:05<00:00,  0.26it/s, v_num=pmyy]train step 2180; scene = [['91599919681fac69'], ['f9cece9ebde532d0']]; loss = 0.054370
+Epoch 0: |                                                                                                                                                     | 2180/? [2:19:09<00:00,  0.26it/s, v_num=pmyy]context = [[46, 63, 73, 75, 76, 83, 99, 102], [17, 31, 32, 39, 47, 52, 58, 68], [83, 91, 110, 113, 116, 117, 121, 133]]target = [[72, 92, 67, 101, 68, 55, 49, 74], [30, 31, 53, 26, 41, 35, 20, 62], [93, 131, 116, 96, 125, 107, 127, 109]]
+Epoch 0: |                                                                                                                                                     | 2189/? [2:19:44<00:00,  0.26it/s, v_num=pmyy]train step 2190; scene = [['4d48befa72535f0a']]; loss = 0.042860
+Epoch 0: |                                                                                                                                                     | 2190/? [2:19:48<00:00,  0.26it/s, v_num=pmyy]context = [[113, 123, 125, 126, 132, 133, 135, 138, 149, 150, 154, 160, 167, 172, 175, 191, 194, 197, 198, 200, 201, 205, 206, 210]]target = [[120, 130, 147, 187, 142, 115, 202, 171, 117, 163, 191, 170, 139, 169, 181, 155, 150, 141, 192, 146, 188, 174, 205, 179]]
+Epoch 0: |                                                                                                                                                     | 2199/? [2:20:21<00:00,  0.26it/s, v_num=pmyy]train step 2200; scene = [['76ce9bd95ed81200'], ['91ac7d7027dcc46d'], ['3c3e1619744887ca'], ['003d2563b3c1023e']]; loss = 0.050293
+Epoch 0: |                                                                                                                                                     | 2200/? [2:20:25<00:00,  0.26it/s, v_num=pmyy]context = [[11, 13, 17, 19, 42, 43, 50, 51, 58, 60, 61, 63, 65, 71, 74, 77, 82, 85, 89, 95, 99, 101, 103, 108]]target = [[61, 31, 48, 27, 88, 78, 23, 53, 24, 93, 63, 73, 20, 32, 55, 35, 29, 75, 21, 76, 49, 42, 28, 40]]
+[2026-02-25 00:48:21,267][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2209/? [2:21:06<00:00,  0.26it/s, v_num=pmyy]train step 2210; scene = [['40873337bb9786ab'], ['4f354a704f70f53b'], ['57cea8a71cab48cf'], ['19c383ad9f21d0d3'], ['a98ec507d448dea2'], ['b663a474fd1e4ca4']]; loss = 0.074518
+Epoch 0: |                                                                                                                                                     | 2210/? [2:21:09<00:00,  0.26it/s, v_num=pmyy]context = [[9, 12, 24, 32, 37, 38, 44, 46, 51, 65, 71, 75], [87, 89, 90, 92, 104, 114, 118, 119, 120, 122, 143, 144]]target = [[61, 31, 40, 37, 52, 13, 14, 68, 57, 25, 70, 54], [95, 130, 128, 134, 91, 114, 133, 102, 129, 98, 138, 116]]
+Epoch 0: |                                                                                                                                                     | 2219/? [2:21:43<00:00,  0.26it/s, v_num=pmyy]train step 2220; scene = [['d85f0322f215aa54'], ['1318e656455d56d2'], ['e5f41c9021b0b7c1']]; loss = 0.051681
+Epoch 0: |                                                                                                                                                     | 2220/? [2:21:47<00:00,  0.26it/s, v_num=pmyy]context = [[12, 17, 27, 29, 36, 37, 44, 50, 52, 54, 55, 65, 75, 80, 81, 83, 84, 85, 86, 89, 91, 94, 95, 109]]target = [[83, 45, 76, 92, 104, 41, 30, 59, 79, 28, 86, 38, 85, 64, 66, 51, 106, 15, 89, 16, 46, 108, 99, 33]]
+Epoch 0: |                                                                                                                                                     | 2229/? [2:22:22<00:00,  0.26it/s, v_num=pmyy]train step 2230; scene = [['9af18a1c4c45a179']]; loss = 0.048140
+Epoch 0: |                                                                                                                                                     | 2230/? [2:22:26<00:00,  0.26it/s, v_num=pmyy]context = [[40, 43, 44, 61, 67, 70, 74, 77, 79, 85, 87, 92, 93, 98, 100, 103, 108, 115, 117, 123, 128, 134, 135, 137]]target = [[119, 88, 102, 62, 79, 131, 53, 122, 46, 128, 130, 59, 54, 76, 71, 57, 107, 126, 93, 65, 72, 115, 69, 66]]
+Epoch 0: |                                                                                                                                                     | 2239/? [2:23:00<00:00,  0.26it/s, v_num=pmyy]train step 2240; scene = [['e454026f5348630e'], ['d6a1f3e13c45df99'], ['4dd9c5fab7e6ec75'], ['352d2bdc1900b5e0']]; loss = 0.037614
+Epoch 0: |                                                                                                                                                     | 2240/? [2:23:03<00:00,  0.26it/s, v_num=pmyy]context = [[5, 53], [14, 60], [21, 91], [60, 113], [2, 85], [31, 99], [27, 81], [3, 54], [15, 101], [11, 69], [0, 45], [193, 272]]target = [[39, 37], [57, 23], [71, 41], [71, 104], [30, 23], [39, 89], [58, 57], [31, 40], [83, 33], [61, 53], [28, 1], [235, 232]]
+Epoch 0: |                                                                                                                                                     | 2249/? [2:23:38<00:00,  0.26it/s, v_num=pmyy]train step 2250; scene = [['5afa3097bb38b159'], ['c0f67af5cd34e8d8']]; loss = 0.046574
+Epoch 0: |                                                                                                                                                     | 2250/? [2:23:42<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 2250; scene = ['651a7f83ed093001'];
+target intrinsic:  tensor(0.8796, device='cuda:0') tensor(0.8798, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9091, device='cuda:0') tensor(0.9109, device='cuda:0')
+[2026-02-25 00:51:35,176][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2250/? [2:23:43<00:00,  0.26it/s, v_num=pmyy]context = [[0, 21, 34, 39, 59, 63, 83, 84], [23, 26, 33, 34, 53, 58, 59, 68], [118, 137, 141, 150, 154, 170, 188, 192]]target = [[14, 79, 44, 19, 83, 18, 29, 23], [48, 24, 60, 52, 35, 61, 29, 64], [136, 187, 135, 171, 137, 131, 147, 125]]
+Epoch 0: |                                                                                                                                                     | 2259/? [2:24:16<00:00,  0.26it/s, v_num=pmyy]train step 2260; scene = [['39cafa06c9c431dc']]; loss = 0.056422
+Epoch 0: |                                                                                                                                                     | 2260/? [2:24:20<00:00,  0.26it/s, v_num=pmyy]context = [[46, 51, 80, 81, 83, 97, 105, 114], [2, 9, 10, 12, 28, 29, 51, 70], [38, 41, 42, 46, 69, 72, 118, 124]]target = [[60, 58, 103, 89, 80, 51, 99, 109], [25, 68, 19, 67, 43, 14, 52, 15], [91, 55, 98, 96, 74, 116, 83, 113]]
+Epoch 0: |                                                                                                                                                     | 2269/? [2:24:55<00:00,  0.26it/s, v_num=pmyy]train step 2270; scene = [['10a2d0db5c8c0962']]; loss = 0.042458
+Epoch 0: |                                                                                                                                                     | 2270/? [2:24:58<00:00,  0.26it/s, v_num=pmyy]context = [[33, 39, 52, 59, 85, 90, 93, 112], [16, 18, 22, 27, 40, 46, 55, 70], [2, 32, 40, 41, 42, 46, 51, 71]]target = [[40, 55, 82, 107, 111, 35, 102, 36], [25, 31, 19, 58, 55, 64, 62, 28], [48, 13, 68, 46, 30, 12, 42, 40]]
+Epoch 0: |                                                                                                                                                     | 2279/? [2:25:33<00:00,  0.26it/s, v_num=pmyy]train step 2280; scene = [['bcbd422c5506815b'], ['dd1a25142bac8d29'], ['ba3e38a071a523f0'], ['2248f7eaf18cca06'], ['d0edef000b9ca743'], ['d9aacea21e110a52']]; loss = 0.059979
+Epoch 0: |                                                                                                                                                     | 2280/? [2:25:37<00:00,  0.26it/s, v_num=pmyy]context = [[103, 107, 133, 140, 142, 146, 151, 159, 168, 169, 172, 177], [2, 4, 9, 11, 12, 17, 28, 38, 40, 46, 56, 59]]target = [[132, 133, 104, 118, 122, 149, 173, 175, 125, 120, 168, 164], [39, 49, 57, 47, 31, 13, 28, 52, 21, 16, 55, 29]]
+Epoch 0: |                                                                                                                                                     | 2289/? [2:26:12<00:00,  0.26it/s, v_num=pmyy]train step 2290; scene = [['fa6cb78e1503d22c'], ['f592013dce557cce']]; loss = 0.051641
+Epoch 0: |                                                                                                                                                     | 2290/? [2:26:15<00:00,  0.26it/s, v_num=pmyy]context = [[30, 35, 41, 49, 50, 59, 65, 78, 89, 91, 113, 116], [126, 127, 128, 134, 152, 160, 162, 166, 171, 172, 176, 177]]target = [[51, 61, 81, 100, 48, 110, 54, 112, 60, 113, 86, 55], [140, 169, 144, 127, 174, 150, 129, 143, 163, 132, 173, 152]]
+Epoch 0: |                                                                                                                                                     | 2299/? [2:26:49<00:00,  0.26it/s, v_num=pmyy]train step 2300; scene = [['e23f945703f61bfc']]; loss = 0.050785
+Epoch 0: |                                                                                                                                                     | 2300/? [2:26:53<00:00,  0.26it/s, v_num=pmyy]context = [[73, 94, 106, 123, 128, 130], [6, 22, 33, 42, 45, 51], [165, 166, 167, 196, 205, 228], [98, 109, 112, 134, 135, 152]]target = [[128, 88, 106, 120, 129, 82], [23, 48, 38, 26, 11, 14], [208, 193, 188, 218, 183, 210], [134, 143, 122, 104, 117, 137]]
+Epoch 0: |                                                                                                                                                     | 2309/? [2:27:27<00:00,  0.26it/s, v_num=pmyy]train step 2310; scene = [['347f4e7f9f627a12'], ['d1d992e581136ac6'], ['59e7800dec3fa9f5'], ['cf6a4349d0ffdfcf']]; loss = 0.041969
+Epoch 0: |                                                                                                                                                     | 2310/? [2:27:31<00:00,  0.26it/s, v_num=pmyy]context = [[0, 5, 8, 13, 15, 43, 47, 51, 61, 68, 72, 74], [8, 13, 17, 27, 29, 38, 49, 53, 56, 60, 63, 65]]target = [[5, 42, 38, 57, 31, 53, 36, 35, 59, 55, 2, 11], [23, 48, 28, 17, 61, 47, 12, 64, 36, 9, 30, 57]]
+Epoch 0: |                                                                                                                                                     | 2319/? [2:28:05<00:00,  0.26it/s, v_num=pmyy]train step 2320; scene = [['60c37b519a01205d'], ['b0ee123a8cfc8e62'], ['c5b4562390525d10'], ['8b950107c02ffaa9']]; loss = 0.066945
+Epoch 0: |                                                                                                                                                     | 2320/? [2:28:09<00:00,  0.26it/s, v_num=pmyy]context = [[8, 11, 13, 15, 17, 25, 29, 39, 40, 45, 48, 58], [4, 18, 21, 36, 47, 50, 60, 63, 71, 76, 79, 82]]target = [[56, 37, 21, 47, 26, 51, 36, 12, 49, 27, 24, 16], [62, 43, 80, 40, 30, 39, 37, 19, 6, 13, 35, 10]]
+Epoch 0: |                                                                                                                                                     | 2329/? [2:28:41<00:00,  0.26it/s, v_num=pmyy]train step 2330; scene = [['5c9b898102b16eae'], ['583ab14553881ee8'], ['e204af0947f704ad'], ['39edfd183c4f5b5b'], ['a85c79f15c396d71'], ['3d7a200dab472990']]; loss = 0.053345
+Epoch 0: |                                                                                                                                                     | 2330/? [2:28:45<00:00,  0.26it/s, v_num=pmyy]context = [[9, 23, 91], [5, 23, 61], [94, 164, 165], [149, 165, 237], [73, 137, 159], [1, 13, 48], [5, 55, 59], [25, 55, 75]]target = [[33, 46, 62], [42, 29, 22], [125, 112, 119], [187, 202, 170], [139, 87, 131], [7, 2, 23], [57, 39, 50], [62, 34, 56]]
+Epoch 0: |                                                                                                                                                     | 2339/? [2:29:20<00:00,  0.26it/s, v_num=pmyy]train step 2340; scene = [['c4005922f59686ae']]; loss = 0.043539
+Epoch 0: |                                                                                                                                                     | 2340/? [2:29:24<00:00,  0.26it/s, v_num=pmyy]context = [[9, 13, 14, 15, 29, 42, 47, 48, 52, 54, 58, 60, 61, 68, 76, 77, 81, 88, 89, 97, 98, 99, 105, 106]]target = [[72, 73, 28, 65, 82, 67, 96, 56, 75, 94, 16, 101, 63, 80, 83, 33, 41, 61, 93, 100, 14, 86, 32, 11]]
+Epoch 0: |                                                                                                                                                     | 2349/? [2:29:58<00:00,  0.26it/s, v_num=pmyy]train step 2350; scene = [['009e573e59c8c393'], ['7c48a30f23ea42c3'], ['739123afdcc19a64'], ['ddc11c891f471dd0'], ['38b7b864ae7b21e9'], ['997656bdb430ad43']]; loss = 0.044650
+Epoch 0: |                                                                                                                                                     | 2350/? [2:30:02<00:00,  0.26it/s, v_num=pmyy]context = [[115, 120, 127, 130, 134, 143, 147, 148, 149, 150, 159, 163, 173, 175, 184, 186, 187, 188, 190, 195, 199, 202, 205, 212]]target = [[148, 147, 199, 184, 190, 152, 134, 189, 127, 120, 201, 163, 186, 165, 171, 193, 117, 210, 183, 211, 195, 156, 204, 145]]
+Epoch 0: |                                                                                                                                                     | 2359/? [2:30:36<00:00,  0.26it/s, v_num=pmyy]train step 2360; scene = [['9a447c89080e9b56'], ['9eb8e7f262b10c23'], ['c2b8b3e74c64553a'], ['b6f9cfe435a0fde7']]; loss = 0.053840
+Epoch 0: |                                                                                                                                                     | 2360/? [2:30:39<00:00,  0.26it/s, v_num=pmyy]context = [[1, 4, 8, 16, 17, 28, 40, 45, 51, 53, 60, 62], [1, 4, 5, 11, 12, 13, 14, 15, 17, 29, 40, 82]]target = [[8, 52, 12, 39, 55, 43, 33, 16, 53, 18, 25, 2], [71, 5, 21, 6, 62, 55, 45, 49, 37, 14, 27, 56]]
+Epoch 0: |                                                                                                                                                     | 2369/? [2:31:14<00:00,  0.26it/s, v_num=pmyy]train step 2370; scene = [['090ced9a667843cb']]; loss = 0.033160
+Epoch 0: |                                                                                                                                                     | 2370/? [2:31:18<00:00,  0.26it/s, v_num=pmyy]context = [[13, 19, 35, 37, 43, 61, 92, 94], [0, 8, 19, 21, 24, 25, 27, 45], [1, 5, 38, 42, 45, 47, 75, 82]]target = [[46, 77, 90, 47, 83, 21, 76, 54], [3, 11, 9, 33, 20, 41, 42, 4], [72, 28, 74, 38, 52, 40, 35, 34]]
+Epoch 0: |                                                                                                                                                     | 2379/? [2:31:52<00:00,  0.26it/s, v_num=pmyy]train step 2380; scene = [['8ef9ff3189c85eee'], ['7e8630a890a85545']]; loss = 0.044951
+Epoch 0: |                                                                                                                                                     | 2380/? [2:31:56<00:00,  0.26it/s, v_num=pmyy]context = [[48, 55, 58, 63, 67, 68, 71, 73, 75, 76, 79, 80, 86, 93, 107, 108, 109, 118, 119, 120, 129, 133, 135, 145]]target = [[133, 59, 121, 87, 96, 110, 67, 101, 49, 142, 60, 134, 109, 132, 98, 122, 103, 100, 80, 57, 65, 86, 128, 143]]
+Epoch 0: |                                                                                                                                                     | 2389/? [2:32:31<00:00,  0.26it/s, v_num=pmyy]train step 2390; scene = [['343a98bd8cfda2de'], ['c09d7898ef37ba32']]; loss = 0.045056
+Epoch 0: |                                                                                                                                                     | 2390/? [2:32:35<00:00,  0.26it/s, v_num=pmyy]context = [[19, 20, 31, 32, 41, 46, 49, 50, 51, 53, 59, 60, 62, 63, 74, 77, 82, 84, 85, 91, 93, 97, 98, 116]]target = [[56, 31, 77, 102, 39, 54, 21, 57, 66, 75, 100, 115, 26, 53, 34, 113, 97, 90, 25, 82, 108, 60, 67, 38]]
+Epoch 0: |                                                                                                                                                     | 2399/? [2:33:09<00:00,  0.26it/s, v_num=pmyy]train step 2400; scene = [['8d758914077e5926']]; loss = 0.044172
+Epoch 0: |                                                                                                                                                     | 2400/? [2:33:13<00:00,  0.26it/s, v_num=pmyy]context = [[114, 117, 119, 120, 131, 136, 141, 148, 149, 151, 163, 168], [44, 47, 52, 76, 77, 78, 85, 88, 89, 90, 97, 99]]target = [[141, 153, 135, 128, 126, 119, 140, 147, 161, 120, 155, 127], [82, 75, 49, 62, 84, 54, 78, 70, 55, 72, 96, 92]]
+[2026-02-25 01:01:09,352][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2409/? [2:33:48<00:00,  0.26it/s, v_num=pmyy]train step 2410; scene = [['d35650707515c2cf']]; loss = 0.046008
+Epoch 0: |                                                                                                                                                     | 2410/? [2:33:52<00:00,  0.26it/s, v_num=pmyy]context = [[4, 8, 16, 30, 45, 53, 70, 77], [123, 139, 144, 149, 166, 184, 187, 207], [38, 44, 51, 66, 68, 90, 94, 98]]target = [[7, 17, 15, 70, 47, 48, 35, 72], [190, 168, 157, 135, 200, 160, 161, 128], [84, 75, 51, 67, 64, 46, 95, 49]]
+Epoch 0: |                                                                                                                                                     | 2419/? [2:34:27<00:00,  0.26it/s, v_num=pmyy]train step 2420; scene = [['ea5def9e91076788']]; loss = 0.033486
+Epoch 0: |                                                                                                                                                     | 2420/? [2:34:31<00:00,  0.26it/s, v_num=pmyy]context = [[93, 107, 110, 121, 127, 148], [165, 169, 170, 208, 209, 214], [0, 6, 7, 23, 45, 83], [12, 33, 38, 45, 81, 91]]target = [[131, 99, 135, 127, 142, 120], [167, 177, 176, 213, 171, 191], [22, 45, 64, 29, 67, 51], [47, 31, 65, 87, 60, 85]]
+Epoch 0: |                                                                                                                                                     | 2429/? [2:35:05<00:00,  0.26it/s, v_num=pmyy]train step 2430; scene = [['45fe4bbb2526ca6e'], ['0db2c3f2775880de'], ['4a0f95a3db913b56']]; loss = 0.028192
+Epoch 0: |                                                                                                                                                     | 2430/? [2:35:08<00:00,  0.26it/s, v_num=pmyy]context = [[22, 66, 68, 80, 92, 99], [26, 44, 52, 76, 80, 86], [3, 38, 40, 69, 77, 89], [43, 83, 90, 109, 121, 126]]target = [[93, 85, 61, 37, 46, 63], [48, 57, 64, 35, 27, 84], [43, 41, 48, 23, 25, 67], [111, 55, 61, 105, 114, 116]]
+Epoch 0: |                                                                                                                                                     | 2439/? [2:35:43<00:00,  0.26it/s, v_num=pmyy]train step 2440; scene = [['512f1df8266be984'], ['50c8a233bd82a613']]; loss = 0.054314
+Epoch 0: |                                                                                                                                                     | 2440/? [2:35:47<00:00,  0.26it/s, v_num=pmyy]context = [[96, 106, 114, 124, 131, 158], [48, 54, 56, 57, 87, 102], [2, 8, 16, 43, 50, 59], [22, 27, 37, 53, 69, 77]]target = [[112, 148, 103, 120, 104, 132], [72, 96, 86, 90, 52, 98], [46, 13, 56, 3, 57, 55], [58, 57, 71, 67, 28, 45]]
+Epoch 0: |                                                                                                                                                     | 2449/? [2:36:20<00:00,  0.26it/s, v_num=pmyy]train step 2450; scene = [['ac584e6fc77676a4'], ['19a3235d680e11c4'], ['61b9def6cdf00024'], ['fc86266e2fcb72fd']]; loss = 0.041586
+Epoch 0: |                                                                                                                                                     | 2450/? [2:36:24<00:00,  0.26it/s, v_num=pmyy]context = [[16, 41, 71], [43, 46, 97], [10, 18, 90], [3, 16, 53], [143, 187, 198], [25, 36, 70], [98, 136, 179], [19, 65, 78]]target = [[45, 18, 43], [95, 96, 51], [32, 21, 70], [5, 8, 27], [146, 154, 151], [28, 31, 48], [133, 137, 151], [33, 53, 64]]
+Epoch 0: |                                                                                                                                                     | 2459/? [2:36:58<00:00,  0.26it/s, v_num=pmyy]train step 2460; scene = [['dc5f38f005c3ebd6']]; loss = 0.066451
+Epoch 0: |                                                                                                                                                     | 2460/? [2:37:02<00:00,  0.26it/s, v_num=pmyy]context = [[4, 5, 11, 12, 16, 17, 19, 22, 23, 30, 39, 44, 49, 50, 51, 61, 77, 82, 87, 89, 95, 96, 97, 101]]target = [[53, 36, 13, 57, 86, 80, 97, 72, 84, 27, 61, 94, 35, 51, 54, 67, 43, 66, 62, 89, 7, 92, 10, 47]]
+Epoch 0: |                                                                                                                                                     | 2469/? [2:37:37<00:00,  0.26it/s, v_num=pmyy]train step 2470; scene = [['c7620995ebe9c4e6'], ['06241ebed1658f34'], ['1f214117250f089a'], ['5e31e0691d426537'], ['11c4a7d67bc2629e'], ['715e8695976cdb61'], ['6888f7ca14081419'], ['8fe341dcd0880bd5']]; loss = 0.065372
+Epoch 0: |                                                                                                                                                     | 2470/? [2:37:40<00:00,  0.26it/s, v_num=pmyy]context = [[17, 28, 38, 42, 68, 86], [2, 4, 28, 37, 51, 58], [7, 21, 24, 54, 64, 70], [41, 42, 61, 71, 81, 89]]target = [[55, 30, 52, 28, 44, 76], [22, 36, 3, 57, 34, 9], [16, 35, 34, 61, 56, 43], [56, 87, 85, 73, 68, 50]]
+Epoch 0: |                                                                                                                                                     | 2479/? [2:38:14<00:00,  0.26it/s, v_num=pmyy]train step 2480; scene = [['cf98d3219d144500']]; loss = 0.069438
+Epoch 0: |                                                                                                                                                     | 2480/? [2:38:18<00:00,  0.26it/s, v_num=pmyy]context = [[151, 158, 171, 173, 176, 205, 206, 230], [74, 88, 94, 98, 119, 130, 146, 163], [5, 8, 33, 42, 52, 53, 57, 59]]target = [[177, 223, 212, 220, 205, 207, 225, 222], [102, 132, 115, 112, 145, 156, 119, 100], [12, 23, 49, 20, 55, 32, 10, 39]]
+Epoch 0: |                                                                                                                                                     | 2489/? [2:38:52<00:00,  0.26it/s, v_num=pmyy]train step 2490; scene = [['ebff5d05f1bb086f'], ['880427ff150f7b4d'], ['9dd0efd4b4626604'], ['f52c70025f40e56d']]; loss = 0.051605
+Epoch 0: |                                                                                                                                                     | 2490/? [2:38:56<00:00,  0.26it/s, v_num=pmyy]context = [[19, 27, 31, 35, 36, 42, 57, 60, 63, 64, 65, 70], [21, 26, 31, 32, 50, 56, 62, 67, 74, 81, 90, 93]]target = [[51, 33, 64, 68, 53, 29, 54, 27, 26, 69, 35, 48], [59, 60, 83, 76, 86, 68, 61, 36, 71, 22, 69, 82]]
+Epoch 0: |                                                                                                                                                     | 2499/? [2:39:29<00:00,  0.26it/s, v_num=pmyy]train step 2500; scene = [['7dc5c394263df267']]; loss = 0.062322
+Epoch 0: |                                                                                                                                                     | 2500/? [2:39:32<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 2500; scene = ['97ef4323919c5e8a'];
+target intrinsic:  tensor(0.8889, device='cuda:0') tensor(0.8892, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9352, device='cuda:0') tensor(0.9300, device='cuda:0')
+[2026-02-25 01:07:25,490][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2500/? [2:39:33<00:00,  0.26it/s, v_num=pmyy]context = [[17, 21, 30, 44, 52, 84], [24, 34, 39, 45, 53, 84], [159, 163, 179, 192, 194, 207], [115, 116, 136, 154, 172, 188]]target = [[31, 20, 44, 37, 73, 59], [47, 48, 59, 36, 52, 44], [198, 165, 170, 193, 188, 175], [173, 128, 152, 148, 158, 172]]
+Epoch 0: |                                                                                                                                                     | 2509/? [2:40:07<00:00,  0.26it/s, v_num=pmyy]train step 2510; scene = [['e806eb496df1dfa0'], ['3caa89905a59e7e6'], ['3fc603b4c4531c11'], ['08da23838ee6e23b'], ['89a0f0553f7f2bdb'], ['5080347450bbcb08']]; loss = 0.047784
+Epoch 0: |                                                                                                                                                     | 2510/? [2:40:11<00:00,  0.26it/s, v_num=pmyy]context = [[7, 16, 17, 33, 34, 42, 43, 45, 48, 49, 52, 60], [16, 27, 29, 31, 38, 42, 58, 64, 71, 89, 93, 104]]target = [[59, 11, 58, 34, 22, 51, 55, 46, 9, 12, 31, 56], [90, 95, 100, 19, 82, 44, 96, 91, 97, 28, 88, 77]]
+Epoch 0: |                                                                                                                                                     | 2519/? [2:40:46<00:00,  0.26it/s, v_num=pmyy]train step 2520; scene = [['0a94493d6f4cd4be'], ['1c3a8fcf1547dcf7'], ['c34950273c4d538b'], ['397652e91be52496'], ['d9a133a4747493d2'], ['46dc3900c302593a'], ['e281213868014707'], ['83c959de5ae5b86c']]; loss = 0.068255
+Epoch 0: |                                                                                                                                                     | 2520/? [2:40:49<00:00,  0.26it/s, v_num=pmyy]context = [[22, 26, 34, 62, 68, 89, 103, 111], [16, 23, 45, 57, 69, 77, 101, 105], [9, 21, 23, 24, 56, 60, 66, 68]]target = [[29, 88, 85, 106, 103, 79, 65, 23], [61, 100, 46, 104, 23, 66, 18, 40], [25, 38, 15, 16, 64, 45, 34, 30]]
+Epoch 0: |                                                                                                                                                     | 2529/? [2:41:24<00:00,  0.26it/s, v_num=pmyy]train step 2530; scene = [['871ec8d951a81c62']]; loss = 0.050936
+Epoch 0: |                                                                                                                                                     | 2530/? [2:41:28<00:00,  0.26it/s, v_num=pmyy]context = [[39, 41, 46, 48, 52, 53, 55, 66, 68, 70, 83, 92, 93, 97, 103, 106, 110, 111, 118, 124, 125, 131, 133, 136]]target = [[125, 130, 53, 128, 124, 75, 61, 112, 64, 70, 99, 120, 57, 47, 69, 60, 43, 85, 79, 63, 45, 135, 104, 129]]
+Epoch 0: |                                                                                                                                                     | 2539/? [2:42:00<00:00,  0.26it/s, v_num=pmyy]train step 2540; scene = [['8e78ce55e3180547']]; loss = 0.039403
+Epoch 0: |                                                                                                                                                     | 2540/? [2:42:04<00:00,  0.26it/s, v_num=pmyy]context = [[19, 22, 26, 27, 29, 35, 40, 42, 54, 59, 60, 61, 64, 74, 77, 86, 92, 102, 106, 109, 110, 111, 115, 116]]target = [[89, 69, 46, 77, 65, 57, 110, 49, 70, 97, 38, 58, 31, 40, 101, 74, 20, 88, 92, 90, 80, 28, 94, 37]]
+Epoch 0: |                                                                                                                                                     | 2549/? [2:42:38<00:00,  0.26it/s, v_num=pmyy]train step 2550; scene = [['6dfaec91f745fdd9'], ['3f72021b93b6224a'], ['36be66d194d57ec8']]; loss = 0.091536
+Epoch 0: |                                                                                                                                                     | 2550/? [2:42:41<00:00,  0.26it/s, v_num=pmyy]context = [[47, 51, 57, 66, 73, 85, 93, 96, 97, 107, 113, 121], [26, 41, 43, 44, 54, 64, 69, 75, 78, 82, 85, 86]]target = [[83, 71, 53, 103, 110, 111, 55, 70, 94, 72, 74, 106], [32, 47, 85, 36, 49, 51, 41, 62, 35, 84, 31, 28]]
+Epoch 0: |                                                                                                                                                     | 2559/? [2:43:15<00:00,  0.26it/s, v_num=pmyy]train step 2560; scene = [['7c6d160c26de6887'], ['5ef6fe9ef5309457'], ['66de8729fe760dd8'], ['faba60084d22aa27']]; loss = 0.051185
+Epoch 0: |                                                                                                                                                     | 2560/? [2:43:19<00:00,  0.26it/s, v_num=pmyy]context = [[44, 89, 97], [157, 230, 235], [6, 8, 64], [19, 33, 76], [24, 39, 91], [14, 17, 65], [34, 59, 83], [19, 28, 77]]target = [[93, 77, 73], [195, 166, 202], [62, 61, 20], [61, 65, 30], [82, 90, 72], [32, 60, 21], [67, 74, 43], [39, 27, 37]]
+Epoch 0: |                                                                                                                                                     | 2569/? [2:43:53<00:00,  0.26it/s, v_num=pmyy]train step 2570; scene = [['678d8464781a3de2'], ['9ab12a31f2a3b9fb']]; loss = 0.043218
+Epoch 0: |                                                                                                                                                     | 2570/? [2:43:57<00:00,  0.26it/s, v_num=pmyy]context = [[41, 92], [2, 64], [133, 190], [7, 81], [6, 63], [33, 93], [65, 146], [18, 69], [15, 63], [13, 78], [4, 79], [66, 126]]target = [[90, 71], [12, 37], [175, 173], [56, 25], [7, 56], [61, 69], [132, 93], [28, 45], [17, 31], [21, 61], [69, 76], [120, 115]]
+Epoch 0: |                                                                                                                                                     | 2579/? [2:44:30<00:00,  0.26it/s, v_num=pmyy]train step 2580; scene = [['c40b23830f7437d5'], ['a0c7cabb66c795a4'], ['df63cd7eb6e92486']]; loss = 0.054028
+Epoch 0: |                                                                                                                                                     | 2580/? [2:44:33<00:00,  0.26it/s, v_num=pmyy]context = [[110, 115, 117, 128, 136, 137, 138, 141, 143, 145, 150, 168], [7, 10, 23, 30, 35, 48, 57, 59, 63, 64, 65, 69]]target = [[128, 149, 154, 126, 167, 132, 162, 151, 112, 134, 133, 131], [18, 25, 50, 30, 12, 19, 47, 39, 54, 35, 37, 56]]
+Epoch 0: |                                                                                                                                                     | 2589/? [2:45:08<00:00,  0.26it/s, v_num=pmyy]train step 2590; scene = [['65cdcbcb16f0ebe5'], ['4b12a530a5ab03d0']]; loss = 0.031062
+Epoch 0: |                                                                                                                                                     | 2590/? [2:45:12<00:00,  0.26it/s, v_num=pmyy]context = [[32, 42, 45, 59, 89, 100], [65, 89, 96, 104, 111, 139], [177, 197, 217, 235, 253, 264], [81, 94, 117, 131, 132, 142]]target = [[41, 75, 46, 34, 67, 60], [125, 90, 94, 114, 92, 95], [196, 223, 250, 229, 245, 259], [124, 91, 129, 94, 85, 131]]
+Epoch 0: |                                                                                                                                                     | 2599/? [2:45:47<00:00,  0.26it/s, v_num=pmyy]train step 2600; scene = [['cb3bac70297d52c0']]; loss = 0.030959
+Epoch 0: |                                                                                                                                                     | 2600/? [2:45:51<00:00,  0.26it/s, v_num=pmyy]context = [[3, 4, 7, 15, 37, 44, 46, 49], [71, 83, 85, 107, 113, 118, 123, 135], [4, 6, 26, 29, 32, 45, 51, 52]]target = [[38, 11, 27, 26, 16, 48, 39, 30], [81, 90, 94, 125, 75, 97, 110, 98], [22, 27, 29, 35, 7, 9, 49, 21]]
+[2026-02-25 01:13:46,903][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2609/? [2:46:24<00:00,  0.26it/s, v_num=pmyy]train step 2610; scene = [['262404ef111a096b'], ['5ac9dbde9cf64eed'], ['c32ee937dbec7221'], ['201e44ad015728c7'], ['1fb562b09fc361ea'], ['e454026f5348630e'], ['d1772c09b4b6d95f'], ['8e7cc5c956a4cf95']]; loss = 0.046926
+Epoch 0: |                                                                                                                                                     | 2610/? [2:46:28<00:00,  0.26it/s, v_num=pmyy]context = [[70, 73, 74, 76, 77, 79, 82, 88, 109, 110, 115, 124, 125, 136, 141, 148, 150, 151, 155, 158, 159, 160, 161, 167]]target = [[156, 121, 159, 150, 128, 138, 162, 81, 102, 71, 103, 74, 137, 144, 108, 166, 72, 161, 129, 110, 165, 136, 106, 99]]
+Epoch 0: |                                                                                                                                                     | 2619/? [2:47:02<00:00,  0.26it/s, v_num=pmyy]train step 2620; scene = [['4dd74d4d53abb812'], ['9029d8d8f6f0e98a'], ['360253c55aef692f'], ['b07a54eda9cdaccb'], ['ca553e70eff3aa6f'], ['70b45bbd1147fbf0']]; loss = 0.039066
+Epoch 0: |                                                                                                                                                     | 2620/? [2:47:05<00:00,  0.26it/s, v_num=pmyy]context = [[27, 55, 94], [5, 48, 83], [56, 94, 118], [121, 137, 181], [0, 61, 78], [36, 80, 108], [2, 18, 55], [8, 60, 69]]target = [[62, 80, 57], [21, 23, 34], [84, 71, 87], [139, 178, 159], [59, 19, 76], [104, 40, 51], [43, 20, 8], [47, 68, 60]]
+Epoch 0: |                                                                                                                                                     | 2629/? [2:47:40<00:00,  0.26it/s, v_num=pmyy]train step 2630; scene = [['c13076a68aeeb481'], ['0bafd80c9e2ae41b']]; loss = 0.050368
+Epoch 0: |                                                                                                                                                     | 2630/? [2:47:44<00:00,  0.26it/s, v_num=pmyy]context = [[135, 142, 145, 147, 149, 150, 158, 162, 163, 169, 172, 203], [22, 29, 55, 60, 62, 63, 64, 68, 69, 70, 71, 72]]target = [[190, 143, 159, 194, 198, 175, 197, 146, 166, 181, 174, 187], [47, 36, 59, 40, 28, 46, 65, 24, 29, 26, 69, 51]]
+Epoch 0: |                                                                                                                                                     | 2639/? [2:48:19<00:00,  0.26it/s, v_num=pmyy]train step 2640; scene = [['eae08bd3d5c892e7'], ['60acba38edb03894']]; loss = 0.026371
+Epoch 0: |                                                                                                                                                     | 2640/? [2:48:22<00:00,  0.26it/s, v_num=pmyy]context = [[166, 176, 251], [156, 173, 220], [69, 137, 148], [197, 204, 277], [28, 29, 100], [44, 111, 118], [0, 56, 61], [4, 68, 81]]target = [[170, 244, 200], [193, 173, 188], [75, 86, 116], [253, 268, 251], [49, 30, 48], [106, 54, 50], [24, 3, 29], [24, 58, 37]]
+Epoch 0: |                                                                                                                                                     | 2649/? [2:48:57<00:00,  0.26it/s, v_num=pmyy]train step 2650; scene = [['c520ae3405948a0a'], ['94340815d8708eea']]; loss = 0.026477
+Epoch 0: |                                                                                                                                                     | 2650/? [2:49:01<00:00,  0.26it/s, v_num=pmyy]context = [[6, 10, 15, 16, 20, 26, 31, 32, 37, 40, 42, 47, 55, 57, 66, 69, 70, 80, 82, 85, 88, 93, 101, 103]]target = [[72, 23, 99, 11, 66, 101, 81, 102, 30, 39, 54, 38, 17, 84, 69, 32, 22, 18, 26, 41, 79, 40, 57, 46]]
+Epoch 0: |                                                                                                                                                     | 2659/? [2:49:35<00:00,  0.26it/s, v_num=pmyy]train step 2660; scene = [['6666ae9375b1656e'], ['64619aa7bd88899d']]; loss = 0.069251
+Epoch 0: |                                                                                                                                                     | 2660/? [2:49:39<00:00,  0.26it/s, v_num=pmyy]context = [[13, 14, 27, 29, 31, 32, 39, 48, 50, 52, 54, 57, 58, 61, 67, 73, 79, 90, 93, 95, 105, 108, 109, 110]]target = [[83, 48, 82, 104, 25, 53, 89, 72, 93, 51, 41, 86, 24, 49, 32, 98, 78, 73, 88, 45, 44, 99, 58, 79]]
+Epoch 0: |                                                                                                                                                     | 2669/? [2:50:13<00:00,  0.26it/s, v_num=pmyy]train step 2670; scene = [['f0712581f6277ffc'], ['5c4442779124ec3d']]; loss = 0.039076
+Epoch 0: |                                                                                                                                                     | 2670/? [2:50:16<00:00,  0.26it/s, v_num=pmyy]context = [[0, 1, 4, 5, 6, 8, 9, 11, 13, 29, 32, 34, 35, 38, 43, 47, 50, 58, 61, 67, 86, 94, 96, 97]]target = [[38, 81, 77, 72, 58, 43, 36, 5, 7, 23, 27, 62, 13, 30, 52, 55, 19, 54, 84, 76, 49, 65, 73, 91]]
+Epoch 0: |                                                                                                                                                     | 2679/? [2:50:51<00:00,  0.26it/s, v_num=pmyy]train step 2680; scene = [['f276b95e48af6e36'], ['e5db691627ea5357']]; loss = 0.057141
+Epoch 0: |                                                                                                                                                     | 2680/? [2:50:55<00:00,  0.26it/s, v_num=pmyy]context = [[108, 120, 121, 130, 145, 149, 158, 169], [9, 17, 36, 43, 44, 69, 77, 78], [107, 120, 131, 140, 141, 142, 147, 152]]target = [[139, 144, 124, 138, 161, 143, 120, 110], [22, 74, 33, 18, 69, 50, 38, 23], [145, 132, 140, 147, 126, 119, 148, 128]]
+Epoch 0: |                                                                                                                                                     | 2689/? [2:51:29<00:00,  0.26it/s, v_num=pmyy]train step 2690; scene = [['2059d1d6c79bc51b'], ['6b1950140a598578'], ['a20b4125ede06429']]; loss = 0.081449
+Epoch 0: |                                                                                                                                                     | 2690/? [2:51:33<00:00,  0.26it/s, v_num=pmyy]context = [[3, 15, 21, 22, 26, 27, 30, 34, 45, 64, 67, 73], [6, 7, 10, 12, 24, 39, 43, 45, 47, 55, 58, 64]]target = [[49, 18, 47, 23, 50, 70, 62, 17, 42, 46, 27, 68], [59, 12, 7, 26, 50, 58, 57, 39, 23, 49, 55, 24]]
+Epoch 0: |                                                                                                                                                     | 2699/? [2:52:08<00:00,  0.26it/s, v_num=pmyy]train step 2700; scene = [['319bd9ea90f25ea3'], ['12879245713d8124'], ['9bb2a6670058b7b2']]; loss = 0.061293
+Epoch 0: |                                                                                                                                                     | 2700/? [2:52:11<00:00,  0.26it/s, v_num=pmyy]context = [[39, 40, 41, 44, 48, 49, 52, 58, 63, 67, 82, 83, 85, 87, 88, 93, 102, 104, 108, 109, 113, 118, 131, 136]]target = [[43, 106, 81, 41, 105, 135, 98, 78, 103, 44, 47, 61, 45, 52, 124, 48, 91, 59, 58, 46, 85, 109, 74, 56]]
+Epoch 0: |                                                                                                                                                     | 2709/? [2:52:43<00:00,  0.26it/s, v_num=pmyy]train step 2710; scene = [['753238098c2307cc']]; loss = 0.074948
+Epoch 0: |                                                                                                                                                     | 2710/? [2:52:47<00:00,  0.26it/s, v_num=pmyy]context = [[6, 7, 8, 9, 17, 19, 29, 32, 39, 48, 54, 55], [110, 116, 118, 121, 122, 128, 134, 138, 144, 149, 170, 178]]target = [[9, 51, 23, 19, 49, 54, 38, 7, 37, 53, 17, 16], [147, 156, 124, 177, 118, 162, 160, 143, 140, 119, 154, 133]]
+Epoch 0: |                                                                                                                                                     | 2719/? [2:53:22<00:00,  0.26it/s, v_num=pmyy]train step 2720; scene = [['7194b8d204f4a0b6'], ['5c0ddb9de8c16f05'], ['1a87a846ba692048']]; loss = 0.044996
+Epoch 0: |                                                                                                                                                     | 2720/? [2:53:26<00:00,  0.26it/s, v_num=pmyy]context = [[73, 88, 89, 91, 93, 96, 104, 111, 114, 115, 118, 120, 121, 125, 133, 136, 137, 143, 144, 145, 151, 161, 165, 170]]target = [[94, 126, 164, 115, 101, 103, 78, 120, 75, 153, 88, 86, 74, 106, 116, 141, 119, 144, 140, 165, 146, 167, 158, 139]]
+Epoch 0: |                                                                                                                                                     | 2729/? [2:54:00<00:00,  0.26it/s, v_num=pmyy]train step 2730; scene = [['b7a4f7a6d35961d4'], ['b7003ac834dc298b'], ['05596054e7569f2b'], ['b8aed6b43cd738c9']]; loss = 0.050474
+Epoch 0: |                                                                                                                                                     | 2730/? [2:54:04<00:00,  0.26it/s, v_num=pmyy]context = [[117, 186, 194, 196], [34, 38, 40, 79], [10, 13, 37, 77], [102, 103, 113, 167], [9, 42, 52, 80], [136, 174, 192, 218]]target = [[187, 180, 158, 134], [57, 76, 69, 70], [49, 37, 70, 57], [158, 143, 120, 136], [15, 34, 75, 65], [138, 214, 217, 143]]
+Epoch 0: |                                                                                                                                                     | 2739/? [2:54:38<00:00,  0.26it/s, v_num=pmyy]train step 2740; scene = [['d9641b3f6ca0b13d'], ['61ef38380e8172c7'], ['3ec462170f378b3f'], ['bc0b4df5aef0a622'], ['6bef29b74b93e80a'], ['8ef643c4e1cb9baf'], ['aa8259399a115c5f'], ['fa35453daaa8e408']]; loss = 0.058113
+Epoch 0: |                                                                                                                                                     | 2740/? [2:54:42<00:00,  0.26it/s, v_num=pmyy]context = [[0, 1, 7, 8, 9, 10, 11, 18, 22, 31, 32, 39, 49, 53, 59, 66, 69, 72, 74, 78, 81, 82, 93, 97]]target = [[74, 55, 50, 21, 47, 85, 2, 42, 41, 40, 39, 93, 49, 45, 23, 66, 73, 38, 89, 36, 87, 30, 48, 9]]
+Epoch 0: |                                                                                                                                                     | 2749/? [2:55:16<00:00,  0.26it/s, v_num=pmyy]train step 2750; scene = [['2225123ef31a93e4']]; loss = 0.089595
+Epoch 0: |                                                                                                                                                     | 2750/? [2:55:20<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 2750; scene = ['3e07add8413f8157'];
+target intrinsic:  tensor(0.8521, device='cuda:0') tensor(0.8523, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8887, device='cuda:0') tensor(0.8874, device='cuda:0')
+[2026-02-25 01:23:13,175][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2750/? [2:55:21<00:00,  0.26it/s, v_num=pmyy]context = [[12, 14, 17, 18, 20, 21, 32, 34, 38, 40, 54, 57, 65, 67, 72, 77, 82, 85, 92, 96, 98, 102, 105, 109]]target = [[101, 39, 104, 51, 15, 33, 16, 24, 94, 58, 72, 48, 20, 17, 22, 37, 107, 63, 40, 79, 69, 27, 28, 60]]
+Epoch 0: |                                                                                                                                                     | 2759/? [2:55:55<00:00,  0.26it/s, v_num=pmyy]train step 2760; scene = [['03a4631365a95cb2'], ['9f622f50132c1efe'], ['4ec29bd9dfd30c2e'], ['314d09d3997725a7'], ['c437043ac9ced6c1'], ['accfed2fa6ff849d'], ['57a1c0730778cddd'], ['bea0e295ee56d42c'], ['75e5ba7e7fb64fc0'], ['71968459b3168e4f'], ['a08c8a37e64ce67d'], ['bf27ccdada0c2373']]; loss = 0.077878
+Epoch 0: |                                                                                                                                                     | 2760/? [2:55:59<00:00,  0.26it/s, v_num=pmyy]context = [[41, 48, 51, 52, 59, 61, 68, 69, 72, 73, 75, 103], [9, 10, 23, 35, 37, 38, 41, 43, 48, 69, 76, 84]]target = [[56, 74, 53, 76, 101, 97, 93, 42, 89, 83, 91, 79], [14, 17, 28, 50, 58, 32, 54, 30, 61, 34, 83, 15]]
+Epoch 0: |                                                                                                                                                     | 2769/? [2:56:33<00:00,  0.26it/s, v_num=pmyy]train step 2770; scene = [['539e0d3713447384'], ['c49e7882e04566c0'], ['69c9010e3d4209bc']]; loss = 0.032319
+Epoch 0: |                                                                                                                                                     | 2770/? [2:56:37<00:00,  0.26it/s, v_num=pmyy]context = [[0, 9, 19, 31, 34, 68], [7, 20, 31, 36, 41, 53], [1, 15, 29, 56, 58, 74], [0, 19, 21, 35, 43, 50]]target = [[10, 2, 39, 12, 11, 26], [21, 31, 11, 28, 27, 52], [16, 13, 66, 17, 65, 6], [39, 21, 41, 27, 40, 25]]
+Epoch 0: |                                                                                                                                                     | 2779/? [2:57:11<00:00,  0.26it/s, v_num=pmyy]train step 2780; scene = [['ede896927ea91dd6'], ['7badba95b5be610e']]; loss = 0.048641
+Epoch 0: |                                                                                                                                                     | 2780/? [2:57:15<00:00,  0.26it/s, v_num=pmyy]context = [[0, 7, 26, 27, 51, 57], [5, 15, 18, 23, 55, 81], [27, 32, 44, 76, 87, 98], [8, 16, 18, 21, 24, 58]]target = [[48, 19, 22, 26, 32, 17], [57, 65, 27, 75, 51, 22], [85, 50, 79, 82, 30, 89], [41, 25, 12, 52, 32, 47]]
+Epoch 0: |                                                                                                                                                     | 2789/? [2:57:50<00:00,  0.26it/s, v_num=pmyy]train step 2790; scene = [['1e7d7ef1404597f0'], ['ea02d0f42c603c21']]; loss = 0.031374
+Epoch 0: |                                                                                                                                                     | 2790/? [2:57:53<00:00,  0.26it/s, v_num=pmyy]context = [[99, 100, 101, 107, 108, 110, 111, 113, 116, 128, 131, 139, 145, 149, 155, 162, 167, 170, 172, 179, 183, 191, 194, 196]]target = [[122, 186, 110, 138, 149, 133, 161, 112, 141, 184, 189, 178, 100, 131, 188, 155, 174, 195, 185, 171, 137, 160, 176, 170]]
+Epoch 0: |                                                                                                                                                     | 2799/? [2:58:28<00:00,  0.26it/s, v_num=pmyy]train step 2800; scene = [['c8e92789f25baec1'], ['b477406d6064f1a3']]; loss = 0.024536
+Epoch 0: |                                                                                                                                                     | 2800/? [2:58:32<00:00,  0.26it/s, v_num=pmyy]context = [[2, 4, 5, 18, 22, 24, 31, 37, 39, 45, 54, 57, 60, 64, 67, 73, 82, 83, 84, 86, 87, 97, 98, 99]]target = [[23, 55, 39, 36, 80, 52, 28, 29, 89, 63, 56, 61, 98, 78, 40, 9, 77, 82, 69, 47, 43, 68, 81, 44]]
+[2026-02-25 01:26:28,369][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2809/? [2:59:11<00:00,  0.26it/s, v_num=pmyy]train step 2810; scene = [['31907a6ebec6ef2e']]; loss = 0.036044
+Epoch 0: |                                                                                                                                                     | 2810/? [2:59:14<00:00,  0.26it/s, v_num=pmyy]context = [[96, 97, 145, 156, 165, 173], [147, 150, 159, 169, 173, 196], [21, 33, 34, 46, 56, 75], [7, 17, 47, 50, 51, 58]]target = [[144, 115, 146, 171, 129, 101], [194, 177, 170, 186, 155, 193], [34, 60, 52, 63, 32, 22], [37, 45, 50, 22, 27, 35]]
+Epoch 0: |                                                                                                                                                     | 2819/? [2:59:49<00:00,  0.26it/s, v_num=pmyy]train step 2820; scene = [['b60f8abb905bb1e1']]; loss = 0.070468
+Epoch 0: |                                                                                                                                                     | 2820/? [2:59:53<00:00,  0.26it/s, v_num=pmyy]context = [[11, 22, 72], [43, 82, 129], [0, 56, 58], [40, 93, 105], [0, 28, 57], [3, 75, 76], [51, 109, 110], [26, 85, 112]]target = [[21, 31, 45], [57, 82, 123], [45, 30, 31], [94, 95, 98], [27, 53, 38], [43, 50, 49], [102, 80, 69], [106, 107, 98]]
+Epoch 0: |                                                                                                                                                     | 2829/? [3:00:28<00:00,  0.26it/s, v_num=pmyy]train step 2830; scene = [['ad1ad6648a30592c'], ['448e12a47b369490'], ['d3dddf816450d1d0'], ['d7fd0247a853f44b'], ['27cf7ba27eed1eae'], ['fd73259e7c2801e8'], ['9869d33078eac049'], ['06779b91ed31eaf1']]; loss = 0.058636
+Epoch 0: |                                                                                                                                                     | 2830/? [3:00:31<00:00,  0.26it/s, v_num=pmyy]context = [[15, 17, 19, 20, 21, 26, 36, 40, 42, 44, 47, 49, 50, 57, 67, 77, 78, 87, 89, 90, 102, 104, 105, 112]]target = [[103, 47, 82, 105, 98, 89, 52, 35, 33, 55, 65, 60, 107, 87, 38, 34, 20, 40, 81, 102, 88, 53, 93, 16]]
+Epoch 0: |                                                                                                                                                     | 2839/? [3:01:06<00:00,  0.26it/s, v_num=pmyy]train step 2840; scene = [['a911c16d8056a34d']]; loss = 0.029494
+Epoch 0: |                                                                                                                                                     | 2840/? [3:01:10<00:00,  0.26it/s, v_num=pmyy]context = [[12, 15, 17, 19, 47, 48, 51, 58], [32, 38, 40, 56, 57, 72, 75, 78], [75, 90, 96, 97, 109, 115, 129, 139]]target = [[53, 44, 45, 42, 50, 17, 36, 54], [71, 72, 37, 48, 54, 39, 74, 52], [104, 134, 124, 132, 129, 127, 137, 121]]
+Epoch 0: |                                                                                                                                                     | 2849/? [3:01:45<00:00,  0.26it/s, v_num=pmyy]train step 2850; scene = [['9753a8291766b5da'], ['3a8f14b855f7f4ee'], ['ba3f2372b7959e95']]; loss = 0.043844
+Epoch 0: |                                                                                                                                                     | 2850/? [3:01:48<00:00,  0.26it/s, v_num=pmyy]context = [[134, 142, 143, 146, 150, 151, 154, 155, 177, 179, 184, 208], [70, 78, 83, 88, 97, 106, 108, 112, 115, 126, 133, 139]]target = [[160, 140, 154, 139, 135, 166, 141, 167, 155, 136, 164, 199], [72, 104, 100, 117, 114, 103, 112, 123, 133, 115, 110, 81]]
+Epoch 0: |                                                                                                                                                     | 2859/? [3:02:22<00:00,  0.26it/s, v_num=pmyy]train step 2860; scene = [['8f4f366720645ea0'], ['fef48b769b17f0ed']]; loss = 0.030962
+Epoch 0: |                                                                                                                                                     | 2860/? [3:02:26<00:00,  0.26it/s, v_num=pmyy]context = [[115, 116, 123, 135, 137, 144, 146, 152, 155, 161, 162, 168, 176, 177, 179, 184, 187, 190, 198, 199, 200, 202, 209, 212]]target = [[195, 155, 181, 142, 154, 202, 144, 161, 136, 132, 137, 192, 148, 189, 188, 174, 178, 133, 156, 134, 124, 197, 199, 166]]
+Epoch 0: |                                                                                                                                                     | 2869/? [3:03:00<00:00,  0.26it/s, v_num=pmyy]train step 2870; scene = [['dff9f50bbc0d0d5d'], ['3baf335bfb54b9c0'], ['c74adb4381e65e99']]; loss = 0.045201
+Epoch 0: |                                                                                                                                                     | 2870/? [3:03:04<00:00,  0.26it/s, v_num=pmyy]context = [[30, 41, 45, 47, 51, 57, 66, 68, 77, 83, 85, 90, 95, 96, 103, 105, 110, 111, 116, 120, 122, 123, 126, 127]]target = [[118, 32, 75, 92, 109, 89, 43, 36, 72, 115, 96, 111, 76, 68, 90, 70, 69, 66, 38, 77, 81, 45, 34, 42]]
+Epoch 0: |                                                                                                                                                     | 2879/? [3:03:39<00:00,  0.26it/s, v_num=pmyy]train step 2880; scene = [['d6bd4b0784843fdd'], ['bd1bc8c20a660a96'], ['f8f0491a2268ee5e'], ['90bcc3b4b3d551a8'], ['57ccd86b6a8979b3'], ['9d66dba4551f79f8'], ['a30f898a6e455745'], ['73973f88f24769be']]; loss = 0.048511
+Epoch 0: |                                                                                                                                                     | 2880/? [3:03:43<00:00,  0.26it/s, v_num=pmyy]context = [[2, 4, 5, 6, 10, 14, 18, 21, 29, 36, 41, 54], [21, 24, 33, 45, 50, 54, 58, 64, 75, 78, 87, 97]]target = [[11, 35, 12, 39, 53, 18, 48, 26, 24, 29, 9, 50], [42, 27, 84, 60, 87, 49, 94, 52, 68, 54, 44, 76]]
+Epoch 0: |                                                                                                                                                     | 2889/? [3:04:17<00:00,  0.26it/s, v_num=pmyy]train step 2890; scene = [['294bfa0dd8a9eada'], ['a069f40a4a017b66']]; loss = 0.052708
+Epoch 0: |                                                                                                                                                     | 2890/? [3:04:21<00:00,  0.26it/s, v_num=pmyy]context = [[3, 8, 10, 16, 25, 28, 29, 35, 36, 37, 43, 55, 60, 69, 72, 75, 82, 84, 87, 90, 91, 94, 95, 100]]target = [[34, 96, 98, 26, 86, 19, 46, 11, 4, 54, 8, 29, 87, 56, 9, 25, 82, 49, 6, 37, 32, 43, 21, 38]]
+Epoch 0: |                                                                                                                                                     | 2899/? [3:04:55<00:00,  0.26it/s, v_num=pmyy]train step 2900; scene = [['cf6618aadac4ddd9'], ['f7e27052900e847e'], ['0b173d0c5951a7f5'], ['3551ff5b8a497fb7']]; loss = 0.032815
+Epoch 0: |                                                                                                                                                     | 2900/? [3:04:59<00:00,  0.26it/s, v_num=pmyy]context = [[38, 95, 96], [42, 65, 102], [13, 22, 77], [198, 232, 261], [13, 56, 82], [7, 21, 79], [0, 49, 80], [83, 129, 157]]target = [[85, 41, 90], [74, 61, 87], [54, 41, 58], [242, 233, 231], [20, 63, 48], [10, 63, 69], [44, 7, 71], [89, 154, 86]]
+Epoch 0: |                                                                                                                                                     | 2909/? [3:05:33<00:00,  0.26it/s, v_num=pmyy]train step 2910; scene = [['b5924605972475e2'], ['64554b0854be0a81']]; loss = 0.036814
+Epoch 0: |                                                                                                                                                     | 2910/? [3:05:37<00:00,  0.26it/s, v_num=pmyy]context = [[0, 15, 38, 47], [103, 127, 147, 175], [174, 186, 243, 250], [14, 30, 57, 62], [88, 115, 122, 134], [64, 65, 75, 127]]target = [[11, 38, 21, 27], [165, 125, 140, 149], [233, 239, 215, 225], [41, 57, 27, 47], [94, 121, 104, 109], [65, 69, 83, 107]]
+Epoch 0: |                                                                                                                                                     | 2919/? [3:06:12<00:00,  0.26it/s, v_num=pmyy]train step 2920; scene = [['d53f3d87d749e474']]; loss = 0.042799
+Epoch 0: |                                                                                                                                                     | 2920/? [3:06:16<00:00,  0.26it/s, v_num=pmyy]context = [[94, 95, 106, 107, 112, 114, 118, 121, 124, 126, 128, 132, 137, 138, 147, 148, 153, 156, 173, 175, 176, 184, 188, 191]]target = [[164, 111, 98, 129, 114, 99, 141, 146, 144, 153, 96, 149, 109, 122, 151, 187, 136, 167, 132, 176, 95, 148, 126, 181]]
+Epoch 0: |                                                                                                                                                     | 2929/? [3:06:48<00:00,  0.26it/s, v_num=pmyy]train step 2930; scene = [['e61532beb3fa8b63'], ['c0803f4d1cb0eeea']]; loss = 0.026138
+Epoch 0: |                                                                                                                                                     | 2930/? [3:06:52<00:00,  0.26it/s, v_num=pmyy]context = [[16, 17, 22, 25, 31, 34, 37, 38, 41, 46, 54, 55, 57, 66, 75, 82, 84, 85, 92, 101, 104, 107, 111, 113]]target = [[109, 112, 73, 65, 111, 102, 34, 28, 20, 49, 46, 85, 17, 81, 50, 58, 22, 89, 52, 19, 63, 87, 59, 25]]
+Epoch 0: |                                                                                                                                                     | 2939/? [3:07:26<00:00,  0.26it/s, v_num=pmyy]train step 2940; scene = [['73aee8654106974f']]; loss = 0.037553
+Epoch 0: |                                                                                                                                                     | 2940/? [3:07:30<00:00,  0.26it/s, v_num=pmyy]context = [[80, 83, 85, 91, 93, 101, 104, 109, 110, 118, 125, 131, 133, 138, 140, 142, 147, 160, 163, 164, 167, 168, 173, 177]]target = [[172, 125, 159, 124, 146, 83, 138, 151, 164, 144, 84, 121, 153, 91, 105, 107, 109, 163, 131, 114, 89, 94, 98, 108]]
+Epoch 0: |                                                                                                                                                     | 2949/? [3:08:04<00:00,  0.26it/s, v_num=pmyy]train step 2950; scene = [['79d5ea6cd3f0fdb2'], ['85fb2f64303a1388'], ['9e3c241e9f50165d'], ['6a94bfa75e7988c8'], ['270022f0b06e71d5'], ['02f1d3b1d43877df'], ['6f4cc17690dcdd2e'], ['70c5a81e8b7868cc'], ['63982f095d5089b0'], ['f7498ea452fed198'], ['d07aa4d1691ccf58'], ['5474e6cd7ccd6d1a']]; loss = 0.055115
+Epoch 0: |                                                                                                                                                     | 2950/? [3:08:07<00:00,  0.26it/s, v_num=pmyy]context = [[1, 20, 21, 24, 25, 26, 27, 30, 36, 39, 50, 53], [60, 74, 75, 77, 82, 90, 99, 106, 111, 113, 114, 133]]target = [[12, 37, 4, 3, 45, 18, 13, 11, 51, 27, 44, 24], [74, 66, 127, 95, 119, 112, 130, 78, 75, 77, 114, 97]]
+Epoch 0: |                                                                                                                                                     | 2959/? [3:08:42<00:00,  0.26it/s, v_num=pmyy]train step 2960; scene = [['0e5f43adc84b0435']]; loss = 0.084964
+Epoch 0: |                                                                                                                                                     | 2960/? [3:08:46<00:00,  0.26it/s, v_num=pmyy]context = [[73, 79, 98, 99, 100, 103, 107, 126], [40, 42, 46, 63, 67, 92, 93, 118], [3, 7, 12, 36, 38, 39, 41, 59]]target = [[113, 108, 119, 82, 81, 84, 86, 124], [90, 106, 43, 103, 69, 87, 113, 111], [18, 41, 24, 30, 7, 20, 53, 54]]
+Epoch 0: |                                                                                                                                                     | 2969/? [3:09:20<00:00,  0.26it/s, v_num=pmyy]train step 2970; scene = [['d3784c9108c25d42']]; loss = 0.037028
+Epoch 0: |                                                                                                                                                     | 2970/? [3:09:24<00:00,  0.26it/s, v_num=pmyy]context = [[10, 11, 17, 29, 30, 31, 33, 39, 48, 60, 63, 64, 69, 70, 72, 87, 91, 92, 93, 95, 99, 100, 104, 107]]target = [[63, 97, 21, 50, 94, 32, 27, 70, 60, 78, 65, 45, 53, 85, 56, 33, 38, 80, 98, 12, 44, 95, 43, 34]]
+Epoch 0: |                                                                                                                                                     | 2979/? [3:09:59<00:00,  0.26it/s, v_num=pmyy]train step 2980; scene = [['058bac6c226fc1a9'], ['7d900c809e896e32']]; loss = 0.039224
+Epoch 0: |                                                                                                                                                     | 2980/? [3:10:02<00:00,  0.26it/s, v_num=pmyy]context = [[64, 69, 70, 71, 83, 85, 86, 93, 94, 96, 111, 112, 113, 117, 123, 138, 139, 140, 141, 148, 149, 153, 159, 161]]target = [[97, 127, 104, 150, 132, 129, 78, 128, 116, 125, 134, 130, 149, 99, 137, 79, 88, 140, 94, 98, 142, 153, 155, 73]]
+Epoch 0: |                                                                                                                                                     | 2989/? [3:10:36<00:00,  0.26it/s, v_num=pmyy]train step 2990; scene = [['51f581faf9000425'], ['89bc971b5dfbd294'], ['511207c07d553599']]; loss = 0.040224
+Epoch 0: |                                                                                                                                                     | 2990/? [3:10:39<00:00,  0.26it/s, v_num=pmyy]context = [[27, 42, 44, 84, 95, 99], [168, 178, 219, 221, 222, 224], [50, 56, 89, 100, 116, 131], [117, 131, 156, 157, 160, 171]]target = [[41, 88, 49, 83, 97, 46], [194, 197, 187, 203, 199, 177], [130, 88, 129, 93, 52, 103], [130, 162, 157, 127, 156, 133]]
+Epoch 0: |                                                                                                                                                     | 2999/? [3:11:13<00:00,  0.26it/s, v_num=pmyy]train step 3000; scene = [['d77284dc9b9d1031']]; loss = 0.101457
+Epoch 0: |                                                                                                                                                     | 3000/? [3:11:17<00:00,  0.26it/s, v_num=pmyy]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 3000; scene = ['a76028640ffa1ef9'];
+target intrinsic:  tensor(0.8569, device='cuda:0') tensor(0.8571, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8821, device='cuda:0') tensor(0.8824, device='cuda:0')
+[2026-02-25 01:39:43,040][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 3000/? [3:11:51<00:00,  0.26it/s, v_num=pmyy]context = [[176, 195, 196, 203, 214, 228, 231, 237], [75, 81, 89, 91, 113, 117, 152, 155], [54, 61, 76, 79, 90, 95, 104, 105]]target = [[207, 201, 216, 186, 203, 231, 198, 179], [144, 150, 79, 87, 147, 123, 115, 138], [86, 82, 99, 72, 100, 74, 66, 62]]
+[2026-02-25 01:39:46,583][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 3001/? [3:11:56<00:00,  0.26it/s, v_num=pmyy]
+`Trainer.fit` stopped: `max_steps=3001` reached.
+Peak VRAM: 103.357 GB (allocated), 134.680 GB (reserved)
+Total elapsed: 3.22 hours
+Saved memory info to: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_FreqSelect/peak_vram_memory.json
diff --git a/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/requirements.txt b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..7fbf9096f92b53f8bb2a7e5467c79ecbe64faca5
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/requirements.txt
@@ -0,0 +1,172 @@
+wheel==0.45.1
+pytz==2025.2
+easydict==1.13
+antlr4-python3-runtime==4.9.3
+wadler_lindig==0.1.7
+urllib3==2.5.0
+tzdata==2025.2
+typing-inspection==0.4.1
+tabulate==0.9.0
+smmap==5.0.2
+kornia_rs==0.1.9
+setuptools==78.1.1
+safetensors==0.5.3
+PyYAML==6.0.2
+PySocks==1.7.1
+pyparsing==3.2.5
+pydantic_core==2.33.2
+pycparser==2.23
+protobuf==6.32.1
+propcache==0.3.2
+proglog==0.1.12
+fsspec==2024.6.1
+platformdirs==4.4.0
+pip==25.2
+pillow==10.4.0
+frozenlist==1.7.0
+packaging==24.2
+opt_einsum==3.4.0
+numpy==1.26.4
+ninja==1.13.0
+fonttools==4.60.0
+networkx==3.4.2
+multidict==6.6.4
+mdurl==0.1.2
+MarkupSafe==3.0.2
+kiwisolver==1.4.9
+imageio-ffmpeg==0.6.0
+idna==3.7
+hf-xet==1.1.10
+gmpy2==2.2.1
+einops==0.8.1
+filelock==3.17.0
+decorator==4.4.2
+dacite==1.9.2
+cycler==0.12.1
+colorama==0.4.6
+click==8.3.0
+nvidia-nvtx-cu12==12.8.90
+charset-normalizer==3.3.2
+certifi==2025.8.3
+beartype==0.19.0
+attrs==25.3.0
+async-timeout==5.0.1
+annotated-types==0.7.0
+aiohappyeyeballs==2.6.1
+yarl==1.20.1
+tifffile==2025.5.10
+sentry-sdk==2.39.0
+scipy==1.15.3
+pydantic==2.11.9
+pandas==2.3.2
+opencv-python==4.11.0.86
+omegaconf==2.3.0
+markdown-it-py==4.0.0
+lightning-utilities==0.14.3
+lazy_loader==0.4
+jaxtyping==0.2.37
+imageio==2.37.0
+gitdb==4.0.12
+contourpy==1.3.2
+colorspacious==1.1.2
+cffi==1.17.1
+aiosignal==1.4.0
+scikit-video==1.1.11
+scikit-image==0.25.2
+rich==14.1.0
+moviepy==1.0.3
+matplotlib==3.10.6
+hydra-core==1.3.2
+nvidia-nccl-cu12==2.27.3
+huggingface-hub==0.35.1
+GitPython==3.1.45
+brotlicffi==1.0.9.2
+aiohttp==3.12.15
+torchmetrics==1.8.2
+opt-einsum-fx==0.1.4
+kornia==0.8.1
+pytorch-lightning==2.5.1
+lpips==0.1.4
+e3nn==0.6.0
+lightning==2.5.1
+nvidia-cusparselt-cu12==0.7.1
+triton==3.4.0
+nvidia-nvjitlink-cu12==12.8.93
+nvidia-curand-cu12==10.3.9.90
+nvidia-cufile-cu12==1.13.1.3
+nvidia-cuda-runtime-cu12==12.8.90
+nvidia-cuda-nvrtc-cu12==12.8.93
+nvidia-cuda-cupti-cu12==12.8.90
+nvidia-cublas-cu12==12.8.4.1
+nvidia-cusparse-cu12==12.5.8.93
+nvidia-cufft-cu12==11.3.3.83
+nvidia-cudnn-cu12==9.10.2.21
+nvidia-cusolver-cu12==11.7.3.90
+torch==2.8.0+cu128
+torchvision==0.23.0+cu128
+torchaudio==2.8.0+cu128
+torch_scatter==2.1.2+pt28cu128
+gsplat==1.5.3
+wandb==0.25.0
+cuda-bindings==13.0.3
+cuda-pathfinder==1.3.3
+Jinja2==3.1.6
+mpmath==1.3.0
+nvidia-cublas==13.1.0.3
+nvidia-cuda-cupti==13.0.85
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-runtime==13.0.96
+nvidia-cudnn-cu13==9.15.1.9
+nvidia-cufft==12.0.0.61
+nvidia-cufile==1.15.1.6
+nvidia-curand==10.4.0.35
+nvidia-cusolver==12.0.4.66
+nvidia-cusparse==12.6.3.3
+nvidia-cusparselt-cu13==0.8.0
+nvidia-nccl-cu13==2.28.9
+nvidia-nvjitlink==13.0.88
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvtx==13.0.85
+requests==2.32.5
+sentencepiece==0.2.1
+sympy==1.14.0
+torchcodec==0.10.0
+torchdata==0.10.0
+torchtext==0.6.0
+anyio==4.12.0
+asttokens==3.0.1
+comm==0.2.3
+debugpy==1.8.19
+executing==2.2.1
+h11==0.16.0
+httpcore==1.0.9
+httpx==0.28.1
+ipykernel==7.1.0
+ipython==9.8.0
+ipython_pygments_lexers==1.1.1
+ipywidgets==8.1.8
+jedi==0.19.2
+jupyter_client==8.7.0
+jupyter_core==5.9.1
+jupyterlab_widgets==3.0.16
+matplotlib-inline==0.2.1
+nest-asyncio==1.6.0
+parso==0.8.5
+pexpect==4.9.0
+prompt_toolkit==3.0.52
+psutil==7.2.1
+ptyprocess==0.7.0
+pure_eval==0.2.3
+Pygments==2.19.2
+python-dateutil==2.9.0.post0
+pyzmq==27.1.0
+shellingham==1.5.4
+six==1.17.0
+stack-data==0.6.3
+tornado==6.5.4
+tqdm==4.67.1
+traitlets==5.14.3
+typer-slim==0.21.0
+typing_extensions==4.15.0
+wcwidth==0.2.14
+widgetsnbextension==4.0.15
diff --git a/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/wandb-metadata.json b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/wandb-metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..8d45ee62b8e0de636699f629712e047f6f6b2f86
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/wandb-metadata.json
@@ -0,0 +1,93 @@
+{
+  "os":  "Linux-6.8.0-90-generic-x86_64-with-glibc2.39",
+  "python":  "CPython 3.12.12",
+  "startedAt":  "2026-02-24T22:27:39.584882Z",
+  "args":  [
+    "+experiment=re10k_ablation_24v",
+    "wandb.mode=online",
+    "wandb.name=ABLATION_0225_FreqSelect",
+    "model.density_control.score_mode=frequency"
+  ],
+  "program":  "-m src.main",
+  "git":  {
+    "remote":  "git@github.com:K-nowing/CVPR2026.git",
+    "commit":  "2512754c6c27ca5150bf17fbcbdde3f192fd53cc"
+  },
+  "email":  "dna9041@korea.ac.kr",
+  "root":  "/workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_FreqSelect",
+  "host":  "27d18dedec6d",
+  "executable":  "/venv/main/bin/python",
+  "cpu_count":  128,
+  "cpu_count_logical":  256,
+  "gpu":  "NVIDIA H200",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "1170378588160",
+      "used":  "636725506048"
+    }
+  },
+  "memory":  {
+    "total":  "1622948257792"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-2649ab80-a3a6-5a1c-0fa5-12bc11bd75e9"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-e92921d9-c681-246f-af93-637e0dc938ca"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-ffe12ffc-9bb7-82de-5692-1ec0ee2e68d8"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-499e5acd-b6ab-2010-c51b-ee9b5aa65825"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-3b2522d9-1c72-e49b-2c30-96165680b74a"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-a9a280c5-b2f9-dc1e-a8a9-7326a74001ff"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-07d0167b-a6a1-1900-2d27-7c6c11598409"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-8362a999-20d1-c27b-5d18-032d23f859ab"
+    }
+  ],
+  "cudaVersion":  "13.1",
+  "writerId":  "1aoh34iwmaamch760bz6silmn5l3ie5b"
+}
\ No newline at end of file
diff --git a/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/wandb-summary.json b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/wandb-summary.json
new file mode 100644
index 0000000000000000000000000000000000000000..4fb5aaa9d7e57a8eddf7c763458d98d8a7725ca2
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/files/wandb-summary.json
@@ -0,0 +1 @@
+{"loss/total":0.10145657509565353,"loss/final_3dgs/lpips":0.009992067702114582,"val/lpips":0.15560707449913025,"loss/camera":0.00027895145467482507,"lr-AdamW/pg1-momentum":0.9,"loss/aux_0/lpips":0.011460522189736366,"loss/aux_2/mse":0.013906704261898994,"loss/scene_scale_reg":0.00029438614728860557,"loss/aux_0/mse":0.014569984748959541,"lr-AdamW/pg2":2e-05,"val/psnr":22.323665618896484,"loss/aux_0/error_score":0.8076989054679871,"loss/aux_2/lpips":0.0103166364133358,"epoch":0,"train/psnr_probabilistic":18.699142456054688,"_runtime":11534,"train/error_scores":{"filenames":["media/images/train/error_scores_201_6255176ede93e5c4c605.png"],"captions":[["0621c7675fab1418"]],"_type":"images/separated","width":1328,"height":2120,"format":"png","count":1},"loss/aux_1/mse":0.014023929834365845,"train/comparison":{"height":2154,"format":"png","count":1,"filenames":["media/images/train/comparison_202_2d515c3482668baeba0f.png"],"captions":[["0621c7675fab1418"]],"_type":"images/separated","width":1328},"error_scores":{"format":"png","count":1,"filenames":["media/images/error_scores_199_bbf557521907e54e9e40.png"],"captions":["a76028640ffa1ef9"],"_type":"images/separated","width":800,"height":536},"loss/aux_1/lpips":0.010416326113045216,"train/scene_scale":1.0072107315063477,"_step":202,"_timestamp":1.771983588695968e+09,"val/gaussian_num_ratio":0.3998870849609375,"trainer/global_step":3001,"loss/final_3dgs/mse":0.013686501421034336,"val/ssim":0.8440837860107422,"loss/aux_1/error_score":0.4816555380821228,"active_mask_imgs":{"filenames":["media/images/active_mask_imgs_198_24c7ded6b719c7a30450.png"],"captions":["a76028640ffa1ef9"],"_type":"images/separated","width":536,"height":800,"format":"png","count":1},"comparison":{"width":1064,"height":1098,"format":"png","count":1,"filenames":["media/images/comparison_197_e0879eb637c4b3dfe984.png"],"captions":["a76028640ffa1ef9"],"_type":"images/separated"},"_wandb":{"runtime":11534},"lr-AdamW/pg1":2.003594834351718e-05,"info/global_step":3000,"lr-AdamW/pg2-momentum":0.9}
\ No newline at end of file
diff --git a/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug-core.log b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug-core.log
new file mode 100644
index 0000000000000000000000000000000000000000..3396e8edcc8679483b725500cd5688e4f55a8d93
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug-core.log
@@ -0,0 +1,15 @@
+{"time":"2026-02-24T22:27:39.691505272Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmphvccr9ry/port-113743.txt","pid":113743,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-02-24T22:27:39.692335245Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":113743}
+{"time":"2026-02-24T22:27:39.692317115Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-113743-116175-1521879483/socket","Net":"unix"}}
+{"time":"2026-02-24T22:27:39.872966329Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-02-24T22:27:39.882057082Z","level":"INFO","msg":"handleInformInit: received","streamId":"y7wvpmyy","id":"1(@)"}
+{"time":"2026-02-24T22:27:40.294862883Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"y7wvpmyy","id":"1(@)"}
+{"time":"2026-02-24T22:27:46.739505276Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"ml9idastztfo"}
+{"time":"2026-02-25T01:39:55.96564956Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
+{"time":"2026-02-25T01:39:55.965789472Z","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-02-25T01:39:55.965784692Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
+{"time":"2026-02-25T01:39:55.965834972Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
+{"time":"2026-02-25T01:39:55.965861283Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-113743-116175-1521879483/socket","Net":"unix"}}
+{"time":"2026-02-25T01:39:57.156626442Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
+{"time":"2026-02-25T01:39:57.156689353Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
+{"time":"2026-02-25T01:39:57.156724243Z","level":"INFO","msg":"server is closed"}
diff --git a/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug-internal.log b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..177c83d79642e27d5ccf1769f6566b523253107a
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug-internal.log
@@ -0,0 +1,12 @@
+{"time":"2026-02-24T22:27:39.882209485Z","level":"INFO","msg":"stream: starting","core version":"0.25.0"}
+{"time":"2026-02-24T22:27:40.294571378Z","level":"INFO","msg":"stream: created new stream","id":"y7wvpmyy"}
+{"time":"2026-02-24T22:27:40.2947114Z","level":"INFO","msg":"handler: started","stream_id":"y7wvpmyy"}
+{"time":"2026-02-24T22:27:40.294855053Z","level":"INFO","msg":"stream: started","id":"y7wvpmyy"}
+{"time":"2026-02-24T22:27:40.294904223Z","level":"INFO","msg":"sender: started","stream_id":"y7wvpmyy"}
+{"time":"2026-02-24T22:27:40.294940724Z","level":"INFO","msg":"writer: started","stream_id":"y7wvpmyy"}
+{"time":"2026-02-25T01:00:56.785103175Z","level":"INFO","msg":"api: retrying HTTP error","status":502,"url":"https://api.wandb.ai/files/know/DCSplat/y7wvpmyy/file_stream","body":"\n<html><head>\n<meta http-equiv=\"content-type\" content=\"text/html;charset=utf-8\">\n<title>502 Server Error</title>\n</head>\n<body text=#000000 bgcolor=#ffffff>\n<h1>Error: Server Error</h1>\n<h2>The server encountered a temporary error and could not complete your request.<p>Please try again in 30 seconds.</h2>\n<h2></h2>\n</body></html>\n"}
+{"time":"2026-02-25T01:39:55.965783052Z","level":"INFO","msg":"stream: closing","id":"y7wvpmyy"}
+{"time":"2026-02-25T01:39:56.929029575Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-02-25T01:39:57.1548805Z","level":"INFO","msg":"handler: closed","stream_id":"y7wvpmyy"}
+{"time":"2026-02-25T01:39:57.155103083Z","level":"INFO","msg":"sender: closed","stream_id":"y7wvpmyy"}
+{"time":"2026-02-25T01:39:57.155127144Z","level":"INFO","msg":"stream: closed","id":"y7wvpmyy"}
diff --git a/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug.log b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..159b8e67b209c045d745b123179f99e28311277a
--- /dev/null
+++ b/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug.log
@@ -0,0 +1,21 @@
+2026-02-24 22:27:39,587 INFO    MainThread:113743 [wandb_setup.py:_flush():81] Current SDK version is 0.25.0
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_setup.py:_flush():81] Configure stats pid to 113743
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug.log
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_FreqSelect/wandb/run-20260224_222739-y7wvpmyy/logs/debug-internal.log
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_init.py:init():844] calling init triggers
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
+config: {'model': {'encoder': {'name': 'dcsplat', 'input_image_shape': [518, 518], 'head_mode': 'pcd', 'num_level': 3, 'gs_param_dim': 256, 'align_corners': False, 'use_voxelize': True}, 'decoder': {'name': 'splatting_cuda', 'background_color': [0.0, 0.0, 0.0], 'make_scale_invariant': False}, 'density_control': {'name': 'density_control_module', 'mean_dim': 32, 'gs_param_dim': 256, 'refinement_layer_num': 1, 'num_level': 3, 'grad_mode': 'absgrad', 'use_mean_features': True, 'refinement_type': 'voxelize', 'refinement_hidden_dim': 32, 'aggregation_mode': 'mean', 'num_heads': 1, 'score_mode': 'frequency', 'latent_dim': 128, 'num_latents': 64, 'num_self_attn_per_block': 2, 'voxel_size': 0.001, 'aux_refine': False, 'refine_error': False, 'use_refine_module': True, 'voxelize_activate': True, 'use_depth': False}}, 'render_loss': {'mse': {'weight': 1.0}, 'lpips': {'weight': 0.05, 'apply_after_step': 0}}, 'density_control_loss': {'error_score': {'weight': 0.01, 'log_scale': False, 'grad_scale': 10000.0, 'mode': 'original'}}, 'direct_loss': {'l1': {'weight': 0.8}, 'ssim': {'weight': 0.2}}, 'wandb': {'project': 'DCSplat', 'entity': 'scene-representation-group', 'name': 'ABLATION_0225_FreqSelect', 'mode': 'online', 'tags': ['re10k', '256x256']}, 'mode': 'train', 'data_loader': {'train': {'num_workers': 16, 'persistent_workers': True, 'batch_size': 16, 'seed': 1234}, 'test': {'num_workers': 4, 'persistent_workers': False, 'batch_size': 1, 'seed': 2345}, 'val': {'num_workers': 1, 'persistent_workers': True, 'batch_size': 1, 'seed': 3456}}, 'optimizer': {'lr': 0.0002, 'warm_up_steps': 25, 'backbone_lr_multiplier': 0.1, 'backbone_trainable': 'T+H', 'accumulate': 1}, 'checkpointing': {'load': None, 'every_n_train_steps': 1500, 'save_top_k': 2, 'save_weights_only': False}, 'train': {'extended_visualization': False, 'print_log_every_n_steps': 10, 'camera_loss': 10.0, 'one_sample_validation': None, 'align_corners': False, 'intrinsic_scaling': False, 'verbose': False, 'beta_dist_param': [0.5, 4.0], 'use_refine_aux': False, 'train_target_set': True, 'train_gs_num': 1, 'ext_scale_detach': False, 'cam_scale_mode': 'sum', 'scene_scale_reg_loss': 0.01, 'train_aux': True, 'vggt_cam_loss': True, 'vggt_distil': False, 'context_view_train': False}, 'test': {'output_path': 'test/ablation/re10k', 'align_pose': False, 'pose_align_steps': 100, 'rot_opt_lr': 0.005, 'trans_opt_lr': 0.005, 'compute_scores': True, 'save_image': False, 'save_video': False, 'save_active_mask_image': False, 'save_error_score_image': False, 'save_compare': False, 'pred_intrinsic': False, 'error_threshold': 0.4, 'error_threshold_list': [0.2, 0.4, 0.6, 0.8, 1.0], 'threshold_mode': 'ratio', 'nvs_view_N_list': [3, 6, 16, 32, 64]}, 'seed': 111123, 'trainer': {'max_steps': 3001, 'val_check_interval': 250, 'gradient_clip_val': 0.5, 'num_nodes': 1}, 'dataset': {'re10k': {'make_baseline_1': True, 'relative_pose': True, 'augment': True, 'background_color': [0.0, 0.0, 0.0], 'overfit_to_scene': None, 'skip_bad_shape': True, 'view_sampler': {'name': 'bounded', 'num_target_views': 4, 'num_context_views': 2, 'min_distance_between_context_views': 45, 'max_distance_between_context_views': 90, 'min_distance_to_context_views': 0, 'warm_up_steps': 1000, 'initial_min_distance_between_context_views': 25, 'initial_max_distance_between_context_views': 25, 'same_target_gap': False, 'num_target_set': 3}, 'name': 're10k', 'roots': ['datasets/re10k'], 'input_image_shape': [256, 256], 'original_image_shape': [360, 640], 'cameras_are_circular': False, 'baseline_min': 0.001, 'baseline_max': 10000000000.0, 'max_fov': 100.0, 'dynamic_context_views': True, 'max_context_views_per_gpu': 24}}, '_wandb': {}}
+2026-02-24 22:27:39,588 INFO    MainThread:113743 [wandb_init.py:init():892] starting backend
+2026-02-24 22:27:39,873 INFO    MainThread:113743 [wandb_init.py:init():895] sending inform_init request
+2026-02-24 22:27:39,880 INFO    MainThread:113743 [wandb_init.py:init():903] backend started and connected
+2026-02-24 22:27:39,887 INFO    MainThread:113743 [wandb_init.py:init():973] updated telemetry
+2026-02-24 22:27:39,894 INFO    MainThread:113743 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
+2026-02-24 22:27:41,506 INFO    MainThread:113743 [wandb_init.py:init():1042] starting run threads in backend
+2026-02-24 22:27:41,632 INFO    MainThread:113743 [wandb_run.py:_console_start():2524] atexit reg
+2026-02-24 22:27:41,632 INFO    MainThread:113743 [wandb_run.py:_redirect():2373] redirect: wrap_raw
+2026-02-24 22:27:41,632 INFO    MainThread:113743 [wandb_run.py:_redirect():2442] Wrapping output streams.
+2026-02-24 22:27:41,632 INFO    MainThread:113743 [wandb_run.py:_redirect():2465] Redirects installed.
+2026-02-24 22:27:41,635 INFO    MainThread:113743 [wandb_init.py:init():1082] run started, returning control to user process
+2026-02-25 01:39:55,965 INFO    wandb-AsyncioManager-main:113743 [service_client.py:_forward_responses():134] Reached EOF.
+2026-02-25 01:39:55,965 INFO    wandb-AsyncioManager-main:113743 [mailbox.py:close():155] Closing mailbox, abandoning 1 handles.
diff --git a/ABLATION_0225_OURS/.hydra/config.yaml b/ABLATION_0225_OURS/.hydra/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..151f5b47036394aae3992c17bcb3b0efbb19594c
--- /dev/null
+++ b/ABLATION_0225_OURS/.hydra/config.yaml
@@ -0,0 +1,185 @@
+model:
+  encoder:
+    name: dcsplat
+    input_image_shape:
+    - 518
+    - 518
+    head_mode: pcd
+    num_level: 3
+    gs_param_dim: 256
+    align_corners: false
+    use_voxelize: true
+  decoder:
+    name: splatting_cuda
+    background_color:
+    - 0.0
+    - 0.0
+    - 0.0
+    make_scale_invariant: false
+  density_control:
+    name: density_control_module
+    mean_dim: 32
+    gs_param_dim: 256
+    refinement_layer_num: 1
+    num_level: 3
+    grad_mode: absgrad
+    use_mean_features: true
+    refinement_type: voxelize
+    refinement_hidden_dim: 32
+    aggregation_mode: mean
+    num_heads: 1
+    score_mode: absgrad
+    latent_dim: 128
+    num_latents: 64
+    num_self_attn_per_block: 2
+    voxel_size: 0.001
+    aux_refine: false
+    refine_error: false
+    use_refine_module: true
+    voxelize_activate: true
+    use_depth: false
+render_loss:
+  mse:
+    weight: 1.0
+  lpips:
+    weight: 0.05
+    apply_after_step: 0
+density_control_loss:
+  error_score:
+    weight: 0.01
+    log_scale: false
+    grad_scale: 10000.0
+    mode: original
+direct_loss:
+  l1:
+    weight: 0.8
+  ssim:
+    weight: 0.2
+wandb:
+  project: DCSplat
+  entity: scene-representation-group
+  name: ABLATION_0225_OURS
+  mode: online
+  tags:
+  - re10k
+  - 256x256
+mode: train
+data_loader:
+  train:
+    num_workers: 16
+    persistent_workers: true
+    batch_size: 16
+    seed: 1234
+  test:
+    num_workers: 4
+    persistent_workers: false
+    batch_size: 1
+    seed: 2345
+  val:
+    num_workers: 1
+    persistent_workers: true
+    batch_size: 1
+    seed: 3456
+optimizer:
+  lr: 0.0002
+  warm_up_steps: 25
+  backbone_lr_multiplier: 0.1
+  backbone_trainable: T+H
+  accumulate: 1
+checkpointing:
+  load: null
+  every_n_train_steps: 1500
+  save_top_k: 2
+  save_weights_only: false
+train:
+  extended_visualization: false
+  print_log_every_n_steps: 10
+  camera_loss: 10.0
+  one_sample_validation: null
+  align_corners: false
+  intrinsic_scaling: false
+  verbose: false
+  beta_dist_param:
+  - 0.5
+  - 4.0
+  use_refine_aux: false
+  train_target_set: true
+  train_gs_num: 1
+  ext_scale_detach: false
+  cam_scale_mode: sum
+  scene_scale_reg_loss: 0.01
+  train_aux: true
+  vggt_cam_loss: true
+  vggt_distil: false
+  context_view_train: false
+test:
+  output_path: test/ablation/re10k
+  align_pose: false
+  pose_align_steps: 100
+  rot_opt_lr: 0.005
+  trans_opt_lr: 0.005
+  compute_scores: true
+  save_image: false
+  save_video: false
+  save_active_mask_image: false
+  save_error_score_image: false
+  save_compare: false
+  pred_intrinsic: false
+  error_threshold: 0.4
+  error_threshold_list:
+  - 0.2
+  - 0.4
+  - 0.6
+  - 0.8
+  - 1.0
+  threshold_mode: ratio
+  nvs_view_N_list:
+  - 3
+  - 6
+  - 16
+  - 32
+  - 64
+seed: 111123
+trainer:
+  max_steps: 3001
+  val_check_interval: 250
+  gradient_clip_val: 0.5
+  num_nodes: 1
+dataset:
+  re10k:
+    make_baseline_1: true
+    relative_pose: true
+    augment: true
+    background_color:
+    - 0.0
+    - 0.0
+    - 0.0
+    overfit_to_scene: null
+    skip_bad_shape: true
+    view_sampler:
+      name: bounded
+      num_target_views: 4
+      num_context_views: 2
+      min_distance_between_context_views: 45
+      max_distance_between_context_views: 90
+      min_distance_to_context_views: 0
+      warm_up_steps: 1000
+      initial_min_distance_between_context_views: 25
+      initial_max_distance_between_context_views: 25
+      same_target_gap: false
+      num_target_set: 3
+    name: re10k
+    roots:
+    - datasets/re10k
+    input_image_shape:
+    - 256
+    - 256
+    original_image_shape:
+    - 360
+    - 640
+    cameras_are_circular: false
+    baseline_min: 0.001
+    baseline_max: 10000000000.0
+    max_fov: 100.0
+    dynamic_context_views: true
+    max_context_views_per_gpu: 24
diff --git a/ABLATION_0225_OURS/.hydra/hydra.yaml b/ABLATION_0225_OURS/.hydra/hydra.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..db0569717d59b0bf4238435dd27f9b2da6bc915f
--- /dev/null
+++ b/ABLATION_0225_OURS/.hydra/hydra.yaml
@@ -0,0 +1,164 @@
+hydra:
+  run:
+    dir: outputs/ablation/re10k/${wandb.name}
+  sweep:
+    dir: multirun/${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:
+    - +experiment=re10k_ablation_24v
+    - wandb.mode=online
+    - wandb.name=ABLATION_0225_OURS
+  job:
+    name: main
+    chdir: null
+    override_dirname: +experiment=re10k_ablation_24v,wandb.mode=online,wandb.name=ABLATION_0225_OURS
+    id: ???
+    num: ???
+    config_name: main
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.3.2
+    version_base: '1.3'
+    cwd: /workspace/code/CVPR2026
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/code/CVPR2026/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_OURS
+    choices:
+      experiment: re10k_ablation_24v
+      dataset@dataset.re10k: re10k
+      dataset/view_sampler_dataset_specific_config@dataset.re10k.view_sampler: bounded_re10k
+      dataset/view_sampler@dataset.re10k.view_sampler: bounded
+      model/density_control: density_control_module
+      model/decoder: splatting_cuda
+      model/encoder: dcsplat
+      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
diff --git a/ABLATION_0225_OURS/.hydra/overrides.yaml b/ABLATION_0225_OURS/.hydra/overrides.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..8d375ab444c4f0ab630717d5e9854bae311c9ba2
--- /dev/null
+++ b/ABLATION_0225_OURS/.hydra/overrides.yaml
@@ -0,0 +1,3 @@
+- +experiment=re10k_ablation_24v
+- wandb.mode=online
+- wandb.name=ABLATION_0225_OURS
diff --git a/ABLATION_0225_OURS/wandb/debug-internal.log b/ABLATION_0225_OURS/wandb/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..14163630e92832b08bd4b3f84b08b6cb465a87f7
--- /dev/null
+++ b/ABLATION_0225_OURS/wandb/debug-internal.log
@@ -0,0 +1,11 @@
+{"time":"2026-02-24T19:15:08.591653472Z","level":"INFO","msg":"stream: starting","core version":"0.25.0"}
+{"time":"2026-02-24T19:15:09.22244861Z","level":"INFO","msg":"stream: created new stream","id":"0b125b6z"}
+{"time":"2026-02-24T19:15:09.222653934Z","level":"INFO","msg":"handler: started","stream_id":"0b125b6z"}
+{"time":"2026-02-24T19:15:09.222865877Z","level":"INFO","msg":"stream: started","id":"0b125b6z"}
+{"time":"2026-02-24T19:15:09.222943579Z","level":"INFO","msg":"writer: started","stream_id":"0b125b6z"}
+{"time":"2026-02-24T19:15:09.222946409Z","level":"INFO","msg":"sender: started","stream_id":"0b125b6z"}
+{"time":"2026-02-24T22:26:34.518352356Z","level":"INFO","msg":"stream: closing","id":"0b125b6z"}
+{"time":"2026-02-24T22:26:35.362766174Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-02-24T22:26:35.604459738Z","level":"INFO","msg":"handler: closed","stream_id":"0b125b6z"}
+{"time":"2026-02-24T22:26:35.604786383Z","level":"INFO","msg":"sender: closed","stream_id":"0b125b6z"}
+{"time":"2026-02-24T22:26:35.604815153Z","level":"INFO","msg":"stream: closed","id":"0b125b6z"}
diff --git a/ABLATION_0225_OURS/wandb/debug.log b/ABLATION_0225_OURS/wandb/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..d6cac35968a95acc41624d3c77cf62af7e1e3185
--- /dev/null
+++ b/ABLATION_0225_OURS/wandb/debug.log
@@ -0,0 +1,21 @@
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_setup.py:_flush():81] Current SDK version is 0.25.0
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_setup.py:_flush():81] Configure stats pid to 90349
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug.log
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug-internal.log
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_init.py:init():844] calling init triggers
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
+config: {'model': {'encoder': {'name': 'dcsplat', 'input_image_shape': [518, 518], 'head_mode': 'pcd', 'num_level': 3, 'gs_param_dim': 256, 'align_corners': False, 'use_voxelize': True}, 'decoder': {'name': 'splatting_cuda', 'background_color': [0.0, 0.0, 0.0], 'make_scale_invariant': False}, 'density_control': {'name': 'density_control_module', 'mean_dim': 32, 'gs_param_dim': 256, 'refinement_layer_num': 1, 'num_level': 3, 'grad_mode': 'absgrad', 'use_mean_features': True, 'refinement_type': 'voxelize', 'refinement_hidden_dim': 32, 'aggregation_mode': 'mean', 'num_heads': 1, 'score_mode': 'absgrad', 'latent_dim': 128, 'num_latents': 64, 'num_self_attn_per_block': 2, 'voxel_size': 0.001, 'aux_refine': False, 'refine_error': False, 'use_refine_module': True, 'voxelize_activate': True, 'use_depth': False}}, 'render_loss': {'mse': {'weight': 1.0}, 'lpips': {'weight': 0.05, 'apply_after_step': 0}}, 'density_control_loss': {'error_score': {'weight': 0.01, 'log_scale': False, 'grad_scale': 10000.0, 'mode': 'original'}}, 'direct_loss': {'l1': {'weight': 0.8}, 'ssim': {'weight': 0.2}}, 'wandb': {'project': 'DCSplat', 'entity': 'scene-representation-group', 'name': 'ABLATION_0225_OURS', 'mode': 'online', 'tags': ['re10k', '256x256']}, 'mode': 'train', 'data_loader': {'train': {'num_workers': 16, 'persistent_workers': True, 'batch_size': 16, 'seed': 1234}, 'test': {'num_workers': 4, 'persistent_workers': False, 'batch_size': 1, 'seed': 2345}, 'val': {'num_workers': 1, 'persistent_workers': True, 'batch_size': 1, 'seed': 3456}}, 'optimizer': {'lr': 0.0002, 'warm_up_steps': 25, 'backbone_lr_multiplier': 0.1, 'backbone_trainable': 'T+H', 'accumulate': 1}, 'checkpointing': {'load': None, 'every_n_train_steps': 1500, 'save_top_k': 2, 'save_weights_only': False}, 'train': {'extended_visualization': False, 'print_log_every_n_steps': 10, 'camera_loss': 10.0, 'one_sample_validation': None, 'align_corners': False, 'intrinsic_scaling': False, 'verbose': False, 'beta_dist_param': [0.5, 4.0], 'use_refine_aux': False, 'train_target_set': True, 'train_gs_num': 1, 'ext_scale_detach': False, 'cam_scale_mode': 'sum', 'scene_scale_reg_loss': 0.01, 'train_aux': True, 'vggt_cam_loss': True, 'vggt_distil': False, 'context_view_train': False}, 'test': {'output_path': 'test/ablation/re10k', 'align_pose': False, 'pose_align_steps': 100, 'rot_opt_lr': 0.005, 'trans_opt_lr': 0.005, 'compute_scores': True, 'save_image': False, 'save_video': False, 'save_active_mask_image': False, 'save_error_score_image': False, 'save_compare': False, 'pred_intrinsic': False, 'error_threshold': 0.4, 'error_threshold_list': [0.2, 0.4, 0.6, 0.8, 1.0], 'threshold_mode': 'ratio', 'nvs_view_N_list': [3, 6, 16, 32, 64]}, 'seed': 111123, 'trainer': {'max_steps': 3001, 'val_check_interval': 250, 'gradient_clip_val': 0.5, 'num_nodes': 1}, 'dataset': {'re10k': {'make_baseline_1': True, 'relative_pose': True, 'augment': True, 'background_color': [0.0, 0.0, 0.0], 'overfit_to_scene': None, 'skip_bad_shape': True, 'view_sampler': {'name': 'bounded', 'num_target_views': 4, 'num_context_views': 2, 'min_distance_between_context_views': 45, 'max_distance_between_context_views': 90, 'min_distance_to_context_views': 0, 'warm_up_steps': 1000, 'initial_min_distance_between_context_views': 25, 'initial_max_distance_between_context_views': 25, 'same_target_gap': False, 'num_target_set': 3}, 'name': 're10k', 'roots': ['datasets/re10k'], 'input_image_shape': [256, 256], 'original_image_shape': [360, 640], 'cameras_are_circular': False, 'baseline_min': 0.001, 'baseline_max': 10000000000.0, 'max_fov': 100.0, 'dynamic_context_views': True, 'max_context_views_per_gpu': 24}}, '_wandb': {}}
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_init.py:init():892] starting backend
+2026-02-24 19:15:08,582 INFO    MainThread:90349 [wandb_init.py:init():895] sending inform_init request
+2026-02-24 19:15:08,588 INFO    MainThread:90349 [wandb_init.py:init():903] backend started and connected
+2026-02-24 19:15:08,591 INFO    MainThread:90349 [wandb_init.py:init():973] updated telemetry
+2026-02-24 19:15:08,598 INFO    MainThread:90349 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
+2026-02-24 19:15:10,455 INFO    MainThread:90349 [wandb_init.py:init():1042] starting run threads in backend
+2026-02-24 19:15:10,580 INFO    MainThread:90349 [wandb_run.py:_console_start():2524] atexit reg
+2026-02-24 19:15:10,580 INFO    MainThread:90349 [wandb_run.py:_redirect():2373] redirect: wrap_raw
+2026-02-24 19:15:10,580 INFO    MainThread:90349 [wandb_run.py:_redirect():2442] Wrapping output streams.
+2026-02-24 19:15:10,582 INFO    MainThread:90349 [wandb_run.py:_redirect():2465] Redirects installed.
+2026-02-24 19:15:10,584 INFO    MainThread:90349 [wandb_init.py:init():1082] run started, returning control to user process
+2026-02-24 22:26:34,518 INFO    wandb-AsyncioManager-main:90349 [service_client.py:_forward_responses():134] Reached EOF.
+2026-02-24 22:26:34,518 INFO    wandb-AsyncioManager-main:90349 [mailbox.py:close():155] Closing mailbox, abandoning 1 handles.
diff --git a/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/config.yaml b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..5a708aa39a2743e4cf1e70e30e41b2bb700df79e
--- /dev/null
+++ b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/config.yaml
@@ -0,0 +1,306 @@
+_wandb:
+    value:
+        cli_version: 0.25.0
+        e:
+            lma14qrq4ffkxha58hrfyhtyvrmlfx2i:
+                args:
+                    - +experiment=re10k_ablation_24v
+                    - wandb.mode=online
+                    - wandb.name=ABLATION_0225_OURS
+                cpu_count: 128
+                cpu_count_logical: 256
+                cudaVersion: "13.1"
+                disk:
+                    /:
+                        total: "1170378588160"
+                        used: "612674392064"
+                email: dna9041@korea.ac.kr
+                executable: /venv/main/bin/python
+                git:
+                    commit: 2512754c6c27ca5150bf17fbcbdde3f192fd53cc
+                    remote: git@github.com:K-nowing/CVPR2026.git
+                gpu: NVIDIA H200
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-2649ab80-a3a6-5a1c-0fa5-12bc11bd75e9
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-e92921d9-c681-246f-af93-637e0dc938ca
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-ffe12ffc-9bb7-82de-5692-1ec0ee2e68d8
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-499e5acd-b6ab-2010-c51b-ee9b5aa65825
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-3b2522d9-1c72-e49b-2c30-96165680b74a
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-a9a280c5-b2f9-dc1e-a8a9-7326a74001ff
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-07d0167b-a6a1-1900-2d27-7c6c11598409
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-8362a999-20d1-c27b-5d18-032d23f859ab
+                host: 27d18dedec6d
+                memory:
+                    total: "1622948257792"
+                os: Linux-6.8.0-90-generic-x86_64-with-glibc2.39
+                program: -m src.main
+                python: CPython 3.12.12
+                root: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_OURS
+                startedAt: "2026-02-24T19:15:08.304921Z"
+                writerId: lma14qrq4ffkxha58hrfyhtyvrmlfx2i
+        m:
+            - "1": trainer/global_step
+              "6":
+                - 3
+              "7": []
+            - "2": '*'
+              "5": 1
+              "6":
+                - 1
+              "7": []
+        python_version: 3.12.12
+        t:
+            "1":
+                - 1
+                - 41
+                - 49
+                - 50
+                - 106
+            "2":
+                - 1
+                - 41
+                - 49
+                - 50
+                - 106
+            "3":
+                - 7
+                - 13
+                - 15
+                - 16
+                - 66
+            "4": 3.12.12
+            "5": 0.25.0
+            "12": 0.25.0
+            "13": linux-x86_64
+checkpointing:
+    value:
+        every_n_train_steps: 1500
+        load: null
+        save_top_k: 2
+        save_weights_only: false
+data_loader:
+    value:
+        test:
+            batch_size: 1
+            num_workers: 4
+            persistent_workers: false
+            seed: 2345
+        train:
+            batch_size: 16
+            num_workers: 16
+            persistent_workers: true
+            seed: 1234
+        val:
+            batch_size: 1
+            num_workers: 1
+            persistent_workers: true
+            seed: 3456
+dataset:
+    value:
+        re10k:
+            augment: true
+            background_color:
+                - 0
+                - 0
+                - 0
+            baseline_max: 1e+10
+            baseline_min: 0.001
+            cameras_are_circular: false
+            dynamic_context_views: true
+            input_image_shape:
+                - 256
+                - 256
+            make_baseline_1: true
+            max_context_views_per_gpu: 24
+            max_fov: 100
+            name: re10k
+            original_image_shape:
+                - 360
+                - 640
+            overfit_to_scene: null
+            relative_pose: true
+            roots:
+                - datasets/re10k
+            skip_bad_shape: true
+            view_sampler:
+                initial_max_distance_between_context_views: 25
+                initial_min_distance_between_context_views: 25
+                max_distance_between_context_views: 90
+                min_distance_between_context_views: 45
+                min_distance_to_context_views: 0
+                name: bounded
+                num_context_views: 2
+                num_target_set: 3
+                num_target_views: 4
+                same_target_gap: false
+                warm_up_steps: 1000
+density_control_loss:
+    value:
+        error_score:
+            grad_scale: 10000
+            log_scale: false
+            mode: original
+            weight: 0.01
+direct_loss:
+    value:
+        l1:
+            weight: 0.8
+        ssim:
+            weight: 0.2
+mode:
+    value: train
+model:
+    value:
+        decoder:
+            background_color:
+                - 0
+                - 0
+                - 0
+            make_scale_invariant: false
+            name: splatting_cuda
+        density_control:
+            aggregation_mode: mean
+            aux_refine: false
+            grad_mode: absgrad
+            gs_param_dim: 256
+            latent_dim: 128
+            mean_dim: 32
+            name: density_control_module
+            num_heads: 1
+            num_latents: 64
+            num_level: 3
+            num_self_attn_per_block: 2
+            refine_error: false
+            refinement_hidden_dim: 32
+            refinement_layer_num: 1
+            refinement_type: voxelize
+            score_mode: absgrad
+            use_depth: false
+            use_mean_features: true
+            use_refine_module: true
+            voxel_size: 0.001
+            voxelize_activate: true
+        encoder:
+            align_corners: false
+            gs_param_dim: 256
+            head_mode: pcd
+            input_image_shape:
+                - 518
+                - 518
+            name: dcsplat
+            num_level: 3
+            use_voxelize: true
+optimizer:
+    value:
+        accumulate: 1
+        backbone_lr_multiplier: 0.1
+        backbone_trainable: T+H
+        lr: 0.0002
+        warm_up_steps: 25
+render_loss:
+    value:
+        lpips:
+            apply_after_step: 0
+            weight: 0.05
+        mse:
+            weight: 1
+seed:
+    value: 111123
+test:
+    value:
+        align_pose: false
+        compute_scores: true
+        error_threshold: 0.4
+        error_threshold_list:
+            - 0.2
+            - 0.4
+            - 0.6
+            - 0.8
+            - 1
+        nvs_view_N_list:
+            - 3
+            - 6
+            - 16
+            - 32
+            - 64
+        output_path: test/ablation/re10k
+        pose_align_steps: 100
+        pred_intrinsic: false
+        rot_opt_lr: 0.005
+        save_active_mask_image: false
+        save_compare: false
+        save_error_score_image: false
+        save_image: false
+        save_video: false
+        threshold_mode: ratio
+        trans_opt_lr: 0.005
+train:
+    value:
+        align_corners: false
+        beta_dist_param:
+            - 0.5
+            - 4
+        cam_scale_mode: sum
+        camera_loss: 10
+        context_view_train: false
+        ext_scale_detach: false
+        extended_visualization: false
+        intrinsic_scaling: false
+        one_sample_validation: null
+        print_log_every_n_steps: 10
+        scene_scale_reg_loss: 0.01
+        train_aux: true
+        train_gs_num: 1
+        train_target_set: true
+        use_refine_aux: false
+        verbose: false
+        vggt_cam_loss: true
+        vggt_distil: false
+trainer:
+    value:
+        gradient_clip_val: 0.5
+        max_steps: 3001
+        num_nodes: 1
+        val_check_interval: 250
+wandb:
+    value:
+        entity: scene-representation-group
+        mode: online
+        name: ABLATION_0225_OURS
+        project: DCSplat
+        tags:
+            - re10k
+            - 256x256
diff --git a/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/output.log b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/output.log
new file mode 100644
index 0000000000000000000000000000000000000000..602bfe2d03b4320c8a5158e0c7f2713e8c50aedf
--- /dev/null
+++ b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/output.log
@@ -0,0 +1,800 @@
+LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0,1,2,3,4,5,6,7]
+
+  | Name                   | Type                 | Params | Mode
+------------------------------------------------------------------------
+0 | encoder                | OurSplat             | 888 M  | train
+1 | density_control_module | DensityControlModule | 2.6 M  | train
+2 | decoder                | DecoderSplattingCUDA | 0      | train
+3 | render_losses          | ModuleList           | 0      | train
+4 | density_control_losses | ModuleList           | 0      | train
+5 | direct_losses          | ModuleList           | 0      | train
+------------------------------------------------------------------------
+891 M     Trainable params
+0         Non-trainable params
+891 M     Total params
+3,564.328 Total estimated model params size (MB)
+1231      Modules in train mode
+522       Modules in eval mode
+Sanity Checking: |                                                                                             | 0/? [00:00<?, ?it/s][2026-02-24 19:15:14,654][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/data_connector.py:425: The 'val_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=31` in the `DataLoader` to improve performance.
+
+[2026-02-24 19:15:14,655][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user.
+  warnings.warn(  # warn only once
+
+Validation epoch start on rank 0
+Sanity Checking DataLoader 0:   0%|                                                                            | 0/1 [00:00<?, ?it/s]validation step 0; scene = ['306e2b7785657539'];
+target intrinsic:  tensor(0.8595, device='cuda:0') tensor(0.8597, device='cuda:0')
+pred intrinsic:  tensor(0.8779, device='cuda:0') tensor(0.8773, device='cuda:0')
+[rank0]:W0224 19:15:17.139000 90349 site-packages/torch/utils/cpp_extension.py:2425] TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation.
+[rank0]:W0224 19:15:17.139000 90349 site-packages/torch/utils/cpp_extension.py:2425] If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'] to specific architectures.
+[2026-02-24 19:15:17,207][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-24 19:15:17,216][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/utilities/data.py:79: Trying to infer the `batch_size` from an ambiguous collection. The batch size we found is 1. To avoid any miscalculations, use `self.log(..., batch_size=batch_size)`.
+
+Setting up [LPIPS] perceptual loss: trunk [vgg], v[0.1], spatial [off]
+[2026-02-24 19:15:17,217][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-24 19:15:17,217][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+Loading model from: /venv/main/lib/python3.12/site-packages/lpips/weights/v0.1/vgg.pth
+[2026-02-24 19:15:18,865][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/functional.py:554: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /pytorch/aten/src/ATen/native/TensorShape.cpp:4322.)
+  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
+
+Sanity Checking DataLoader 0: 100%|████████████████████████████████████████████████████████████████████| 1/1 [00:03<00:00,  0.25it/s][2026-02-24 19:15:19,148][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/psnr', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-24 19:15:19,149][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/lpips', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-24 19:15:19,150][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/ssim', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-24 19:15:19,150][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/gaussian_num_ratio', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-24 19:15:19,150][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('info/global_step', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+Epoch 0: |                                                                                                     | 0/? [00:00<?, ?it/s]context = [[34, 36, 50, 53, 54, 60, 63, 70, 76, 78, 79, 80, 81, 88, 92, 94, 102, 110, 112, 114, 122, 125, 126, 131]]target = [[126, 96, 109, 55, 99, 116, 43, 60, 113, 85, 103, 90, 130, 62, 76, 123, 35, 102, 125, 128, 98, 67, 129, 79]]
+[2026-02-24 19:15:28,655][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [256, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [256, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-24 19:15:28,731][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                 | 9/? [00:41<00:00,  0.22it/s, v_num=5b6z]train step 10; scene = [['08c26703c4987851']]; loss = 0.932576
+Epoch 0: |                                                                                | 10/? [00:45<00:00,  0.22it/s, v_num=5b6z]context = [[98, 107, 112, 119, 122, 123], [21, 22, 27, 36, 38, 46], [63, 66, 67, 77, 84, 88], [56, 62, 65, 68, 78, 81]]target = [[105, 110, 116, 112, 111, 101], [29, 37, 36, 22, 40, 45], [73, 66, 75, 87, 83, 64], [79, 69, 75, 58, 61, 62]]
+Epoch 0: |                                                                                | 19/? [01:18<00:00,  0.24it/s, v_num=5b6z]train step 20; scene = [['4012c15c8381568b'], ['af08406c5a9a43a0'], ['9f9f9beffb86fad7'], ['fc8d08df6c875cb0']]; loss = 0.228420
+Epoch 0: |                                                                                | 20/? [01:22<00:00,  0.24it/s, v_num=5b6z]context = [[144, 152, 157, 164, 166, 169, 171, 177], [201, 211, 216, 221, 228, 230, 233, 234], [11, 15, 16, 23, 30, 38, 43, 44]]target = [[153, 170, 149, 169, 145, 174, 165, 157], [229, 216, 205, 206, 203, 213, 233, 215], [37, 38, 39, 35, 15, 24, 19, 25]]
+Epoch 0: |                                                                                | 24/? [01:36<00:00,  0.25it/s, v_num=5b6z][2026-02-24 19:17:03,331][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+Epoch 0: |                                                                                | 29/? [01:55<00:00,  0.25it/s, v_num=5b6z]train step 30; scene = [['00980329a3221f1c'], ['1e7c432d2207b6f2'], ['af2748330e5243d0']]; loss = 0.182995
+Epoch 0: |                                                                                | 30/? [01:59<00:00,  0.25it/s, v_num=5b6z]context = [[2, 7, 10, 15, 20, 24, 25, 29, 30, 32, 33, 37, 40, 44, 49, 60, 64, 74, 76, 81, 89, 91, 95, 99]]target = [[79, 84, 35, 43, 89, 44, 63, 58, 48, 13, 65, 7, 96, 27, 51, 20, 60, 15, 85, 59, 22, 66, 17, 62]]
+Epoch 0: |                                                                                | 39/? [02:32<00:00,  0.26it/s, v_num=5b6z]train step 40; scene = [['79a9385753d426bc'], ['593538382d2dc847'], ['c9c67636b9d521be']]; loss = 0.152928
+Epoch 0: |                                                                                | 40/? [02:36<00:00,  0.26it/s, v_num=5b6z]context = [[200, 208, 223, 225], [6, 8, 12, 31], [104, 107, 111, 129], [69, 80, 86, 95], [11, 15, 18, 37], [52, 60, 70, 78]]target = [[213, 203, 214, 201], [29, 11, 15, 28], [111, 122, 120, 115], [77, 90, 89, 78], [18, 24, 29, 14], [64, 67, 63, 74]]
+Epoch 0: |                                                                                | 49/? [03:09<00:00,  0.26it/s, v_num=5b6z]train step 50; scene = [['579a11551b3315d9'], ['c9dd64b7415e788e'], ['6f3fb517d1798d03']]; loss = 0.140180
+Epoch 0: |                                                                                | 50/? [03:13<00:00,  0.26it/s, v_num=5b6z]context = [[20, 22, 25, 28, 36, 47], [10, 13, 15, 22, 26, 35], [31, 47, 49, 51, 52, 56], [109, 110, 115, 116, 125, 135]]target = [[40, 24, 23, 30, 37, 33], [19, 32, 25, 24, 26, 28], [35, 52, 50, 55, 45, 36], [131, 123, 130, 110, 115, 134]]
+Epoch 0: |                                                                                | 59/? [03:46<00:00,  0.26it/s, v_num=5b6z]train step 60; scene = [['07916b8004a8e336'], ['e51ef9945ae527c4'], ['db84f84b1d775bb8'], ['92ed61f8e16b7e67']]; loss = 0.138675
+Epoch 0: |                                                                                | 60/? [03:50<00:00,  0.26it/s, v_num=5b6z]context = [[96, 97, 101, 102, 105, 106, 107, 114, 124, 134, 138, 145], [76, 78, 82, 83, 96, 98, 100, 111, 112, 118, 119, 125]]target = [[144, 124, 137, 102, 107, 130, 119, 129, 118, 123, 133, 126], [98, 108, 96, 100, 113, 123, 77, 78, 112, 109, 80, 102]]
+Epoch 0: |                                                                                | 69/? [04:24<00:00,  0.26it/s, v_num=5b6z]train step 70; scene = [['c34efa1505a0cfaa'], ['a3d0cca9fb57fd85'], ['43d0e6dce7bb1e95'], ['d8c2f0a3734cb493']]; loss = 0.109652
+Epoch 0: |                                                                                | 70/? [04:27<00:00,  0.26it/s, v_num=5b6z]context = [[204, 208, 211, 218, 232, 234, 237, 239, 240, 241, 247, 253], [6, 7, 10, 12, 35, 36, 37, 38, 40, 45, 48, 55]]target = [[230, 236, 222, 227, 220, 252, 216, 221, 213, 217, 249, 240], [18, 24, 25, 12, 9, 42, 34, 52, 41, 33, 43, 48]]
+Epoch 0: |                                                                                | 79/? [05:00<00:00,  0.26it/s, v_num=5b6z]train step 80; scene = [['24d756c820744e31'], ['cd6c21656a85e9b9'], ['f3b24cf238154fc0']]; loss = 0.113916
+Epoch 0: |                                                                                | 80/? [05:04<00:00,  0.26it/s, v_num=5b6z]context = [[4, 30], [52, 79], [61, 87], [12, 40], [83, 109], [3, 29], [221, 249], [198, 227], [9, 38], [46, 72], [0, 26], [123, 150]]target = [[27, 26], [58, 57], [70, 77], [26, 20], [87, 98], [26, 14], [223, 224], [221, 200], [23, 33], [67, 64], [4, 25], [140, 148]]
+Epoch 0: |                                                                                | 89/? [05:39<00:00,  0.26it/s, v_num=5b6z]train step 90; scene = [['617b4bc98d7e0bb6'], ['666e4a9aba27bb64']]; loss = 0.084713
+Epoch 0: |                                                                                | 90/? [05:43<00:00,  0.26it/s, v_num=5b6z]context = [[134, 135, 138, 146, 147, 151, 156, 161, 162, 164, 166, 167, 168, 169, 187, 189, 197, 210, 215, 224, 225, 228, 230, 231]]target = [[143, 149, 191, 182, 151, 226, 165, 140, 208, 171, 179, 223, 168, 136, 194, 207, 227, 144, 187, 185, 145, 218, 139, 170]]
+Epoch 0: |                                                                                | 99/? [06:17<00:00,  0.26it/s, v_num=5b6z]train step 100; scene = [['12fee7f1978d52f1'], ['c963bb60939e2d81']]; loss = 0.118093
+Epoch 0: |                                                                               | 100/? [06:21<00:00,  0.26it/s, v_num=5b6z]context = [[40, 44, 47, 48, 49, 50, 56, 63, 77, 79, 88, 89], [21, 33, 34, 35, 36, 40, 43, 46, 47, 60, 62, 70]]target = [[58, 81, 76, 64, 68, 72, 51, 87, 77, 65, 88, 45], [31, 47, 25, 35, 55, 22, 48, 65, 29, 40, 63, 67]]
+Epoch 0: |                                                                               | 109/? [06:54<00:00,  0.26it/s, v_num=5b6z]train step 110; scene = [['47396d5a5299873e']]; loss = 0.134408
+Epoch 0: |                                                                               | 110/? [06:58<00:00,  0.26it/s, v_num=5b6z]context = [[20, 22, 29, 30, 31, 34, 37, 45, 47, 51, 56, 69], [18, 22, 24, 27, 31, 37, 44, 48, 49, 60, 61, 67]]target = [[49, 22, 36, 59, 63, 60, 45, 66, 38, 28, 26, 64], [61, 65, 25, 20, 63, 60, 26, 22, 33, 45, 37, 35]]
+Epoch 0: |                                                                               | 119/? [07:33<00:00,  0.26it/s, v_num=5b6z]train step 120; scene = [['9bd7044e7cbf8e60'], ['76e44cf6b5658b26']]; loss = 0.076970
+Epoch 0: |                                                                               | 120/? [07:37<00:00,  0.26it/s, v_num=5b6z]context = [[7, 14, 26, 28, 31, 34, 37, 40], [10, 20, 26, 32, 35, 39, 41, 43], [17, 23, 26, 28, 44, 46, 49, 50]]target = [[14, 26, 10, 22, 30, 13, 31, 11], [24, 31, 36, 14, 40, 34, 41, 13], [24, 44, 34, 48, 46, 36, 26, 38]]
+Epoch 0: |                                                                               | 129/? [08:10<00:00,  0.26it/s, v_num=5b6z]train step 130; scene = [['a8cef6a851fbea3c'], ['b6699f4d039a5b06'], ['55cf2bbe9e017ea4'], ['6b0dd861e1ab1fec'], ['14db202c335af709'], ['8b6ff6c5153a7794'], ['b75f3820760d835c'], ['f7dbc855fd2a7669'], ['cfb20f8971e6a591'], ['95f2be7bb8303f50'], ['ff422469e034ae11'], ['5a2ad43377e9d18d']]; loss = 0.125457
+Epoch 0: |                                                                               | 130/? [08:14<00:00,  0.26it/s, v_num=5b6z]context = [[8, 15, 26, 30, 32, 35, 40, 46, 47, 48, 50, 53, 54, 62, 69, 70, 72, 76, 80, 86, 90, 99, 100, 105]]target = [[26, 12, 49, 100, 89, 24, 10, 81, 37, 63, 52, 17, 39, 70, 16, 56, 40, 55, 43, 34, 72, 28, 48, 45]]
+Epoch 0: |                                                                               | 139/? [08:47<00:00,  0.26it/s, v_num=5b6z]train step 140; scene = [['f62a962df5c26a1a'], ['b076420679a04731']]; loss = 0.083262
+Epoch 0: |                                                                               | 140/? [08:51<00:00,  0.26it/s, v_num=5b6z]context = [[90, 95, 97, 104, 118, 121], [14, 29, 34, 35, 36, 44], [12, 14, 22, 27, 34, 41], [6, 10, 16, 18, 34, 35]]target = [[99, 104, 93, 113, 110, 96], [29, 37, 22, 21, 39, 41], [19, 25, 39, 27, 29, 13], [26, 28, 9, 7, 13, 34]]
+Epoch 0: |                                                                               | 149/? [09:25<00:00,  0.26it/s, v_num=5b6z]train step 150; scene = [['a52d26a78b04aebd']]; loss = 0.074416
+Epoch 0: |                                                                               | 150/? [09:29<00:00,  0.26it/s, v_num=5b6z]context = [[115, 132, 134, 145], [16, 32, 39, 44], [57, 63, 78, 88], [9, 11, 23, 36], [15, 26, 28, 44], [1, 8, 22, 33]]target = [[124, 132, 116, 119], [27, 24, 33, 29], [81, 61, 85, 79], [28, 26, 29, 16], [28, 29, 37, 22], [9, 4, 32, 27]]
+Epoch 0: |                                                                               | 159/? [10:03<00:00,  0.26it/s, v_num=5b6z]train step 160; scene = [['268fbffc6c479d5b']]; loss = 0.071149
+Epoch 0: |                                                                               | 160/? [10:06<00:00,  0.26it/s, v_num=5b6z]context = [[18, 25, 26, 37, 42, 46, 49, 51, 53, 64, 65, 67], [69, 75, 78, 79, 82, 84, 94, 95, 104, 108, 117, 118]]target = [[53, 27, 22, 32, 41, 38, 50, 43, 47, 48, 23, 19], [74, 70, 114, 115, 90, 89, 88, 92, 94, 110, 107, 101]]
+Epoch 0: |                                                                               | 169/? [10:41<00:00,  0.26it/s, v_num=5b6z]train step 170; scene = [['719e2e8912e4eed3'], ['a3e51565a737569f']]; loss = 0.121969
+Epoch 0: |                                                                               | 170/? [10:45<00:00,  0.26it/s, v_num=5b6z]context = [[14, 18, 20, 21, 24, 27, 33, 40, 45, 47, 48, 51, 52, 60, 64, 70, 75, 77, 80, 85, 90, 98, 102, 111]]target = [[29, 93, 32, 39, 81, 108, 72, 107, 51, 35, 16, 36, 70, 18, 34, 92, 94, 47, 23, 74, 50, 77, 19, 37]]
+Epoch 0: |                                                                               | 179/? [11:18<00:00,  0.26it/s, v_num=5b6z]train step 180; scene = [['f44b9aa76a94a0a3']]; loss = 0.077210
+Epoch 0: |                                                                               | 180/? [11:22<00:00,  0.26it/s, v_num=5b6z]context = [[0, 6, 17, 22, 26, 28, 33, 41, 50, 55, 56, 57, 71, 76, 79, 81, 84, 85, 86, 87, 89, 95, 96, 97]]target = [[37, 49, 12, 78, 9, 16, 84, 13, 5, 4, 6, 38, 80, 51, 43, 68, 64, 46, 56, 24, 25, 72, 36, 21]]
+Epoch 0: |                                                                               | 189/? [11:57<00:00,  0.26it/s, v_num=5b6z]train step 190; scene = [['71bb669d936a5718'], ['a47203cfd5e0a478'], ['4b009f82cf5c7098']]; loss = 0.079976
+Epoch 0: |                                                                               | 190/? [12:01<00:00,  0.26it/s, v_num=5b6z]context = [[9, 11, 19, 25, 33, 43, 46, 47, 48, 54, 57, 58, 63, 70, 72, 75, 79, 80, 83, 84, 85, 96, 99, 106]]target = [[30, 53, 82, 31, 55, 12, 20, 72, 104, 70, 24, 52, 21, 32, 102, 71, 35, 11, 46, 10, 15, 74, 33, 26]]
+Epoch 0: |                                                                               | 199/? [12:34<00:00,  0.26it/s, v_num=5b6z]train step 200; scene = [['dd5ec950a01c42a0'], ['6d0db0358f7e051e'], ['983fe650a925ec1b']]; loss = 0.090733
+Epoch 0: |                                                                               | 200/? [12:38<00:00,  0.26it/s, v_num=5b6z]context = [[8, 10, 14, 15, 22, 23, 27, 30, 36, 38, 39, 47, 64, 65, 67, 78, 79, 80, 83, 86, 93, 96, 98, 105]]target = [[73, 63, 10, 27, 89, 35, 44, 58, 97, 71, 17, 24, 66, 87, 50, 12, 23, 11, 31, 45, 69, 96, 98, 94]]
+[2026-02-24 19:28:04,865][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                               | 209/? [13:19<00:00,  0.26it/s, v_num=5b6z]train step 210; scene = [['9be9b273b3c22c61'], ['4b5883872c9b860c']]; loss = 0.078428
+Epoch 0: |                                                                               | 210/? [13:22<00:00,  0.26it/s, v_num=5b6z]context = [[34, 35, 37, 54, 56, 59, 61, 68, 69, 76, 90, 96, 98, 101, 103, 107, 116, 119, 120, 121, 122, 125, 126, 131]]target = [[95, 50, 119, 121, 84, 107, 72, 52, 80, 42, 127, 94, 79, 98, 46, 128, 73, 75, 106, 92, 37, 110, 96, 56]]
+Epoch 0: |                                                                               | 219/? [13:57<00:00,  0.26it/s, v_num=5b6z]train step 220; scene = [['a3b6faa8d238d993'], ['df9ba36fbe753843']]; loss = 0.069444
+Epoch 0: |                                                                               | 220/? [14:01<00:00,  0.26it/s, v_num=5b6z]context = [[39, 71, 74], [21, 41, 51], [28, 55, 59], [2, 31, 37], [15, 25, 48], [57, 64, 93], [76, 87, 105], [31, 53, 64]]target = [[52, 43, 59], [28, 25, 49], [48, 58, 51], [26, 9, 6], [30, 32, 47], [84, 88, 59], [104, 87, 88], [42, 52, 59]]
+Epoch 0: |                                                                               | 229/? [14:36<00:00,  0.26it/s, v_num=5b6z]train step 230; scene = [['ca04de3c55cd1ca0'], ['3d90d586b33daa63'], ['d1772c09b4b6d95f'], ['03d05f69a1cab4f8'], ['60d296908f43a97a'], ['37c400e282bc481e']]; loss = 0.080689
+Epoch 0: |                                                                               | 230/? [14:40<00:00,  0.26it/s, v_num=5b6z]context = [[203, 204, 208, 209, 210, 233], [41, 42, 43, 52, 60, 76], [2, 8, 9, 23, 29, 31], [70, 74, 92, 97, 98, 100]]target = [[226, 223, 210, 217, 228, 222], [46, 70, 71, 75, 43, 56], [24, 23, 29, 25, 3, 21], [95, 81, 74, 73, 98, 92]]
+Epoch 0: |                                                                               | 239/? [15:14<00:00,  0.26it/s, v_num=5b6z]train step 240; scene = [['9794641b7e015578']]; loss = 0.128111
+Epoch 0: |                                                                               | 240/? [15:18<00:00,  0.26it/s, v_num=5b6z]context = [[62, 65, 68, 71, 80, 86, 87, 96, 99, 101, 103, 111], [134, 138, 139, 142, 151, 163, 171, 172, 173, 174, 181, 183]]target = [[87, 89, 96, 93, 103, 71, 65, 77, 63, 98, 102, 105], [165, 177, 159, 147, 138, 152, 171, 141, 181, 146, 161, 176]]
+Epoch 0: |                                                                               | 249/? [15:53<00:00,  0.26it/s, v_num=5b6z]train step 250; scene = [['93dff1b985f2c7f9']]; loss = 0.084227
+Epoch 0: |                                                                               | 250/? [15:57<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                  | 0/? [00:00<?, ?it/s]validation step 250; scene = ['49b8f80c849dc341'];
+target intrinsic:  tensor(0.8891, device='cuda:0') tensor(0.8894, device='cuda:0')                             | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8711, device='cuda:0') tensor(0.8713, device='cuda:0')
+[2026-02-24 19:31:20,418][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                               | 250/? [15:58<00:00,  0.26it/s, v_num=5b6z]context = [[113, 115, 120, 131, 135, 137, 144, 152], [16, 20, 23, 28, 32, 35, 39, 52], [14, 18, 25, 26, 28, 29, 51, 53]]target = [[149, 139, 118, 151, 137, 141, 121, 130], [45, 51, 28, 50, 35, 37, 27, 23], [29, 26, 22, 51, 24, 34, 47, 45]]
+Epoch 0: |                                                                               | 259/? [16:31<00:00,  0.26it/s, v_num=5b6z]train step 260; scene = [['b2288bf7003d5d4d']]; loss = 0.082914
+Epoch 0: |                                                                               | 260/? [16:35<00:00,  0.26it/s, v_num=5b6z]context = [[206, 208, 212, 216, 221, 236], [58, 60, 68, 69, 70, 93], [15, 19, 26, 34, 41, 44], [15, 18, 23, 34, 42, 44]]target = [[225, 216, 234, 214, 207, 232], [88, 76, 91, 67, 64, 70], [20, 29, 22, 36, 23, 40], [42, 28, 31, 30, 16, 37]]
+Epoch 0: |                                                                               | 269/? [17:09<00:00,  0.26it/s, v_num=5b6z]train step 270; scene = [['013ec74a4fde6737'], ['78e816776b064fc4'], ['1b778f72bbee1f27'], ['c71549de92ecb2e4'], ['8e16c8644efeec52'], ['35c5fc80e85db7cd'], ['34c8c62d878eca66'], ['203a5fd3a45ac4a7']]; loss = 0.070059
+Epoch 0: |                                                                               | 270/? [17:13<00:00,  0.26it/s, v_num=5b6z]context = [[160, 174, 177, 178, 192, 200], [34, 52, 54, 56, 58, 67], [78, 85, 101, 102, 109, 118], [123, 135, 136, 144, 150, 155]]target = [[169, 195, 175, 174, 194, 189], [55, 53, 57, 48, 40, 66], [90, 108, 110, 84, 109, 83], [131, 130, 127, 124, 139, 133]]
+Epoch 0: |                                                                               | 279/? [17:48<00:00,  0.26it/s, v_num=5b6z]train step 280; scene = [['75335793f866b96d'], ['e9d9dc952f5bbd83']]; loss = 0.046761
+Epoch 0: |                                                                               | 280/? [17:52<00:00,  0.26it/s, v_num=5b6z]context = [[2, 31], [39, 73], [69, 98], [58, 92], [109, 146], [27, 59], [83, 114], [199, 231], [13, 50], [77, 109], [72, 105], [7, 45]]target = [[20, 26], [71, 57], [92, 77], [61, 72], [125, 129], [44, 51], [110, 101], [211, 202], [27, 17], [98, 104], [80, 89], [36, 37]]
+Epoch 0: |                                                                               | 289/? [18:25<00:00,  0.26it/s, v_num=5b6z]train step 290; scene = [['51252022ddf74fb9'], ['8dd73309b133b8bf'], ['9e8db62a9b3cbd5e'], ['d41e59ee023e977b'], ['ce1a9465dc08ef4c'], ['e7887dec76685627']]; loss = 0.067491
+Epoch 0: |                                                                               | 290/? [18:29<00:00,  0.26it/s, v_num=5b6z]context = [[74, 83, 85, 88, 89, 93, 100, 101, 108, 119, 122, 123], [17, 20, 23, 24, 28, 30, 32, 35, 48, 52, 64, 66]]target = [[110, 79, 95, 92, 111, 109, 99, 89, 77, 98, 81, 102], [28, 29, 52, 36, 54, 64, 49, 23, 40, 58, 31, 27]]
+Epoch 0: |                                                                               | 299/? [19:03<00:00,  0.26it/s, v_num=5b6z]train step 300; scene = [['0c5d83212982c0ec'], ['00793a8a3b268d7c'], ['47a9b1e96499a466'], ['a1fb990016d7b3af']]; loss = 0.072660
+Epoch 0: |                                                                               | 300/? [19:07<00:00,  0.26it/s, v_num=5b6z]context = [[83, 88, 89, 91, 92, 100, 102, 104, 107, 111, 113, 115, 129, 132, 134, 135, 140, 142, 145, 150, 154, 173, 175, 180]]target = [[115, 140, 130, 169, 114, 163, 175, 177, 104, 152, 151, 89, 154, 123, 132, 85, 122, 119, 155, 91, 117, 171, 158, 157]]
+Epoch 0: |                                                                                                                                                        | 309/? [19:41<00:00,  0.26it/s, v_num=5b6z]train step 310; scene = [['9b73ab94b5c43711'], ['8c845b940aa8244c'], ['b2789c1a5c127a02'], ['3db6c0e172d18826']]; loss = 0.084836
+Epoch 0: |                                                                                                                                                        | 310/? [19:45<00:00,  0.26it/s, v_num=5b6z]context = [[30, 47, 66], [16, 31, 46], [10, 28, 40], [43, 68, 75], [21, 43, 63], [15, 29, 46], [36, 44, 66], [17, 39, 59]]target = [[51, 41, 31], [29, 43, 44], [23, 35, 33], [64, 71, 63], [48, 59, 27], [45, 29, 27], [60, 42, 50], [34, 26, 47]]
+Epoch 0: |                                                                                                                                                        | 319/? [20:19<00:00,  0.26it/s, v_num=5b6z]train step 320; scene = [['591cd9d079cd7842'], ['3dd7802a2c93a865']]; loss = 0.085465
+Epoch 0: |                                                                                                                                                        | 320/? [20:23<00:00,  0.26it/s, v_num=5b6z]context = [[4, 22, 28, 31, 34, 37, 45, 46], [26, 28, 38, 41, 46, 55, 57, 59], [129, 139, 145, 150, 151, 158, 167, 169]]target = [[7, 41, 13, 5, 20, 27, 32, 44], [29, 32, 54, 28, 40, 52, 43, 33], [137, 143, 146, 152, 130, 131, 145, 133]]
+Epoch 0: |                                                                                                                                                        | 329/? [20:58<00:00,  0.26it/s, v_num=5b6z]train step 330; scene = [['30d9f6321281dade'], ['2a08fac923c9e50d']]; loss = 0.062580
+Epoch 0: |                                                                                                                                                        | 330/? [21:02<00:00,  0.26it/s, v_num=5b6z]context = [[4, 6, 12, 14, 24, 41, 42, 44, 45, 48, 49, 53], [102, 106, 112, 113, 118, 122, 125, 130, 134, 140, 149, 151]]target = [[50, 42, 7, 28, 11, 33, 20, 45, 25, 8, 22, 16], [135, 134, 113, 128, 139, 112, 117, 108, 122, 127, 138, 120]]
+Epoch 0: |                                                                                                                                                        | 339/? [21:36<00:00,  0.26it/s, v_num=5b6z]train step 340; scene = [['bd9f2096d355b1b8'], ['07d3325178e7a790'], ['8204d757ce43dda8']]; loss = 0.060256
+Epoch 0: |                                                                                                                                                        | 340/? [21:40<00:00,  0.26it/s, v_num=5b6z]context = [[87, 88, 93, 105, 107, 111, 116, 118, 123, 131, 132, 136], [104, 111, 116, 121, 122, 127, 131, 132, 136, 137, 142, 153]]target = [[114, 108, 107, 130, 95, 129, 118, 104, 116, 96, 109, 99], [127, 106, 145, 129, 114, 109, 143, 150, 111, 133, 146, 122]]
+Epoch 0: |                                                                                                                                                        | 349/? [22:14<00:00,  0.26it/s, v_num=5b6z]train step 350; scene = [['9d0bfbe5b7f98545'], ['06a16655c8e8ad9c']]; loss = 0.096433
+Epoch 0: |                                                                                                                                                        | 350/? [22:18<00:00,  0.26it/s, v_num=5b6z]context = [[33, 52, 53, 56, 58, 64, 73, 75], [100, 102, 115, 117, 130, 134, 139, 142], [223, 225, 240, 242, 255, 256, 263, 267]]target = [[41, 60, 53, 63, 72, 38, 37, 73], [122, 141, 116, 138, 110, 119, 118, 135], [260, 237, 261, 264, 238, 255, 243, 253]]
+Epoch 0: |                                                                                                                                                        | 359/? [22:52<00:00,  0.26it/s, v_num=5b6z]train step 360; scene = [['73b27f4f150327af'], ['169aaaf51ef3849c'], ['068a8406f1a383d8'], ['a9936b77895f33b3']]; loss = 0.064843
+Epoch 0: |                                                                                                                                                        | 360/? [22:56<00:00,  0.26it/s, v_num=5b6z]context = [[0, 14, 18, 23, 25, 32, 42, 45, 46, 49, 53, 55, 59, 60, 61, 74, 78, 80, 84, 85, 89, 90, 92, 97]]target = [[50, 81, 69, 64, 38, 86, 91, 16, 27, 33, 44, 90, 48, 23, 8, 79, 6, 39, 42, 36, 82, 78, 59, 54]]
+Epoch 0: |                                                                                                                                                        | 369/? [23:29<00:00,  0.26it/s, v_num=5b6z]train step 370; scene = [['8673faf0a9d48165'], ['99a0790d72e6c2af'], ['6cbbe9075b0d2138']]; loss = 0.061921
+Epoch 0: |                                                                                                                                                        | 370/? [23:33<00:00,  0.26it/s, v_num=5b6z]context = [[61, 63, 65, 72, 73, 74, 76, 89, 92, 94, 100, 101, 120, 124, 126, 127, 136, 140, 144, 145, 150, 152, 156, 158]]target = [[121, 100, 147, 98, 148, 143, 126, 63, 73, 141, 79, 119, 115, 106, 153, 101, 120, 71, 91, 62, 105, 84, 151, 145]]
+Epoch 0: |                                                                                                                                                        | 379/? [24:07<00:00,  0.26it/s, v_num=5b6z]train step 380; scene = [['656330f47c5df010'], ['6dfb89a98e14ca66']]; loss = 0.064836
+Epoch 0: |                                                                                                                                                        | 380/? [24:11<00:00,  0.26it/s, v_num=5b6z]context = [[210, 215, 223, 225, 231, 234, 235, 249], [28, 32, 35, 53, 54, 56, 57, 68], [149, 150, 154, 155, 174, 177, 181, 183]]target = [[241, 236, 218, 230, 223, 228, 239, 227], [55, 46, 44, 49, 30, 33, 60, 45], [157, 160, 161, 164, 170, 181, 168, 166]]
+Epoch 0: |                                                                                                                                                        | 389/? [24:46<00:00,  0.26it/s, v_num=5b6z]train step 390; scene = [['723f94d150ab09f2'], ['393cdfb7e832d285'], ['14900b71ac66b7bd'], ['452625cd6b071b87'], ['281599bbab3e73dd'], ['0a2b42e240751d33']]; loss = 0.069923
+Epoch 0: |                                                                                                                                                        | 390/? [24:50<00:00,  0.26it/s, v_num=5b6z]context = [[48, 52, 80], [14, 47, 60], [142, 159, 187], [78, 120, 121], [6, 24, 45], [47, 68, 90], [15, 19, 48], [9, 20, 52]]target = [[51, 50, 70], [45, 15, 56], [172, 148, 157], [120, 98, 109], [20, 33, 11], [83, 84, 73], [45, 30, 32], [16, 36, 45]]
+Epoch 0: |                                                                                                                                                        | 399/? [25:22<00:00,  0.26it/s, v_num=5b6z]train step 400; scene = [['4303746d8f23f16b'], ['0fe8246bb7e2fe40'], ['b7d77240852d6a52'], ['6e5505414fd63528'], ['44985936f68c3a36'], ['1550f1b4fff1f2a4'], ['cea3d842c3285c65'], ['b34bb5f53856d34f']]; loss = 0.096210
+Epoch 0: |                                                                                                                                                        | 400/? [25:26<00:00,  0.26it/s, v_num=5b6z]context = [[9, 20, 28, 31, 34, 53], [189, 196, 216, 223, 230, 232], [74, 77, 86, 88, 98, 118], [132, 145, 157, 168, 179, 181]]target = [[51, 32, 14, 50, 45, 34], [212, 226, 201, 205, 215, 222], [93, 113, 107, 94, 83, 79], [161, 139, 177, 151, 156, 172]]
+[2026-02-24 19:40:53,015][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 409/? [26:01<00:00,  0.26it/s, v_num=5b6z]train step 410; scene = [['144e1ec915e46d29'], ['b290b6a0afa1dac7'], ['b3d84dba6581c3d9']]; loss = 0.058120
+Epoch 0: |                                                                                                                                                        | 410/? [26:05<00:00,  0.26it/s, v_num=5b6z]context = [[26, 40, 41, 50, 55, 58], [159, 174, 179, 192, 198, 201], [2, 3, 5, 7, 24, 40], [187, 204, 205, 208, 217, 221]]target = [[38, 32, 44, 52, 42, 27], [198, 191, 196, 162, 163, 169], [33, 17, 23, 37, 15, 39], [214, 211, 195, 220, 210, 192]]
+Epoch 0: |                                                                                                                                                        | 419/? [26:39<00:00,  0.26it/s, v_num=5b6z]train step 420; scene = [['a1dff9c50d92dc9c']]; loss = 0.065388
+Epoch 0: |                                                                                                                                                        | 420/? [26:43<00:00,  0.26it/s, v_num=5b6z]context = [[11, 13, 14, 26, 27, 41, 46, 51], [151, 155, 161, 167, 172, 176, 191, 195], [49, 52, 55, 57, 59, 74, 79, 90]]target = [[25, 44, 17, 33, 37, 24, 38, 40], [193, 191, 194, 162, 158, 176, 163, 171], [87, 85, 67, 58, 86, 68, 83, 78]]
+Epoch 0: |                                                                                                                                                        | 429/? [27:17<00:00,  0.26it/s, v_num=5b6z]train step 430; scene = [['36664e22fd10a141'], ['0474328f4cefd619']]; loss = 0.047884
+Epoch 0: |                                                                                                                                                        | 430/? [27:20<00:00,  0.26it/s, v_num=5b6z]context = [[123, 130, 131, 132, 136, 154, 169, 173], [35, 36, 47, 49, 65, 67, 76, 86], [33, 34, 39, 41, 59, 64, 65, 73]]target = [[145, 147, 167, 130, 146, 170, 138, 137], [59, 38, 49, 42, 70, 76, 39, 40], [56, 44, 36, 35, 38, 45, 48, 58]]
+Epoch 0: |                                                                                                                                                        | 439/? [27:55<00:00,  0.26it/s, v_num=5b6z]train step 440; scene = [['342099a48847f4f6'], ['5ad0327426e3718b'], ['c25b314716aa6b10'], ['c91e2b5399b14430'], ['e1d9ade67e615bd8'], ['46df912c9748215b']]; loss = 0.068825
+Epoch 0: |                                                                                                                                                        | 440/? [27:59<00:00,  0.26it/s, v_num=5b6z]context = [[0, 33], [10, 57], [36, 82], [46, 79], [62, 99], [33, 79], [11, 52], [1, 34], [0, 37], [26, 62], [18, 67], [96, 138]]target = [[22, 18], [37, 23], [44, 54], [69, 50], [68, 69], [69, 56], [49, 41], [27, 30], [31, 17], [42, 32], [59, 54], [108, 115]]
+Epoch 0: |                                                                                                                                                        | 449/? [28:34<00:00,  0.26it/s, v_num=5b6z]train step 450; scene = [['e19c6facac1c9624'], ['5244830b7357365b'], ['b80c2522b1070e2f'], ['6ea0ff32c8ea695c'], ['2f311b2bbbeb5940'], ['3f7992e72a096099']]; loss = 0.069304
+Epoch 0: |                                                                                                                                                        | 450/? [28:37<00:00,  0.26it/s, v_num=5b6z]context = [[41, 42, 44, 49, 57, 66, 68, 73, 75, 76, 83, 91], [4, 5, 15, 24, 32, 35, 37, 38, 45, 47, 49, 56]]target = [[79, 77, 90, 46, 81, 83, 56, 68, 89, 70, 84, 57], [7, 8, 40, 37, 32, 10, 25, 35, 23, 43, 48, 52]]
+Epoch 0: |                                                                                                                                                        | 459/? [29:12<00:00,  0.26it/s, v_num=5b6z]train step 460; scene = [['46fb6702ed1b9967'], ['bdc3f978b0d3aa8f']]; loss = 0.054985
+Epoch 0: |                                                                                                                                                        | 460/? [29:16<00:00,  0.26it/s, v_num=5b6z]context = [[65, 66, 69, 78, 84, 85, 90, 91, 99, 103, 112, 116], [39, 40, 52, 63, 64, 68, 72, 73, 83, 87, 90, 92]]target = [[99, 86, 101, 76, 84, 77, 105, 66, 89, 75, 100, 106], [84, 61, 68, 82, 59, 47, 51, 71, 85, 56, 57, 50]]
+Epoch 0: |                                                                                                                                                        | 469/? [29:48<00:00,  0.26it/s, v_num=5b6z]train step 470; scene = [['2c88995e05a17d17'], ['2b1f47da224557a3'], ['62216d162b71b5b4'], ['61d39a97cb69d99f'], ['42000d5a83b48ee4'], ['cc8480640599f9f3']]; loss = 0.069207
+Epoch 0: |                                                                                                                                                        | 470/? [29:52<00:00,  0.26it/s, v_num=5b6z]context = [[1, 6, 15, 26, 28, 41, 46, 52], [0, 1, 3, 5, 14, 26, 32, 37], [52, 56, 58, 79, 87, 88, 93, 97]]target = [[24, 22, 32, 14, 42, 21, 12, 37], [14, 10, 11, 25, 12, 34, 7, 32], [62, 60, 80, 87, 63, 55, 78, 88]]
+Epoch 0: |                                                                                                                                                        | 479/? [30:27<00:00,  0.26it/s, v_num=5b6z]train step 480; scene = [['2e3bb7fb33e1ed30'], ['7460f503eb18fa6a'], ['bde49071d2088850'], ['e80016be3043dfa4']]; loss = 0.079180
+Epoch 0: |                                                                                                                                                        | 480/? [30:30<00:00,  0.26it/s, v_num=5b6z]context = [[9, 15, 37, 47], [0, 5, 7, 35], [101, 113, 134, 153], [64, 88, 93, 107], [18, 44, 56, 62], [31, 46, 58, 75]]target = [[45, 27, 28, 18], [10, 8, 14, 3], [128, 143, 110, 129], [87, 84, 90, 94], [29, 31, 40, 36], [52, 72, 33, 56]]
+Epoch 0: |                                                                                                                                                        | 489/? [31:05<00:00,  0.26it/s, v_num=5b6z]train step 490; scene = [['83085493f4bc18d2']]; loss = 0.078027
+Epoch 0: |                                                                                                                                                        | 490/? [31:09<00:00,  0.26it/s, v_num=5b6z]context = [[0, 3, 7, 12, 38, 40, 45, 48], [25, 28, 57, 62, 69, 70, 76, 78], [25, 34, 37, 39, 41, 53, 65, 67]]target = [[15, 7, 39, 47, 42, 26, 25, 5], [63, 40, 59, 31, 75, 34, 47, 39], [58, 35, 53, 33, 65, 44, 38, 31]]
+Epoch 0: |                                                                                                                                                        | 499/? [31:43<00:00,  0.26it/s, v_num=5b6z]train step 500; scene = [['1241bcb5732a9502'], ['d33a9e90e1416efb']]; loss = 0.045151
+Epoch 0: |                                                                                                                                                        | 500/? [31:47<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 500; scene = ['73d6f935f31b3fd4'];
+target intrinsic:  tensor(0.8576, device='cuda:0') tensor(0.8579, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8699, device='cuda:0') tensor(0.8683, device='cuda:0')
+[2026-02-24 19:47:10,684][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 500/? [31:48<00:00,  0.26it/s, v_num=5b6z]context = [[1, 21, 23, 29, 39, 48, 49, 52], [17, 18, 20, 28, 34, 39, 45, 51], [99, 103, 106, 110, 116, 136, 138, 140]]target = [[4, 46, 41, 43, 40, 37, 39, 25], [30, 48, 22, 42, 44, 31, 38, 37], [112, 108, 133, 116, 111, 125, 123, 127]]
+Epoch 0: |                                                                                                                                                        | 509/? [32:21<00:00,  0.26it/s, v_num=5b6z]train step 510; scene = [['0eed4548041bea8e'], ['277a96ce456580f4']]; loss = 0.058767
+Epoch 0: |                                                                                                                                                        | 510/? [32:25<00:00,  0.26it/s, v_num=5b6z]context = [[3, 13, 21, 24, 27, 28, 30, 37, 42, 44, 46, 55, 62, 65, 69, 71, 73, 80, 83, 84, 89, 95, 97, 100]]target = [[51, 81, 54, 89, 55, 45, 10, 19, 21, 68, 48, 67, 33, 69, 8, 25, 64, 83, 37, 29, 24, 82, 99, 46]]
+Epoch 0: |                                                                                                                                                        | 519/? [32:59<00:00,  0.26it/s, v_num=5b6z]train step 520; scene = [['625e3aa0ff734714'], ['395802511d26f32e'], ['39343936591c28de']]; loss = 0.080314
+Epoch 0: |                                                                                                                                                        | 520/? [33:03<00:00,  0.26it/s, v_num=5b6z]context = [[13, 16, 17, 22, 25, 30, 35, 39, 53, 56, 62, 68, 78, 87, 90, 92, 93, 95, 99, 100, 103, 106, 107, 110]]target = [[39, 96, 58, 28, 16, 59, 44, 17, 26, 83, 103, 31, 57, 35, 107, 51, 27, 77, 46, 30, 100, 91, 93, 97]]
+Epoch 0: |                                                                                                                                                        | 529/? [33:37<00:00,  0.26it/s, v_num=5b6z]train step 530; scene = [['0c199c575b699444'], ['70d878da47f984e4'], ['15f77c76ea744f99'], ['e54b5eec8cc47776'], ['1969ed97e68d83d9'], ['c7cf9b63dc3e5830'], ['bcef3076b93012b1'], ['ab2680bf91942e23']]; loss = 0.095884
+Epoch 0: |                                                                                                                                                        | 530/? [33:41<00:00,  0.26it/s, v_num=5b6z]context = [[78, 84, 85, 92, 124, 127], [14, 15, 24, 52, 64, 68], [1, 6, 16, 20, 22, 40], [33, 51, 61, 65, 66, 70]]target = [[104, 103, 94, 96, 88, 82], [15, 53, 58, 47, 56, 60], [31, 10, 18, 26, 23, 24], [37, 69, 62, 65, 49, 64]]
+Epoch 0: |                                                                                                                                                        | 539/? [34:15<00:00,  0.26it/s, v_num=5b6z]train step 540; scene = [['a071d9276f6a9272']]; loss = 0.070171
+Epoch 0: |                                                                                                                                                        | 540/? [34:19<00:00,  0.26it/s, v_num=5b6z]context = [[35, 46, 47, 49, 52, 54, 55, 56, 57, 65, 66, 85], [51, 54, 55, 59, 68, 77, 82, 86, 87, 93, 105, 106]]target = [[69, 68, 75, 48, 44, 79, 53, 74, 60, 81, 72, 71], [77, 73, 88, 59, 67, 61, 102, 62, 93, 75, 95, 105]]
+Epoch 0: |                                                                                                                                                        | 549/? [34:53<00:00,  0.26it/s, v_num=5b6z]train step 550; scene = [['836250796ea45b6c']]; loss = 0.085105
+Epoch 0: |                                                                                                                                                        | 550/? [34:57<00:00,  0.26it/s, v_num=5b6z]context = [[30, 34, 39, 45, 47, 79, 80, 83], [16, 21, 22, 24, 33, 42, 51, 60], [115, 128, 129, 138, 141, 150, 156, 163]]target = [[79, 51, 62, 77, 46, 54, 33, 49], [27, 59, 20, 41, 21, 48, 58, 30], [133, 125, 124, 159, 122, 129, 155, 119]]
+Epoch 0: |                                                                                                                                                        | 559/? [35:32<00:00,  0.26it/s, v_num=5b6z]train step 560; scene = [['d70ca840b3c5aec9'], ['65c3f29c43dd1e63'], ['d3917d0a1eda2a1f'], ['5c83dfc8f9ab44fa']]; loss = 0.067866
+Epoch 0: |                                                                                                                                                        | 560/? [35:36<00:00,  0.26it/s, v_num=5b6z]context = [[62, 66, 67, 68, 71, 87, 88, 92, 100, 108, 110, 122], [153, 160, 165, 168, 171, 177, 178, 181, 183, 185, 200, 205]]target = [[95, 104, 83, 68, 115, 81, 74, 106, 119, 79, 90, 72], [204, 165, 196, 185, 197, 198, 195, 187, 155, 188, 154, 201]]
+Epoch 0: |                                                                                                                                                        | 569/? [36:10<00:00,  0.26it/s, v_num=5b6z]train step 570; scene = [['9d8ddcdbe1f7ac42'], ['721df0f45094ca34'], ['fdbfe35f5940d3ad']]; loss = 0.051634
+Epoch 0: |                                                                                                                                                        | 570/? [36:14<00:00,  0.26it/s, v_num=5b6z]context = [[23, 34, 41, 42, 48, 54, 56, 59, 60, 63, 66, 71, 74, 79, 88, 90, 92, 93, 95, 97, 110, 111, 119, 120]]target = [[30, 27, 28, 53, 31, 75, 85, 54, 77, 111, 76, 25, 44, 52, 33, 41, 69, 89, 73, 68, 26, 93, 83, 119]]
+Epoch 0: |                                                                                                                                                        | 579/? [36:49<00:00,  0.26it/s, v_num=5b6z]train step 580; scene = [['88a0267e41b851f0'], ['df71fbb70b19cbc3'], ['1c713c10ecf5a0c9']]; loss = 0.069671
+Epoch 0: |                                                                                                                                                        | 580/? [36:52<00:00,  0.26it/s, v_num=5b6z]context = [[9, 22, 36, 39, 43, 68], [70, 85, 95, 98, 116, 121], [10, 28, 45, 47, 50, 57], [132, 144, 154, 158, 168, 188]]target = [[67, 45, 11, 46, 50, 30], [114, 94, 71, 104, 117, 76], [22, 12, 24, 51, 50, 13], [180, 169, 183, 173, 182, 171]]
+Epoch 0: |                                                                                                                                                        | 589/? [37:27<00:00,  0.26it/s, v_num=5b6z]train step 590; scene = [['3f732b63cdd0729e'], ['9be3165beb073d95'], ['42a6c835ff830674'], ['f928d960cbfae15a'], ['140b10a4f6bb5aa5'], ['cc8e19c8ad1846f4']]; loss = 0.082504
+Epoch 0: |                                                                                                                                                        | 590/? [37:30<00:00,  0.26it/s, v_num=5b6z]context = [[198, 200, 201, 204, 209, 210, 224, 234], [6, 14, 18, 23, 25, 27, 58, 59], [13, 29, 40, 51, 55, 66, 70, 71]]target = [[203, 208, 227, 217, 222, 223, 200, 211], [11, 46, 45, 19, 20, 18, 47, 7], [16, 51, 40, 15, 55, 44, 23, 61]]
+Epoch 0: |                                                                                                                                                        | 599/? [38:05<00:00,  0.26it/s, v_num=5b6z]train step 600; scene = [['cb734fdc69e9900e']]; loss = 0.055922
+Epoch 0: |                                                                                                                                                        | 600/? [38:09<00:00,  0.26it/s, v_num=5b6z]context = [[10, 16, 24, 31, 33, 34, 36, 41, 49, 59, 63, 67], [160, 162, 167, 172, 173, 180, 183, 187, 189, 205, 212, 214]]target = [[32, 54, 42, 14, 39, 60, 66, 33, 62, 22, 47, 49], [182, 199, 163, 179, 203, 200, 174, 209, 164, 172, 194, 167]]
+[2026-02-24 19:53:35,645][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 609/? [38:44<00:00,  0.26it/s, v_num=5b6z]train step 610; scene = [['ed9409fa128e193b'], ['a5c03b0c5fb7203e']]; loss = 0.043273
+Epoch 0: |                                                                                                                                                        | 610/? [38:48<00:00,  0.26it/s, v_num=5b6z]context = [[36, 37, 39, 50, 51, 52, 57, 59, 72, 73, 86, 94], [45, 46, 54, 57, 66, 75, 77, 81, 83, 89, 97, 101]]target = [[47, 37, 89, 79, 92, 78, 81, 90, 38, 56, 84, 69], [84, 91, 85, 69, 79, 53, 55, 48, 87, 49, 92, 63]]
+Epoch 0: |                                                                                                                                                        | 619/? [39:23<00:00,  0.26it/s, v_num=5b6z]train step 620; scene = [['7898a828b7203ca4'], ['9c269fce78f0dd27'], ['e1e317857deb7afc'], ['30124191dafb3383'], ['c39f1a9a73797efe'], ['a640a55439a43108']]; loss = 0.057952
+Epoch 0: |                                                                                                                                                        | 620/? [39:27<00:00,  0.26it/s, v_num=5b6z]context = [[2, 15, 16, 21, 26, 29, 47, 56], [6, 12, 15, 37, 45, 47, 50, 68], [9, 18, 23, 26, 39, 48, 50, 60]]target = [[38, 7, 32, 36, 16, 21, 53, 22], [50, 67, 28, 43, 15, 41, 44, 23], [12, 54, 16, 24, 58, 32, 30, 47]]
+Epoch 0: |                                                                                                                                                        | 629/? [40:01<00:00,  0.26it/s, v_num=5b6z]train step 630; scene = [['dd3bbf1f7f832e83'], ['0ff7277275e55096'], ['5f45c360d76a3b12']]; loss = 0.049226
+Epoch 0: |                                                                                                                                                        | 630/? [40:05<00:00,  0.26it/s, v_num=5b6z]context = [[29, 42, 44, 45, 76, 81, 87, 93], [77, 84, 94, 96, 100, 102, 121, 125], [67, 68, 70, 89, 90, 91, 96, 107]]target = [[83, 55, 36, 89, 45, 62, 90, 63], [106, 92, 114, 99, 117, 82, 119, 107], [99, 68, 75, 92, 79, 85, 101, 90]]
+Epoch 0: |                                                                                                                                                        | 639/? [40:39<00:00,  0.26it/s, v_num=5b6z]train step 640; scene = [['867edbda9bb8ef59'], ['1d83764e77e159d8'], ['e318dafa4071cef9'], ['169f09c33ee35289']]; loss = 0.091717
+Epoch 0: |                                                                                                                                                        | 640/? [40:43<00:00,  0.26it/s, v_num=5b6z]context = [[52, 57, 67, 75, 77, 80, 85, 86, 99, 105, 112, 117], [6, 11, 22, 23, 32, 39, 42, 47, 48, 53, 63, 64]]target = [[114, 60, 111, 110, 57, 69, 68, 80, 62, 99, 77, 90], [27, 15, 46, 56, 51, 60, 41, 63, 49, 16, 9, 43]]
+Epoch 0: |                                                                                                                                                        | 649/? [41:17<00:00,  0.26it/s, v_num=5b6z]train step 650; scene = [['23668135f32e0126'], ['daca15248046e480'], ['174ebd189316bd92']]; loss = 0.049818
+Epoch 0: |                                                                                                                                                        | 650/? [41:21<00:00,  0.26it/s, v_num=5b6z]context = [[6, 23, 25, 26, 28, 32, 37, 38, 39, 40, 50, 57], [0, 2, 3, 11, 17, 19, 24, 25, 38, 40, 45, 53]]target = [[15, 47, 23, 31, 12, 30, 20, 55, 33, 11, 22, 10], [51, 9, 13, 34, 52, 30, 26, 45, 38, 27, 11, 14]]
+Epoch 0: |                                                                                                                                                        | 659/? [41:55<00:00,  0.26it/s, v_num=5b6z]train step 660; scene = [['60499200285c9abe']]; loss = 0.044941
+Epoch 0: |                                                                                                                                                        | 660/? [41:58<00:00,  0.26it/s, v_num=5b6z]context = [[57, 65, 76, 86, 93, 99, 100, 122], [31, 34, 42, 47, 60, 80, 81, 83], [13, 16, 19, 30, 33, 49, 50, 55]]target = [[85, 75, 88, 114, 107, 70, 116, 79], [82, 35, 67, 56, 68, 61, 74, 58], [54, 44, 23, 36, 31, 39, 45, 28]]
+Epoch 0: |                                                                                                                                                        | 669/? [42:32<00:00,  0.26it/s, v_num=5b6z]train step 670; scene = [['7665ff641f430aa5']]; loss = 0.044916
+Epoch 0: |                                                                                                                                                        | 670/? [42:36<00:00,  0.26it/s, v_num=5b6z]context = [[50, 57, 61, 69, 82, 87, 98, 101], [26, 33, 44, 47, 53, 56, 62, 66], [11, 33, 39, 48, 56, 67, 75, 78]]target = [[67, 96, 71, 75, 73, 57, 55, 89], [65, 39, 63, 59, 60, 47, 55, 41], [18, 56, 21, 64, 32, 59, 33, 25]]
+Epoch 0: |                                                                                                                                                        | 679/? [43:10<00:00,  0.26it/s, v_num=5b6z]train step 680; scene = [['43c939b11c5fed4a']]; loss = 0.072316
+Epoch 0: |                                                                                                                                                        | 680/? [43:14<00:00,  0.26it/s, v_num=5b6z]context = [[47, 54, 60, 64, 79, 80, 83, 93], [12, 13, 21, 30, 34, 45, 53, 55], [57, 75, 78, 82, 96, 103, 113, 117]]target = [[80, 62, 75, 54, 57, 92, 53, 58], [38, 32, 15, 44, 42, 51, 27, 39], [81, 87, 106, 94, 99, 103, 73, 79]]
+Epoch 0: |                                                                                                                                                        | 689/? [43:47<00:00,  0.26it/s, v_num=5b6z]train step 690; scene = [['1848b8b363d0d2b9'], ['afe6b05d0554a880']]; loss = 0.057546
+Epoch 0: |                                                                                                                                                        | 690/? [43:51<00:00,  0.26it/s, v_num=5b6z]context = [[12, 18, 20, 27, 36, 37, 38, 43, 48, 49, 52, 58, 59, 62, 67, 81, 83, 86, 93, 95, 101, 104, 108, 109]]target = [[34, 24, 43, 31, 87, 30, 51, 54, 52, 94, 86, 21, 44, 97, 61, 95, 38, 60, 49, 73, 41, 19, 65, 67]]
+Epoch 0: |                                                                                                                                                        | 699/? [44:25<00:00,  0.26it/s, v_num=5b6z]train step 700; scene = [['674ef9fb9cf20f9f'], ['8624ee0839cb6e4c'], ['caed302f388b799f']]; loss = 0.052923
+Epoch 0: |                                                                                                                                                        | 700/? [44:28<00:00,  0.26it/s, v_num=5b6z]context = [[31, 32, 41, 46, 53, 54, 55, 57, 65, 68, 73, 74, 80, 85, 100, 105, 108, 109, 113, 114, 116, 118, 126, 128]]target = [[117, 52, 85, 57, 37, 45, 78, 100, 125, 35, 113, 66, 105, 103, 61, 83, 88, 40, 116, 60, 79, 32, 102, 107]]
+Epoch 0: |                                                                                                                                                        | 709/? [45:03<00:00,  0.26it/s, v_num=5b6z]train step 710; scene = [['db6cd90de8fee2ff'], ['7a20ba81fb778529'], ['970350268b239272']]; loss = 0.060537
+Epoch 0: |                                                                                                                                                        | 710/? [45:07<00:00,  0.26it/s, v_num=5b6z]context = [[3, 12, 16, 31, 41, 45, 49, 53, 54, 55, 67, 69], [208, 211, 217, 221, 222, 227, 230, 231, 239, 247, 252, 267]]target = [[38, 34, 68, 25, 44, 63, 48, 7, 42, 6, 64, 28], [228, 253, 259, 266, 244, 238, 249, 234, 232, 241, 220, 213]]
+Epoch 0: |                                                                                                                                                        | 719/? [45:41<00:00,  0.26it/s, v_num=5b6z]train step 720; scene = [['f63d2df8871ce70c'], ['0fdeda15097ed4a4']]; loss = 0.054131
+Epoch 0: |                                                                                                                                                        | 720/? [45:45<00:00,  0.26it/s, v_num=5b6z]context = [[55, 57, 71, 72, 75, 85, 90, 92, 98, 99, 102, 108, 112, 116, 125, 126, 130, 131, 135, 141, 144, 145, 146, 152]]target = [[61, 145, 86, 66, 62, 119, 100, 139, 105, 125, 58, 101, 140, 132, 118, 128, 65, 141, 151, 78, 104, 107, 150, 138]]
+Epoch 0: |                                                                                                                                                        | 729/? [46:20<00:00,  0.26it/s, v_num=5b6z]train step 730; scene = [['232abb354c423e81'], ['d34926c73ae1277e']]; loss = 0.041493
+Epoch 0: |                                                                                                                                                        | 730/? [46:23<00:00,  0.26it/s, v_num=5b6z]context = [[45, 50, 54, 58, 66, 70, 72, 78, 85, 88, 94, 105, 109, 110, 111, 117, 120, 122, 126, 127, 128, 133, 135, 142]]target = [[95, 96, 50, 114, 135, 124, 104, 100, 49, 119, 139, 62, 92, 123, 58, 46, 57, 112, 116, 90, 54, 101, 85, 81]]
+Epoch 0: |                                                                                                                                                        | 739/? [46:57<00:00,  0.26it/s, v_num=5b6z]train step 740; scene = [['19f7966006ad778d'], ['dde0212418df7ca9'], ['ad75e36b74f6b033'], ['ea97e5ae55e56208'], ['9d29b0289133ab4e'], ['282938f90821bdef']]; loss = 0.109286
+Epoch 0: |                                                                                                                                                        | 740/? [47:01<00:00,  0.26it/s, v_num=5b6z]context = [[8, 13, 19, 22, 30, 31, 36, 45, 46, 48, 54, 57, 58, 59, 60, 61, 65, 73, 77, 83, 86, 88, 93, 105]]target = [[46, 101, 90, 61, 31, 23, 37, 95, 18, 67, 32, 100, 93, 35, 89, 45, 10, 70, 60, 97, 85, 81, 66, 79]]
+Epoch 0: |                                                                                                                                                        | 749/? [47:35<00:00,  0.26it/s, v_num=5b6z]train step 750; scene = [['f85921f42c5d98d7'], ['a95dacbd3ea3db36']]; loss = 0.046105
+Epoch 0: |                                                                                                                                                        | 750/? [47:39<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 750; scene = ['91fda69e1cda4602'];
+target intrinsic:  tensor(0.8937, device='cuda:0') tensor(0.8939, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9013, device='cuda:0') tensor(0.8987, device='cuda:0')
+[2026-02-24 20:03:02,783][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 750/? [47:41<00:00,  0.26it/s, v_num=5b6z]context = [[1, 38, 46, 49, 52, 59], [213, 218, 227, 233, 252, 257], [99, 105, 109, 124, 125, 153], [8, 17, 20, 37, 62, 65]]target = [[54, 19, 39, 57, 7, 4], [252, 219, 223, 216, 248, 237], [148, 136, 102, 103, 149, 132], [12, 52, 13, 44, 63, 39]]
+Epoch 0: |                                                                                                                                                        | 759/? [48:14<00:00,  0.26it/s, v_num=5b6z]train step 760; scene = [['75617c97bff1e873'], ['ff02f88545dfa566']]; loss = 0.036503
+Epoch 0: |                                                                                                                                                        | 760/? [48:18<00:00,  0.26it/s, v_num=5b6z]context = [[71, 73, 111], [2, 14, 56], [29, 82, 88], [120, 144, 168], [181, 214, 250], [0, 10, 71], [12, 53, 54], [14, 27, 87]]target = [[101, 74, 84], [43, 51, 38], [82, 55, 72], [156, 139, 147], [198, 214, 207], [22, 5, 40], [36, 13, 39], [68, 35, 70]]
+Epoch 0: |                                                                                                                                                        | 769/? [48:52<00:00,  0.26it/s, v_num=5b6z]train step 770; scene = [['62b0d4ee613af70f'], ['f7926eb1096de201'], ['c63b37ec347f0d0e'], ['b43d9f7c70f5caa0']]; loss = 0.075437
+Epoch 0: |                                                                                                                                                        | 770/? [48:56<00:00,  0.26it/s, v_num=5b6z]context = [[13, 29, 30, 31, 32, 34, 41, 42, 46, 52, 63, 66], [146, 149, 157, 166, 177, 180, 181, 193, 197, 204, 211, 213]]target = [[35, 55, 52, 59, 24, 51, 63, 18, 42, 44, 33, 61], [183, 158, 171, 175, 163, 196, 177, 181, 210, 167, 187, 174]]
+Epoch 0: |                                                                                                                                                        | 779/? [49:30<00:00,  0.26it/s, v_num=5b6z]train step 780; scene = [['b41f4db8b8a42a71']]; loss = 0.084910
+Epoch 0: |                                                                                                                                                        | 780/? [49:34<00:00,  0.26it/s, v_num=5b6z]context = [[43, 50, 53, 55, 57, 58, 59, 60, 68, 81, 94, 96, 102, 106, 108, 110, 112, 121, 125, 126, 130, 131, 136, 140]]target = [[74, 122, 93, 98, 70, 84, 49, 136, 77, 117, 135, 138, 123, 89, 119, 45, 129, 105, 50, 58, 63, 103, 82, 121]]
+Epoch 0: |                                                                                                                                                        | 789/? [50:08<00:00,  0.26it/s, v_num=5b6z]train step 790; scene = [['d79666d294813d8e']]; loss = 0.155817
+Epoch 0: |                                                                                                                                                        | 790/? [50:12<00:00,  0.26it/s, v_num=5b6z]context = [[59, 79, 88, 93, 100, 107], [3, 16, 19, 31, 46, 51], [41, 49, 54, 56, 72, 83], [226, 229, 235, 244, 255, 272]]target = [[103, 98, 81, 75, 74, 72], [48, 44, 20, 31, 15, 46], [64, 82, 70, 50, 47, 60], [251, 249, 271, 250, 259, 260]]
+Epoch 0: |                                                                                                                                                        | 799/? [50:47<00:00,  0.26it/s, v_num=5b6z]train step 800; scene = [['cb797cd30542e55c']]; loss = 0.057307
+Epoch 0: |                                                                                                                                                        | 800/? [50:51<00:00,  0.26it/s, v_num=5b6z]context = [[10, 19, 20, 39, 42, 55], [1, 6, 11, 20, 30, 42], [0, 3, 4, 19, 42, 45], [21, 35, 52, 82, 85, 88]]target = [[38, 39, 26, 21, 37, 40], [11, 40, 18, 27, 3, 24], [43, 33, 8, 24, 5, 19], [83, 71, 66, 32, 63, 47]]
+[2026-02-24 20:06:17,398][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 809/? [51:25<00:00,  0.26it/s, v_num=5b6z]train step 810; scene = [['9bd08fc9288bef8b']]; loss = 0.068685
+Epoch 0: |                                                                                                                                                        | 810/? [51:29<00:00,  0.26it/s, v_num=5b6z]context = [[5, 6, 14, 17, 19, 23, 30, 37, 40, 49, 53, 55], [3, 6, 15, 17, 27, 34, 38, 40, 42, 46, 56, 65]]target = [[16, 8, 48, 23, 44, 49, 10, 7, 18, 50, 22, 15], [34, 18, 9, 42, 31, 38, 45, 26, 43, 6, 49, 44]]
+Epoch 0: |                                                                                                                                                        | 819/? [52:03<00:00,  0.26it/s, v_num=5b6z]train step 820; scene = [['49952f737be91dd2']]; loss = 0.080855
+Epoch 0: |                                                                                                                                                        | 820/? [52:07<00:00,  0.26it/s, v_num=5b6z]context = [[22, 23, 26, 28, 31, 43, 51, 53, 57, 65, 67, 68, 71, 76, 77, 78, 80, 84, 88, 90, 107, 112, 116, 119]]target = [[35, 45, 31, 72, 24, 71, 89, 54, 25, 61, 44, 97, 56, 47, 82, 87, 50, 110, 108, 75, 103, 67, 86, 117]]
+Epoch 0: |                                                                                                                                                        | 829/? [52:39<00:00,  0.26it/s, v_num=5b6z]train step 830; scene = [['2f9c9d1b56eb7f75'], ['1c392cc98b3a7642']]; loss = 0.063032
+Epoch 0: |                                                                                                                                                        | 830/? [52:43<00:00,  0.26it/s, v_num=5b6z]context = [[42, 67, 90, 119], [18, 54, 72, 90], [43, 67, 73, 102], [2, 22, 40, 43], [23, 40, 52, 67], [19, 37, 75, 78]]target = [[86, 112, 87, 94], [59, 40, 20, 47], [58, 97, 66, 77], [34, 19, 32, 17], [62, 37, 35, 54], [54, 22, 31, 50]]
+Epoch 0: |                                                                                                                                                        | 839/? [53:17<00:00,  0.26it/s, v_num=5b6z]train step 840; scene = [['c51c7bc0c8151abb'], ['a0d16e79ab441c4f']]; loss = 0.184030
+Epoch 0: |                                                                                                                                                        | 840/? [53:21<00:00,  0.26it/s, v_num=5b6z]context = [[1, 18, 22, 27, 29, 35, 38, 49], [3, 12, 30, 36, 43, 47, 54, 55], [17, 36, 55, 65, 72, 77, 86, 88]]target = [[5, 36, 48, 11, 2, 39, 35, 37], [6, 54, 43, 18, 30, 15, 10, 11], [52, 59, 56, 77, 25, 68, 22, 27]]
+Epoch 0: |                                                                                                                                                        | 849/? [53:55<00:00,  0.26it/s, v_num=5b6z]train step 850; scene = [['818df63ddd1cf294'], ['3002f9cbe7f00e6c'], ['8ec42c5dfea6823b'], ['64ae75e57c6aa0a4']]; loss = 0.119801
+Epoch 0: |                                                                                                                                                        | 850/? [53:59<00:00,  0.26it/s, v_num=5b6z]context = [[31, 75, 86], [20, 22, 73], [7, 55, 63], [3, 62, 68], [10, 43, 54], [52, 92, 111], [0, 34, 42], [51, 90, 102]]target = [[40, 57, 36], [39, 67, 34], [12, 9, 52], [29, 12, 39], [49, 13, 17], [106, 93, 85], [39, 20, 9], [79, 60, 53]]
+Epoch 0: |                                                                                                                                                        | 859/? [54:34<00:00,  0.26it/s, v_num=5b6z]train step 860; scene = [['eb0aa1a4fb58c50c']]; loss = 0.044347
+Epoch 0: |                                                                                                                                                        | 860/? [54:38<00:00,  0.26it/s, v_num=5b6z]context = [[147, 155, 165, 189, 194, 195], [9, 15, 44, 47, 49, 82], [5, 22, 23, 51, 61, 64], [97, 98, 140, 149, 160, 170]]target = [[194, 192, 155, 151, 174, 158], [22, 27, 54, 18, 81, 31], [25, 27, 29, 23, 6, 59], [137, 162, 98, 111, 152, 164]]
+Epoch 0: |                                                                                                                                                        | 869/? [55:12<00:00,  0.26it/s, v_num=5b6z]train step 870; scene = [['c357fbd8aca05570']]; loss = 0.051297
+Epoch 0: |                                                                                                                                                        | 870/? [55:16<00:00,  0.26it/s, v_num=5b6z]context = [[44, 45, 82, 100], [70, 93, 96, 147], [144, 176, 184, 205], [0, 51, 65, 67], [105, 106, 130, 152], [5, 6, 55, 60]]target = [[81, 66, 67, 97], [116, 118, 137, 91], [200, 193, 197, 176], [11, 13, 22, 32], [120, 117, 127, 130], [17, 14, 58, 27]]
+Epoch 0: |                                                                                                                                                        | 879/? [55:50<00:00,  0.26it/s, v_num=5b6z]train step 880; scene = [['c672fa3960b73528'], ['a60e4127f167ac93'], ['a9739ec3a34012af']]; loss = 0.062172
+Epoch 0: |                                                                                                                                                        | 880/? [55:54<00:00,  0.26it/s, v_num=5b6z]context = [[15, 85, 91], [0, 68, 74], [73, 85, 121], [4, 53, 80], [15, 36, 63], [42, 81, 92], [0, 45, 49], [65, 97, 132]]target = [[49, 21, 35], [47, 44, 69], [119, 89, 106], [57, 76, 6], [28, 32, 55], [62, 43, 49], [18, 44, 24], [90, 104, 71]]
+Epoch 0: |                                                                                                                                                        | 889/? [56:27<00:00,  0.26it/s, v_num=5b6z]train step 890; scene = [['40a3f4f9389dd20c'], ['b14ec6f019932d8d'], ['4b9ed7532c875dab'], ['10c36bd5ef5f5a6b'], ['bc9a64096787007d'], ['d58a26d24f2776b2'], ['46f2228076e6f3f7'], ['399668567ff33ad7']]; loss = 0.072246
+Epoch 0: |                                                                                                                                                        | 890/? [56:31<00:00,  0.26it/s, v_num=5b6z]context = [[32, 34, 35, 37, 39, 44, 52, 59, 73, 84, 87, 93], [20, 21, 22, 34, 43, 47, 52, 57, 64, 69, 72, 96]]target = [[87, 80, 41, 73, 35, 37, 86, 66, 72, 39, 50, 83], [28, 22, 50, 33, 26, 79, 75, 73, 29, 77, 60, 37]]
+Epoch 0: |                                                                                                                                                        | 899/? [57:05<00:00,  0.26it/s, v_num=5b6z]train step 900; scene = [['711ade236bebd618']]; loss = 0.075575
+Epoch 0: |                                                                                                                                                        | 900/? [57:09<00:00,  0.26it/s, v_num=5b6z]context = [[6, 7, 14, 15, 17, 19, 39, 43, 44, 45, 47, 48, 49, 53, 57, 61, 67, 69, 72, 73, 78, 90, 92, 103]]target = [[18, 19, 38, 66, 14, 96, 50, 55, 41, 57, 45, 83, 12, 100, 24, 53, 85, 40, 54, 58, 20, 44, 86, 95]]
+Epoch 0: |                                                                                                                                                        | 909/? [57:43<00:00,  0.26it/s, v_num=5b6z]train step 910; scene = [['6b495ce9634d2fbb']]; loss = 0.097048
+Epoch 0: |                                                                                                                                                        | 910/? [57:47<00:00,  0.26it/s, v_num=5b6z]context = [[59, 61, 62, 63, 73, 74, 77, 90, 100, 105, 111, 112], [8, 15, 18, 21, 26, 35, 37, 40, 41, 56, 68, 70]]target = [[76, 98, 64, 84, 69, 87, 100, 99, 67, 108, 77, 78], [22, 24, 41, 15, 45, 44, 34, 28, 13, 39, 11, 55]]
+Epoch 0: |                                                                                                                                                        | 919/? [58:21<00:00,  0.26it/s, v_num=5b6z]train step 920; scene = [['013264a550df794f'], ['4203a06d618eeb97']]; loss = 0.092523
+Epoch 0: |                                                                                                                                                        | 920/? [58:25<00:00,  0.26it/s, v_num=5b6z]context = [[14, 16, 18, 28, 30, 37, 45, 49, 50, 61, 63, 65], [97, 98, 114, 115, 121, 134, 137, 140, 143, 144, 169, 179]]target = [[38, 62, 50, 58, 42, 55, 57, 46, 43, 49, 56, 40], [158, 178, 130, 108, 145, 120, 168, 164, 156, 177, 125, 111]]
+Epoch 0: |                                                                                                                                                        | 929/? [58:58<00:00,  0.26it/s, v_num=5b6z]train step 930; scene = [['5747f1d12ad10026'], ['48c9bb29482ddf76'], ['a45bddb856f554a1']]; loss = 0.047039
+Epoch 0: |                                                                                                                                                        | 930/? [59:02<00:00,  0.26it/s, v_num=5b6z]context = [[0, 2, 12, 23, 26, 31, 35, 37, 41, 42, 44, 49], [164, 174, 179, 182, 184, 187, 189, 201, 202, 205, 207, 216]]target = [[33, 7, 10, 26, 44, 36, 39, 30, 43, 11, 38, 27], [214, 173, 176, 191, 182, 190, 175, 210, 215, 167, 195, 198]]
+Epoch 0: |                                                                                                                                                        | 939/? [59:37<00:00,  0.26it/s, v_num=5b6z]train step 940; scene = [['db02cd4ba6a027da'], ['8f5e074629cedd06']]; loss = 0.045134
+Epoch 0: |                                                                                                                                                        | 940/? [59:39<00:00,  0.26it/s, v_num=5b6z]context = [[127, 141, 154, 157, 160, 163, 172, 173, 189, 190, 192, 196], [6, 23, 36, 41, 46, 51, 57, 61, 65, 66, 83, 84]]target = [[157, 179, 153, 148, 161, 137, 194, 177, 128, 162, 166, 174], [38, 41, 32, 28, 8, 37, 36, 9, 77, 18, 80, 26]]
+Epoch 0: |                                                                                                                                                      | 949/? [1:00:14<00:00,  0.26it/s, v_num=5b6z]train step 950; scene = [['5b98e84e8e7ffef0'], ['d7ca47da5fac7140'], ['eff98653337775a8'], ['ba00608cd351deb0']]; loss = 0.049714
+Epoch 0: |                                                                                                                                                      | 950/? [1:00:18<00:00,  0.26it/s, v_num=5b6z]context = [[28, 33, 43, 46, 47, 51, 54, 67, 68, 72, 74, 75, 88, 90, 94, 98, 101, 102, 103, 105, 107, 113, 118, 125]]target = [[56, 59, 68, 66, 119, 65, 80, 100, 101, 39, 97, 94, 110, 62, 109, 40, 91, 42, 44, 78, 60, 73, 108, 50]]
+Epoch 0: |                                                                                                                                                      | 959/? [1:00:52<00:00,  0.26it/s, v_num=5b6z]train step 960; scene = [['f6d65c637ff68de3'], ['9b4d466924c40d8b'], ['b6c9aa729ebc703e']]; loss = 0.092930
+Epoch 0: |                                                                                                                                                      | 960/? [1:00:56<00:00,  0.26it/s, v_num=5b6z]context = [[2, 10, 11, 14, 19, 21, 23, 33, 39, 45, 49, 52, 53, 55, 58, 61, 71, 74, 80, 83, 84, 90, 94, 99]]target = [[41, 39, 14, 8, 52, 95, 19, 76, 24, 68, 75, 69, 22, 3, 47, 98, 17, 38, 89, 35, 21, 57, 45, 43]]
+Epoch 0: |                                                                                                                                                      | 969/? [1:01:30<00:00,  0.26it/s, v_num=5b6z]train step 970; scene = [['4b6cbf1f4c87d918']]; loss = 0.044434
+Epoch 0: |                                                                                                                                                      | 970/? [1:01:34<00:00,  0.26it/s, v_num=5b6z]context = [[20, 23, 30, 51, 53, 56, 60, 70], [145, 151, 184, 191, 192, 195, 215, 219], [36, 61, 63, 93, 106, 109, 111, 114]]target = [[52, 61, 27, 43, 44, 68, 67, 65], [210, 148, 165, 147, 166, 153, 170, 176], [108, 92, 76, 81, 71, 104, 41, 54]]
+Epoch 0: |                                                                                                                                                      | 979/? [1:02:08<00:00,  0.26it/s, v_num=5b6z]train step 980; scene = [['0492d6125268e9ae']]; loss = 0.053897
+Epoch 0: |                                                                                                                                                      | 980/? [1:02:12<00:00,  0.26it/s, v_num=5b6z]context = [[2, 4, 66, 70], [30, 96, 104, 105], [2, 13, 43, 65], [6, 42, 79, 91], [48, 85, 95, 96], [43, 62, 80, 101]]target = [[16, 66, 47, 53], [66, 42, 33, 84], [51, 25, 31, 53], [39, 90, 74, 16], [89, 80, 56, 68], [49, 45, 72, 48]]
+Epoch 0: |                                                                                                                                                      | 989/? [1:02:47<00:00,  0.26it/s, v_num=5b6z]train step 990; scene = [['b0413d361b4e8abb']]; loss = 0.035714
+Epoch 0: |                                                                                                                                                      | 990/? [1:02:51<00:00,  0.26it/s, v_num=5b6z]context = [[14, 19, 24, 29, 30, 38, 39, 40, 53, 57, 58, 65], [0, 1, 21, 33, 35, 38, 39, 43, 49, 58, 59, 64]]target = [[38, 37, 40, 27, 39, 45, 15, 52, 55, 22, 23, 31], [12, 42, 48, 30, 10, 58, 52, 59, 8, 46, 19, 44]]
+Epoch 0: |                                                                                                                                                      | 999/? [1:03:25<00:00,  0.26it/s, v_num=5b6z]train step 1000; scene = [['2bf91de5cd028c93'], ['2abe932fd9d76528'], ['8fabc39ad677dece'], ['188398a54205f797']]; loss = 0.063463
+Epoch 0: |                                                                                                                                                     | 1000/? [1:03:29<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1000; scene = ['647f2049bf4cb3f3'];
+target intrinsic:  tensor(0.8998, device='cuda:0') tensor(0.9001, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8811, device='cuda:0') tensor(0.8813, device='cuda:0')
+[2026-02-24 20:18:52,619][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1000/? [1:03:30<00:00,  0.26it/s, v_num=5b6z]context = [[176, 185, 186, 188, 208, 224, 228, 246], [0, 6, 9, 11, 37, 43, 46, 54], [0, 16, 23, 40, 41, 44, 55, 68]]target = [[205, 177, 207, 223, 188, 184, 231, 179], [7, 27, 9, 11, 39, 15, 12, 28], [38, 2, 54, 3, 25, 7, 37, 46]]
+[2026-02-24 20:18:56,264][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1009/? [1:04:06<00:00,  0.26it/s, v_num=5b6z]train step 1010; scene = [['3b42fa1245f6b00b'], ['c472581eb9a351cb'], ['222255ddcacd20cf'], ['b953af75bafccce8'], ['2c862f06019afc7e'], ['4178cb9ef3e0fd6d'], ['7f9480301fa3e38b'], ['8ccc9f1ead9cad5b']]; loss = 0.066614
+Epoch 0: |                                                                                                                                                     | 1010/? [1:04:09<00:00,  0.26it/s, v_num=5b6z]context = [[0, 73], [12, 74], [41, 117], [34, 98], [3, 58], [117, 195], [117, 198], [33, 101], [0, 71], [45, 102], [36, 110], [82, 153]]target = [[44, 27], [42, 20], [116, 46], [88, 73], [5, 49], [190, 129], [127, 171], [75, 70], [61, 20], [97, 100], [51, 43], [151, 139]]
+Epoch 0: |                                                                                                                                                     | 1019/? [1:04:43<00:00,  0.26it/s, v_num=5b6z]train step 1020; scene = [['3a95efa051965ee0']]; loss = 0.108705
+Epoch 0: |                                                                                                                                                     | 1020/? [1:04:47<00:00,  0.26it/s, v_num=5b6z]context = [[8, 11, 13, 14, 15, 20, 21, 31, 34, 43, 46, 50, 55, 61, 71, 75, 76, 77, 81, 83, 85, 87, 93, 105]]target = [[83, 40, 76, 65, 29, 87, 84, 91, 31, 73, 50, 22, 38, 57, 99, 20, 19, 24, 26, 82, 36, 34, 44, 15]]
+Epoch 0: |                                                                                                                                                     | 1029/? [1:05:21<00:00,  0.26it/s, v_num=5b6z]train step 1030; scene = [['6bd6d8270eee0b81'], ['2ac1cf1adda42447'], ['a017de2545c81e26'], ['4d5593ce15e317ed'], ['437188c182ae6a10'], ['86440d86f60644b7'], ['27508406fddac6dd'], ['7feae1e0e2c1701b'], ['a86dd7257f21bc73'], ['a025554e94654c2f'], ['20ee785f1bc6340d'], ['db6cd90de8fee2ff']]; loss = 0.110671
+Epoch 0: |                                                                                                                                                     | 1030/? [1:05:25<00:00,  0.26it/s, v_num=5b6z]context = [[84, 89, 90, 91, 93, 111, 112, 113, 116, 117, 121, 126, 128, 134, 138, 139, 145, 153, 159, 161, 169, 173, 178, 181]]target = [[147, 138, 152, 92, 157, 97, 115, 154, 98, 113, 142, 137, 177, 165, 158, 146, 120, 139, 136, 122, 104, 85, 128, 144]]
+Epoch 0: |                                                                                                                                                     | 1039/? [1:05:59<00:00,  0.26it/s, v_num=5b6z]train step 1040; scene = [['80fd51689742e978'], ['5eb07ee6a9e4fec2']]; loss = 0.054026
+Epoch 0: |                                                                                                                                                     | 1040/? [1:06:03<00:00,  0.26it/s, v_num=5b6z]context = [[98, 99, 101, 110, 114, 116, 123, 126, 131, 137, 139, 150, 153, 155, 156, 158, 165, 177, 178, 184, 185, 187, 192, 195]]target = [[134, 141, 100, 117, 190, 107, 162, 158, 179, 192, 121, 151, 157, 139, 118, 175, 156, 183, 127, 167, 103, 172, 154, 116]]
+Epoch 0: |                                                                                                                                                     | 1049/? [1:06:36<00:00,  0.26it/s, v_num=5b6z]train step 1050; scene = [['77dd1e7595b8f9a1']]; loss = 0.062179
+Epoch 0: |                                                                                                                                                     | 1050/? [1:06:40<00:00,  0.26it/s, v_num=5b6z]context = [[114, 115, 118, 119, 124, 126, 130, 147, 149, 151, 158, 166, 169, 170, 173, 174, 177, 180, 191, 192, 203, 206, 208, 211]]target = [[209, 128, 160, 132, 164, 207, 201, 145, 117, 163, 120, 202, 151, 119, 124, 162, 189, 133, 184, 198, 153, 116, 188, 193]]
+Epoch 0: |                                                                                                                                                     | 1059/? [1:07:14<00:00,  0.26it/s, v_num=5b6z]train step 1060; scene = [['51b6e63c903275b4'], ['2a9baa599c00613f'], ['1d7dc2d489714b09'], ['9f92d7e79d4a35bb']]; loss = 0.056958
+Epoch 0: |                                                                                                                                                     | 1060/? [1:07:18<00:00,  0.26it/s, v_num=5b6z]context = [[15, 17, 24, 33, 36, 40, 69, 74, 77, 82, 83, 84], [1, 5, 9, 16, 25, 29, 32, 33, 44, 47, 49, 51]]target = [[41, 42, 50, 25, 44, 31, 43, 24, 81, 18, 26, 71], [45, 3, 37, 11, 42, 29, 49, 41, 17, 22, 39, 23]]
+Epoch 0: |                                                                                                                                                     | 1069/? [1:07:53<00:00,  0.26it/s, v_num=5b6z]train step 1070; scene = [['a464ce7c8383dbbb']]; loss = 0.068758
+Epoch 0: |                                                                                                                                                     | 1070/? [1:07:57<00:00,  0.26it/s, v_num=5b6z]context = [[6, 10, 28, 32, 57, 71, 89, 92], [34, 52, 65, 75, 88, 89, 114, 117], [119, 123, 131, 133, 137, 160, 180, 190]]target = [[90, 59, 62, 24, 52, 29, 47, 35], [72, 39, 106, 85, 58, 74, 77, 40], [188, 149, 156, 166, 161, 120, 125, 158]]
+Epoch 0: |                                                                                                                                                     | 1079/? [1:08:31<00:00,  0.26it/s, v_num=5b6z]train step 1080; scene = [['b382af5f342061fa'], ['f6a0556897b15d6b']]; loss = 0.045315
+Epoch 0: |                                                                                                                                                     | 1080/? [1:08:34<00:00,  0.26it/s, v_num=5b6z]context = [[127, 150, 162, 189], [25, 41, 98, 102], [6, 17, 20, 68], [24, 48, 78, 100], [154, 155, 175, 235], [41, 55, 86, 109]]target = [[180, 163, 148, 162], [45, 57, 100, 82], [16, 24, 44, 47], [50, 42, 47, 92], [206, 210, 196, 211], [80, 89, 82, 62]]
+Epoch 0: |                                                                                                                                                     | 1089/? [1:09:09<00:00,  0.26it/s, v_num=5b6z]train step 1090; scene = [['5beb85aaf29d1242']]; loss = 0.072918
+Epoch 0: |                                                                                                                                                     | 1090/? [1:09:13<00:00,  0.26it/s, v_num=5b6z]context = [[27, 37, 53, 65, 72, 76, 79, 83, 86, 97, 104, 109], [10, 17, 18, 32, 38, 45, 46, 52, 67, 68, 72, 79]]target = [[103, 63, 65, 74, 40, 85, 44, 107, 57, 58, 53, 29], [18, 61, 49, 74, 69, 21, 73, 59, 30, 67, 11, 43]]
+Epoch 0: |                                                                                                                                                     | 1099/? [1:09:47<00:00,  0.26it/s, v_num=5b6z]train step 1100; scene = [['db811a2460c4f9b5'], ['7db8c4965bba509a'], ['a7a79393e5bb8108'], ['2d77b1dd90856337']]; loss = 0.078126
+Epoch 0: |                                                                                                                                                     | 1100/? [1:09:51<00:00,  0.26it/s, v_num=5b6z]context = [[1, 10, 15, 20, 27, 35, 38, 44, 51, 52, 53, 70], [40, 44, 45, 46, 52, 54, 56, 63, 88, 95, 96, 98]]target = [[18, 63, 10, 69, 55, 26, 40, 21, 59, 51, 42, 48], [84, 92, 51, 54, 77, 50, 59, 47, 86, 43, 89, 87]]
+Epoch 0: |                                                                                                                                                     | 1109/? [1:10:24<00:00,  0.26it/s, v_num=5b6z]train step 1110; scene = [['a815d5a5f2ec9562'], ['3be8c5ae6c95c9b4']]; loss = 0.042122
+Epoch 0: |                                                                                                                                                     | 1110/? [1:10:28<00:00,  0.26it/s, v_num=5b6z]context = [[0, 2, 11, 15, 25, 28, 29, 32, 33, 35, 40, 44, 47, 49, 55, 56, 66, 72, 74, 84, 86, 88, 96, 97]]target = [[15, 85, 74, 83, 43, 47, 21, 40, 39, 66, 3, 26, 29, 9, 38, 23, 94, 36, 62, 19, 12, 44, 76, 31]]
+Epoch 0: |                                                                                                                                                     | 1119/? [1:11:03<00:00,  0.26it/s, v_num=5b6z]train step 1120; scene = [['3ab1ce6779776017']]; loss = 0.091448
+Epoch 0: |                                                                                                                                                     | 1120/? [1:11:06<00:00,  0.26it/s, v_num=5b6z]context = [[9, 16, 17, 21, 23, 30, 31, 43, 55, 57, 59, 61], [11, 25, 51, 66, 72, 76, 82, 87, 91, 94, 97, 99]]target = [[56, 22, 30, 35, 59, 36, 38, 23, 34, 11, 26, 51], [19, 73, 37, 71, 50, 43, 17, 80, 25, 54, 28, 24]]
+Epoch 0: |                                                                                                                                                     | 1129/? [1:11:40<00:00,  0.26it/s, v_num=5b6z]train step 1130; scene = [['fd0ebd5afbfd1acf'], ['edf6636dfd51ba3c'], ['7b2c118f021e6902'], ['b3356e816a130b87'], ['78cbac3ff58f2e41'], ['94c654bd3e031bcb'], ['b281bf93286a0573'], ['d9dce3382830aea6'], ['070a524bacb9aa38'], ['cae0139a521aa052'], ['5645a008715acf0a'], ['23174a6cd65a0731']]; loss = 0.106756
+Epoch 0: |                                                                                                                                                     | 1130/? [1:11:44<00:00,  0.26it/s, v_num=5b6z]context = [[8, 50, 81, 84], [2, 42, 44, 54], [25, 26, 79, 87], [25, 30, 88, 92], [45, 100, 129, 135], [17, 68, 77, 83]]target = [[62, 80, 76, 55], [11, 29, 39, 13], [42, 27, 75, 58], [31, 63, 29, 48], [72, 58, 93, 106], [66, 77, 60, 59]]
+Epoch 0: |                                                                                                                                                     | 1139/? [1:12:18<00:00,  0.26it/s, v_num=5b6z]train step 1140; scene = [['090a038e9d844b4e'], ['d28a2455cc34badb']]; loss = 0.043574
+Epoch 0: |                                                                                                                                                     | 1140/? [1:12:21<00:00,  0.26it/s, v_num=5b6z]context = [[19, 20, 28, 32, 34, 36, 38, 40, 52, 54, 63, 65, 68, 77, 78, 87, 94, 98, 100, 105, 107, 113, 114, 116]]target = [[32, 23, 97, 28, 51, 21, 114, 76, 46, 84, 74, 31, 98, 108, 77, 82, 72, 49, 103, 47, 80, 39, 59, 70]]
+Epoch 0: |                                                                                                                                                     | 1149/? [1:12:56<00:00,  0.26it/s, v_num=5b6z]train step 1150; scene = [['82347d56a4a55c27'], ['ee03755cce11b682']]; loss = 0.070448
+Epoch 0: |                                                                                                                                                     | 1150/? [1:13:00<00:00,  0.26it/s, v_num=5b6z]context = [[13, 16, 50, 52, 55, 62], [55, 56, 79, 89, 111, 117], [35, 43, 73, 74, 82, 85], [30, 56, 58, 80, 102, 108]]target = [[39, 59, 30, 34, 42, 27], [91, 98, 71, 70, 84, 68], [42, 44, 39, 68, 69, 50], [42, 68, 72, 55, 87, 51]]
+Epoch 0: |                                                                                                                                                     | 1159/? [1:13:34<00:00,  0.26it/s, v_num=5b6z]train step 1160; scene = [['255998558abc7172']]; loss = 0.071248
+Epoch 0: |                                                                                                                                                     | 1160/? [1:13:38<00:00,  0.26it/s, v_num=5b6z]context = [[13, 16, 18, 19, 21, 25, 28, 29, 36, 38, 43, 47, 49, 50, 53, 56, 67, 68, 69, 71, 72, 93, 108, 110]]target = [[47, 34, 77, 80, 28, 94, 97, 19, 29, 26, 68, 40, 53, 20, 60, 30, 92, 70, 83, 48, 106, 15, 63, 41]]
+Epoch 0: |                                                                                                                                                     | 1169/? [1:14:12<00:00,  0.26it/s, v_num=5b6z]train step 1170; scene = [['771aa992eae9a574'], ['4f9716bb3dc7feec'], ['fbed2318ae410b31'], ['8e1b4054949b6a46']]; loss = 0.061591
+Epoch 0: |                                                                                                                                                     | 1170/? [1:14:16<00:00,  0.26it/s, v_num=5b6z]context = [[207, 214, 223, 228, 242, 250, 268, 269], [35, 58, 71, 96, 97, 109, 110, 114], [0, 27, 28, 46, 50, 60, 68, 75]]target = [[216, 210, 265, 253, 237, 223, 241, 234], [105, 106, 100, 96, 73, 51, 109, 93], [2, 54, 60, 70, 28, 22, 51, 62]]
+Epoch 0: |                                                                                                                                                     | 1179/? [1:14:51<00:00,  0.26it/s, v_num=5b6z]train step 1180; scene = [['18c880b4b5ef683e']]; loss = 0.070093
+Epoch 0: |                                                                                                                                                     | 1180/? [1:14:55<00:00,  0.26it/s, v_num=5b6z]context = [[48, 60, 70, 80, 90, 109, 110, 114], [3, 8, 11, 23, 28, 34, 62, 82], [14, 17, 19, 24, 30, 34, 59, 67]]target = [[85, 101, 84, 113, 106, 74, 87, 98], [81, 68, 48, 46, 67, 69, 56, 22], [31, 33, 45, 20, 17, 61, 37, 44]]
+Epoch 0: |                                                                                                                                                     | 1189/? [1:15:29<00:00,  0.26it/s, v_num=5b6z]train step 1190; scene = [['019fdd708d7163bd'], ['7046980b0d3d3c63'], ['f1af5d4039ce3a2c']]; loss = 0.052572
+Epoch 0: |                                                                                                                                                     | 1190/? [1:15:32<00:00,  0.26it/s, v_num=5b6z]context = [[0, 7, 8, 11, 15, 20, 21, 28, 30, 33, 38, 42, 47, 54, 59, 63, 66, 71, 73, 79, 81, 88, 90, 97]]target = [[83, 9, 22, 57, 21, 76, 37, 10, 7, 13, 43, 47, 19, 6, 51, 71, 95, 78, 24, 92, 30, 28, 55, 84]]
+Epoch 0: |                                                                                                                                                     | 1199/? [1:16:07<00:00,  0.26it/s, v_num=5b6z]train step 1200; scene = [['5070ea042c65de0d'], ['19f950b6900d6176'], ['02768fad99be3290'], ['ed6c9a5913622c3d'], ['0a4cbe699be68e5c'], ['0892e76375b283ba'], ['efd001b0d5127d61'], ['9adb745c741c85e0'], ['e4774e728791dc20'], ['23710dc8de8b4a49'], ['347d7c1f3516f732'], ['97283dc038203c65']]; loss = 0.115725
+Epoch 0: |                                                                                                                                                     | 1200/? [1:16:11<00:00,  0.26it/s, v_num=5b6z]context = [[44, 48, 49, 54, 63, 66, 67, 70, 76, 77, 78, 84, 90, 100, 101, 105, 106, 111, 113, 116, 120, 123, 125, 141]]target = [[63, 111, 132, 61, 108, 104, 110, 81, 59, 89, 131, 126, 73, 100, 134, 74, 121, 84, 80, 51, 140, 133, 55, 83]]
+[2026-02-24 20:31:37,459][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1209/? [1:16:48<00:00,  0.26it/s, v_num=5b6z]train step 1210; scene = [['7ba6259f378c70f8']]; loss = 0.044792
+Epoch 0: |                                                                                                                                                     | 1210/? [1:16:52<00:00,  0.26it/s, v_num=5b6z]context = [[77, 82, 86, 95, 101, 107, 111, 112, 113, 117, 120, 130], [9, 29, 33, 44, 49, 53, 54, 58, 67, 68, 81, 96]]target = [[99, 107, 90, 119, 117, 113, 92, 120, 121, 123, 86, 94], [32, 87, 79, 70, 51, 71, 33, 67, 53, 82, 41, 56]]
+Epoch 0: |                                                                                                                                                     | 1219/? [1:17:26<00:00,  0.26it/s, v_num=5b6z]train step 1220; scene = [['0049c83bad21bbdf'], ['a73f4bf6ce8b2a00'], ['f7bef32e09ab061e'], ['175e1611df44964f']]; loss = 0.044071
+Epoch 0: |                                                                                                                                                     | 1220/? [1:17:30<00:00,  0.26it/s, v_num=5b6z]context = [[44, 47, 55, 57, 65, 80, 85, 87, 88, 94, 112, 116], [25, 37, 49, 59, 61, 70, 82, 84, 86, 94, 97, 100]]target = [[96, 81, 62, 98, 79, 76, 52, 95, 59, 65, 56, 72], [34, 26, 32, 50, 44, 97, 67, 53, 75, 66, 59, 40]]
+Epoch 0: |                                                                                                                                                     | 1229/? [1:18:04<00:00,  0.26it/s, v_num=5b6z]train step 1230; scene = [['342ff7b7111b53c1'], ['9ebe2434b1d68246'], ['b84cce034b55e1e4']]; loss = 0.048377
+Epoch 0: |                                                                                                                                                     | 1230/? [1:18:08<00:00,  0.26it/s, v_num=5b6z]context = [[5, 8, 11, 16, 19, 23, 25, 34, 35, 36, 43, 49, 51, 57, 60, 68, 77, 78, 86, 88, 97, 98, 101, 102]]target = [[84, 48, 64, 18, 41, 19, 60, 25, 52, 49, 51, 50, 75, 38, 42, 81, 92, 32, 69, 67, 26, 100, 65, 61]]
+Epoch 0: |                                                                                                                                                     | 1239/? [1:18:42<00:00,  0.26it/s, v_num=5b6z]train step 1240; scene = [['d7ac888a45c3c904'], ['60d988ddbc2d04f1'], ['617340307747e227'], ['969c91a2507e2d81'], ['af197296b340b564'], ['d488b5f08aadff0c']]; loss = 0.087887
+Epoch 0: |                                                                                                                                                     | 1240/? [1:18:46<00:00,  0.26it/s, v_num=5b6z]context = [[9, 16, 18, 25, 28, 30, 41, 43, 46, 51, 53, 55, 56, 57, 59, 68, 71, 72, 74, 77, 87, 95, 105, 106]]target = [[72, 53, 26, 17, 11, 65, 42, 97, 60, 75, 93, 10, 45, 15, 102, 38, 101, 54, 23, 98, 13, 68, 16, 46]]
+Epoch 0: |                                                                                                                                                     | 1249/? [1:19:20<00:00,  0.26it/s, v_num=5b6z]train step 1250; scene = [['d51d569c27d0b2b5'], ['e33b2a9076f25c4d'], ['2f330103819454c6'], ['944e92ff3fea78eb'], ['5a6387d05cd51e02'], ['2f269b68e14e256a'], ['55f1b82ae9c5571f'], ['1490c145692a1899'], ['b1b24b049d5a5da4'], ['834e851000651b8f'], ['4bc6b34a301aac73'], ['97caecafbfa7f1f6']]; loss = 0.112026
+Epoch 0: |                                                                                                                                                     | 1250/? [1:19:24<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1250; scene = ['70b0a33083333dc9'];
+target intrinsic:  tensor(0.8872, device='cuda:0') tensor(0.8874, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8856, device='cuda:0') tensor(0.8840, device='cuda:0')
+[2026-02-24 20:34:47,502][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1250/? [1:19:25<00:00,  0.26it/s, v_num=5b6z]context = [[0, 2, 3, 7, 9, 14, 16, 19, 27, 30, 32, 34, 39, 43, 56, 62, 65, 73, 75, 80, 82, 85, 91, 97]]target = [[7, 68, 16, 60, 94, 67, 34, 41, 53, 11, 1, 12, 30, 6, 61, 76, 22, 32, 25, 81, 88, 33, 9, 45]]
+Epoch 0: |                                                                                                                                                     | 1259/? [1:19:59<00:00,  0.26it/s, v_num=5b6z]train step 1260; scene = [['1a1f0a618bad5d30'], ['3046f14dad8f19f2'], ['8b4aecd7318912e3']]; loss = 0.047074
+Epoch 0: |                                                                                                                                                     | 1260/? [1:20:02<00:00,  0.26it/s, v_num=5b6z]context = [[44, 48, 50, 51, 55, 60, 62, 66, 72, 78, 79, 85, 91, 98, 99, 100, 102, 112, 117, 121, 123, 125, 129, 141]]target = [[73, 109, 111, 54, 63, 69, 65, 124, 82, 119, 125, 110, 117, 56, 123, 86, 99, 126, 134, 53, 84, 112, 74, 100]]
+Epoch 0: |                                                                                                                                                     | 1269/? [1:20:36<00:00,  0.26it/s, v_num=5b6z]train step 1270; scene = [['a0438ace9b7619cf'], ['336b3140d9d8bebd']]; loss = 0.070944
+Epoch 0: |                                                                                                                                                     | 1270/? [1:20:40<00:00,  0.26it/s, v_num=5b6z]context = [[20, 22, 25, 26, 40, 44, 45, 51, 55, 63, 67, 70, 72, 76, 80, 81, 82, 84, 88, 96, 104, 109, 110, 117]]target = [[84, 115, 116, 28, 110, 27, 111, 85, 100, 39, 61, 63, 68, 56, 101, 113, 108, 94, 73, 62, 67, 98, 47, 55]]
+Epoch 0: |                                                                                                                                                     | 1279/? [1:21:14<00:00,  0.26it/s, v_num=5b6z]train step 1280; scene = [['118f75d62c9a1f46'], ['3bcbf64a56113736'], ['6ca39802dcef328e'], ['0d754618c77e44f6']]; loss = 0.058882
+Epoch 0: |                                                                                                                                                     | 1280/? [1:21:18<00:00,  0.26it/s, v_num=5b6z]context = [[5, 19, 21, 22, 26, 37, 47, 50, 53, 75, 85, 87], [4, 19, 23, 26, 29, 36, 37, 63, 69, 72, 75, 80]]target = [[65, 33, 14, 32, 52, 25, 54, 34, 15, 6, 60, 30], [11, 5, 9, 35, 70, 15, 51, 38, 20, 42, 68, 22]]
+Epoch 0: |                                                                                                                                                     | 1289/? [1:21:51<00:00,  0.26it/s, v_num=5b6z]train step 1290; scene = [['4738c57794aa47f8']]; loss = 0.056338
+Epoch 0: |                                                                                                                                                     | 1290/? [1:21:55<00:00,  0.26it/s, v_num=5b6z]context = [[1, 4, 14, 17, 27, 30, 32, 37, 44, 52, 64, 80], [0, 8, 12, 14, 24, 28, 29, 37, 49, 54, 56, 67]]target = [[64, 28, 44, 35, 23, 68, 42, 61, 3, 77, 52, 14], [1, 11, 56, 59, 13, 63, 55, 61, 38, 9, 18, 37]]
+Epoch 0: |                                                                                                                                                     | 1299/? [1:22:29<00:00,  0.26it/s, v_num=5b6z]train step 1300; scene = [['f18e524ab2d288c3'], ['9dff9a317da1d49d'], ['11a87d6d490e024d']]; loss = 0.046989
+Epoch 0: |                                                                                                                                                     | 1300/? [1:22:33<00:00,  0.26it/s, v_num=5b6z]context = [[5, 6, 8, 12, 30, 45, 58, 64, 65, 75, 79, 85], [22, 26, 27, 28, 39, 45, 47, 55, 56, 57, 66, 72]]target = [[34, 57, 32, 7, 11, 58, 73, 68, 83, 14, 61, 55], [42, 34, 36, 47, 33, 67, 31, 30, 50, 27, 24, 58]]
+Epoch 0: |                                                                                                                                                     | 1309/? [1:23:07<00:00,  0.26it/s, v_num=5b6z]train step 1310; scene = [['589e362118b32d25'], ['d13522abd38eddfe'], ['f1657c6128d2b332'], ['1cb3ecb30e3e9d0e'], ['827c5f3b3886553d'], ['4d27fb96530fe02b']]; loss = 0.108010
+Epoch 0: |                                                                                                                                                     | 1310/? [1:23:11<00:00,  0.26it/s, v_num=5b6z]context = [[20, 21, 27, 31, 36, 40, 48, 51, 55, 62, 69, 72, 74, 75, 83, 89, 96, 97, 99, 104, 105, 106, 114, 117]]target = [[57, 113, 99, 80, 42, 30, 24, 74, 34, 52, 63, 59, 100, 101, 76, 36, 105, 66, 55, 27, 84, 22, 79, 86]]
+Epoch 0: |                                                                                                                                                     | 1319/? [1:23:45<00:00,  0.26it/s, v_num=5b6z]train step 1320; scene = [['c21766bd51fed5bc'], ['47a2a8b326b9f40e'], ['d0c0d78936b6e5ce'], ['e0ee3878561a5fed']]; loss = 0.065103
+Epoch 0: |                                                                                                                                                     | 1320/? [1:23:48<00:00,  0.26it/s, v_num=5b6z]context = [[194, 242], [34, 99], [51, 136], [156, 207], [36, 100], [0, 48], [10, 99], [22, 79], [18, 99], [2, 74], [39, 128], [201, 272]]target = [[225, 201], [63, 36], [58, 72], [193, 181], [39, 77], [13, 1], [29, 68], [36, 58], [45, 66], [58, 18], [100, 50], [234, 262]]
+Epoch 0: |                                                                                                                                                     | 1329/? [1:24:23<00:00,  0.26it/s, v_num=5b6z]train step 1330; scene = [['07fdb102ee3677f5']]; loss = 0.045392
+Epoch 0: |                                                                                                                                                     | 1330/? [1:24:27<00:00,  0.26it/s, v_num=5b6z]context = [[20, 23, 28, 29, 37, 43, 49, 53, 56, 57, 61, 65, 69, 70, 80, 92, 98, 102, 103, 104, 106, 108, 116, 117]]target = [[37, 57, 73, 79, 30, 102, 42, 115, 86, 33, 28, 25, 114, 53, 76, 107, 36, 106, 52, 49, 59, 66, 61, 101]]
+Epoch 0: |                                                                                                                                                     | 1339/? [1:25:01<00:00,  0.26it/s, v_num=5b6z]train step 1340; scene = [['2ff4f3b2475e0e8c'], ['db062134f9dec5f1']]; loss = 0.038457
+Epoch 0: |                                                                                                                                                     | 1340/? [1:25:05<00:00,  0.26it/s, v_num=5b6z]context = [[20, 22, 50, 84], [0, 36, 48, 51], [10, 38, 52, 66], [0, 24, 25, 53], [0, 4, 19, 45], [54, 57, 73, 112]]target = [[42, 28, 78, 23], [3, 12, 46, 44], [57, 60, 15, 41], [47, 3, 31, 2], [42, 13, 23, 28], [102, 69, 98, 70]]
+Epoch 0: |                                                                                                                                                     | 1349/? [1:25:39<00:00,  0.26it/s, v_num=5b6z]train step 1350; scene = [['87e1164e050e9686'], ['3803b2c8fd539a88']]; loss = 0.068446
+Epoch 0: |                                                                                                                                                     | 1350/? [1:25:42<00:00,  0.26it/s, v_num=5b6z]context = [[52, 53, 55, 63, 64, 66, 74, 75, 77, 79, 85, 89, 90, 99, 121, 122, 123, 126, 133, 134, 135, 136, 147, 149]]target = [[99, 80, 109, 136, 92, 134, 112, 76, 135, 68, 116, 138, 54, 145, 144, 143, 61, 72, 102, 77, 90, 85, 139, 58]]
+Epoch 0: |                                                                                                                                                     | 1359/? [1:26:17<00:00,  0.26it/s, v_num=5b6z]train step 1360; scene = [['7db4ed902d003a63']]; loss = 0.071257
+Epoch 0: |                                                                                                                                                     | 1360/? [1:26:21<00:00,  0.26it/s, v_num=5b6z]context = [[13, 26, 58, 83], [117, 127, 165, 171], [12, 45, 50, 66], [21, 23, 67, 101], [26, 34, 95, 97], [42, 73, 81, 98]]target = [[65, 36, 43, 27], [124, 136, 141, 132], [33, 37, 62, 29], [82, 90, 80, 97], [45, 44, 32, 86], [55, 82, 58, 89]]
+Epoch 0: |                                                                                                                                                     | 1369/? [1:26:55<00:00,  0.26it/s, v_num=5b6z]train step 1370; scene = [['36df585860d0ad88']]; loss = 0.034171
+Epoch 0: |                                                                                                                                                     | 1370/? [1:26:59<00:00,  0.26it/s, v_num=5b6z]context = [[41, 42, 45, 55, 59, 62, 67, 68, 69, 73, 79, 87, 90, 94, 104, 114, 117, 120, 121, 125, 128, 130, 135, 138]]target = [[79, 103, 135, 133, 104, 59, 92, 71, 62, 66, 61, 109, 68, 124, 74, 54, 44, 46, 57, 96, 108, 58, 85, 83]]
+Epoch 0: |                                                                                                                                                     | 1379/? [1:27:33<00:00,  0.26it/s, v_num=5b6z]train step 1380; scene = [['256ae648672281d1'], ['7043e8afce176c8c']]; loss = 0.071870
+Epoch 0: |                                                                                                                                                     | 1380/? [1:27:36<00:00,  0.26it/s, v_num=5b6z]context = [[31, 33, 48, 63, 106, 115], [8, 28, 29, 44, 67, 97], [65, 72, 100, 110, 115, 124], [8, 9, 63, 70, 75, 79]]target = [[57, 76, 83, 113, 107, 38], [51, 64, 72, 39, 18, 77], [89, 108, 95, 114, 84, 116], [20, 59, 14, 56, 15, 38]]
+Epoch 0: |                                                                                                                                                     | 1389/? [1:28:10<00:00,  0.26it/s, v_num=5b6z]train step 1390; scene = [['ca65a604c0f00319'], ['937ea87bb5a2047f'], ['a2b430d7bec915d8'], ['c8c116c28ca108b0']]; loss = 0.046987
+Epoch 0: |                                                                                                                                                     | 1390/? [1:28:14<00:00,  0.26it/s, v_num=5b6z]context = [[0, 3, 5, 26, 31, 34, 38, 40, 44, 53, 62, 70], [1, 4, 21, 29, 36, 37, 39, 44, 52, 54, 59, 64]]target = [[36, 5, 37, 41, 54, 19, 44, 21, 40, 34, 59, 32], [59, 51, 53, 46, 27, 9, 21, 52, 24, 54, 32, 48]]
+Epoch 0: |                                                                                                                                                     | 1399/? [1:28:47<00:00,  0.26it/s, v_num=5b6z]train step 1400; scene = [['73fef5139753e974'], ['4b3e117d4f50b167'], ['0e916e63743f841b']]; loss = 0.056986
+Epoch 0: |                                                                                                                                                     | 1400/? [1:28:51<00:00,  0.26it/s, v_num=5b6z]context = [[23, 26, 27, 29, 31, 32, 41, 47, 49, 50, 55, 56, 58, 71, 77, 78, 87, 91, 97, 100, 101, 102, 113, 120]]target = [[55, 52, 103, 28, 68, 66, 49, 93, 76, 109, 105, 112, 100, 101, 31, 70, 64, 116, 77, 56, 46, 48, 32, 85]]
+[2026-02-24 20:44:17,905][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1409/? [1:29:30<00:00,  0.26it/s, v_num=5b6z]train step 1410; scene = [['d9ce9620cf31e776'], ['36bf8b485c1284dd']]; loss = 0.057916
+Epoch 0: |                                                                                                                                                     | 1410/? [1:29:34<00:00,  0.26it/s, v_num=5b6z]context = [[44, 47, 51, 55, 57, 59, 61, 62, 68, 72, 78, 80, 82, 88, 91, 97, 109, 112, 116, 118, 135, 138, 140, 141]]target = [[54, 91, 112, 61, 115, 80, 93, 45, 105, 51, 131, 102, 55, 60, 57, 136, 129, 99, 125, 113, 84, 89, 58, 65]]
+Epoch 0: |                                                                                                                                                     | 1419/? [1:30:08<00:00,  0.26it/s, v_num=5b6z]train step 1420; scene = [['df4e1717577f82c1'], ['50b38521b86dd7b6'], ['e4fc6fa4bbd6efa1'], ['54d3668b6c7ed4b8']]; loss = 0.041016
+Epoch 0: |                                                                                                                                                     | 1420/? [1:30:11<00:00,  0.26it/s, v_num=5b6z]context = [[38, 43, 53, 54, 68, 87, 97, 101], [52, 53, 74, 84, 86, 96, 132, 133], [157, 173, 184, 190, 200, 203, 210, 211]]target = [[77, 71, 57, 46, 97, 67, 93, 96], [83, 127, 75, 122, 56, 78, 130, 107], [177, 183, 188, 174, 191, 203, 200, 170]]
+Epoch 0: |                                                                                                                                                     | 1429/? [1:30:46<00:00,  0.26it/s, v_num=5b6z]train step 1430; scene = [['e70c21625fceaebd'], ['4cfdc3ed984caecf'], ['dfc00f3016d34131'], ['f5338d00f9c021a3'], ['89fbc4149a5d7348'], ['d2d3474ebd2be7e8'], ['0b42b8cf14bd1e13'], ['358a2d1387de3df8'], ['abe463f94db2c398'], ['d35b7c2c52c45a54'], ['12f629409a280733'], ['6b09a022387bd762']]; loss = 0.104974
+Epoch 0: |                                                                                                                                                     | 1430/? [1:30:50<00:00,  0.26it/s, v_num=5b6z]context = [[146, 159, 179, 187, 191, 200], [3, 11, 37, 39, 40, 48], [3, 16, 25, 27, 47, 49], [31, 36, 39, 84, 87, 94]]target = [[154, 176, 156, 190, 172, 150], [13, 36, 28, 46, 33, 32], [33, 7, 23, 32, 20, 35], [53, 82, 44, 85, 38, 54]]
+Epoch 0: |                                                                                                                                                     | 1439/? [1:31:24<00:00,  0.26it/s, v_num=5b6z]train step 1440; scene = [['83f69d126eb1528d'], ['e92e4346a6224492']]; loss = 0.042914
+Epoch 0: |                                                                                                                                                     | 1440/? [1:31:28<00:00,  0.26it/s, v_num=5b6z]context = [[1, 19, 33, 39, 51, 53, 57, 61], [4, 18, 19, 43, 44, 47, 49, 75], [40, 51, 69, 84, 88, 93, 99, 104]]target = [[3, 30, 34, 24, 46, 8, 13, 4], [19, 24, 28, 56, 55, 10, 66, 21], [54, 88, 79, 100, 81, 84, 83, 72]]
+Epoch 0: |                                                                                                                                                     | 1449/? [1:32:02<00:00,  0.26it/s, v_num=5b6z]train step 1450; scene = [['c05fd148d2da6d26'], ['b49cc4ec7c6b0050'], ['0d158225b3c47682']]; loss = 0.071190
+Epoch 0: |                                                                                                                                                     | 1450/? [1:32:06<00:00,  0.26it/s, v_num=5b6z]context = [[2, 15, 27, 38, 41, 50, 51, 57, 58, 61, 62, 73], [16, 17, 19, 26, 41, 44, 80, 89, 90, 92, 100, 101]]target = [[37, 21, 47, 67, 70, 31, 10, 57, 34, 30, 50, 52], [84, 77, 99, 81, 19, 59, 24, 78, 64, 25, 30, 90]]
+Epoch 0: |                                                                                                                                                     | 1459/? [1:32:40<00:00,  0.26it/s, v_num=5b6z]train step 1460; scene = [['8d583bfb265295b9']]; loss = 0.049028
+Epoch 0: |                                                                                                                                                     | 1460/? [1:32:44<00:00,  0.26it/s, v_num=5b6z]context = [[143, 146, 157, 160, 179, 202], [1, 13, 19, 36, 45, 46], [175, 189, 227, 237, 252, 257], [22, 29, 47, 51, 54, 67]]target = [[156, 146, 191, 160, 196, 187], [28, 40, 8, 23, 33, 44], [234, 248, 199, 178, 220, 226], [38, 63, 50, 35, 66, 33]]
+Epoch 0: |                                                                                                                                                     | 1469/? [1:33:18<00:00,  0.26it/s, v_num=5b6z]train step 1470; scene = [['6ebab888069161eb']]; loss = 0.043518
+Epoch 0: |                                                                                                                                                     | 1470/? [1:33:22<00:00,  0.26it/s, v_num=5b6z]context = [[28, 29, 32, 37, 43, 46, 50, 65, 67, 71, 76, 77], [132, 133, 136, 137, 140, 141, 145, 147, 169, 171, 200, 204]]target = [[31, 53, 62, 64, 65, 67, 34, 72, 39, 58, 57, 68], [174, 172, 164, 145, 162, 199, 202, 142, 163, 159, 176, 200]]
+Epoch 0: |                                                                                                                                                     | 1479/? [1:33:57<00:00,  0.26it/s, v_num=5b6z]train step 1480; scene = [['964d888d8d08f2aa'], ['58901334e2d813d9'], ['ea4146e3386ff1ac']]; loss = 0.087446
+Epoch 0: |                                                                                                                                                     | 1480/? [1:34:00<00:00,  0.26it/s, v_num=5b6z]context = [[50, 57, 67, 77, 79, 101], [66, 67, 75, 83, 96, 117], [53, 65, 70, 81, 130, 138], [19, 45, 53, 57, 65, 66]]target = [[76, 85, 70, 53, 90, 65], [87, 116, 103, 85, 111, 97], [80, 116, 87, 117, 54, 112], [35, 47, 59, 51, 53, 28]]
+Epoch 0: |                                                                                                                                                     | 1489/? [1:34:34<00:00,  0.26it/s, v_num=5b6z]train step 1490; scene = [['f73db02fdfe72073'], ['a4eeae8de8e2e98e']]; loss = 0.039439
+Epoch 0: |                                                                                                                                                     | 1490/? [1:34:37<00:00,  0.26it/s, v_num=5b6z]context = [[10, 19, 23, 29, 35, 40, 41, 42, 43, 53, 54, 57, 63, 65, 66, 69, 72, 73, 77, 84, 91, 94, 95, 107]]target = [[97, 79, 11, 35, 65, 87, 80, 51, 32, 13, 83, 72, 85, 34, 105, 92, 59, 25, 81, 40, 36, 82, 88, 15]]
+Epoch 0: |                                                                                                                                                     | 1499/? [1:35:12<00:00,  0.26it/s, v_num=5b6z]train step 1500; scene = [['fa1ddac84aafd9b7']]; loss = 0.125263
+Epoch 0: |                                                                                                                                                     | 1500/? [1:35:16<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1500; scene = ['45592a7f307bccd0'];
+target intrinsic:  tensor(0.8508, device='cuda:0') tensor(0.8510, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8651, device='cuda:0') tensor(0.8656, device='cuda:0')
+[2026-02-24 20:50:50,712][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1500/? [1:35:28<00:00,  0.26it/s, v_num=5b6z]context = [[5, 10, 19, 23, 27, 33, 40, 57, 60, 63, 64, 68], [97, 102, 107, 111, 113, 120, 137, 148, 157, 163, 168, 169]]target = [[54, 31, 25, 44, 17, 36, 32, 65, 16, 11, 23, 38], [99, 138, 103, 159, 164, 104, 123, 135, 120, 124, 130, 148]]
+Epoch 0: |                                                                                                                                                     | 1509/? [1:36:01<00:00,  0.26it/s, v_num=5b6z]train step 1510; scene = [['289a0ef9678c7c70'], ['f4fd1fee2a8f69ee'], ['a88b44d03528e2f2']]; loss = 0.047381
+Epoch 0: |                                                                                                                                                     | 1510/? [1:36:05<00:00,  0.26it/s, v_num=5b6z]context = [[1, 2, 8, 18, 19, 27, 33, 35, 48, 49, 63, 77], [0, 9, 12, 15, 20, 22, 23, 30, 34, 58, 66, 68]]target = [[12, 51, 13, 69, 65, 38, 6, 30, 37, 18, 53, 2], [8, 14, 60, 49, 63, 46, 47, 3, 35, 13, 21, 44]]
+Epoch 0: |                                                                                                                                                     | 1519/? [1:36:38<00:00,  0.26it/s, v_num=5b6z]train step 1520; scene = [['85cff5568e6f52e7'], ['9de970f9a14770a9'], ['a08eb87db37b694a'], ['0a8fda80930b52ae']]; loss = 0.050095
+Epoch 0: |                                                                                                                                                     | 1520/? [1:36:42<00:00,  0.26it/s, v_num=5b6z]context = [[6, 32, 33, 39, 62, 69, 77, 96], [32, 33, 42, 61, 83, 93, 99, 107], [13, 27, 41, 70, 78, 84, 86, 93]]target = [[28, 30, 62, 7, 57, 86, 77, 36], [101, 43, 41, 68, 46, 36, 47, 105], [27, 16, 29, 71, 40, 45, 74, 65]]
+Epoch 0: |                                                                                                                                                     | 1529/? [1:37:15<00:00,  0.26it/s, v_num=5b6z]train step 1530; scene = [['43cc195276cf0c56']]; loss = 0.068442
+Epoch 0: |                                                                                                                                                     | 1530/? [1:37:19<00:00,  0.26it/s, v_num=5b6z]context = [[86, 89, 91, 102, 109, 136, 173, 175], [1, 12, 32, 33, 36, 37, 43, 47], [15, 24, 40, 41, 44, 70, 73, 102]]target = [[138, 158, 117, 118, 170, 120, 165, 133], [37, 44, 31, 32, 24, 9, 40, 38], [16, 27, 88, 46, 69, 52, 73, 44]]
+Epoch 0: |                                                                                                                                                     | 1539/? [1:37:53<00:00,  0.26it/s, v_num=5b6z]train step 1540; scene = [['4b7f3f58b0838d38']]; loss = 0.046770
+Epoch 0: |                                                                                                                                                     | 1540/? [1:37:57<00:00,  0.26it/s, v_num=5b6z]context = [[17, 21, 24, 40, 42, 44, 49, 50, 60, 62, 63, 64, 65, 72, 73, 76, 77, 85, 89, 93, 96, 104, 108, 114]]target = [[104, 84, 41, 96, 107, 103, 94, 44, 111, 97, 83, 54, 100, 57, 27, 38, 113, 29, 18, 99, 76, 101, 55, 36]]
+Epoch 0: |                                                                                                                                                     | 1549/? [1:38:32<00:00,  0.26it/s, v_num=5b6z]train step 1550; scene = [['d6cc1a3af543a7f8'], ['f3dd12e8d9dd4c20'], ['b74b90e3a87f285c'], ['c51f5219c3d09e33'], ['3f6666062c86b73a'], ['3b7bd7e723f069b2'], ['b3992ad0aff60272'], ['30d52bc66d89221d']]; loss = 0.072423
+Epoch 0: |                                                                                                                                                     | 1550/? [1:38:36<00:00,  0.26it/s, v_num=5b6z]context = [[9, 79], [29, 75], [29, 95], [124, 170], [42, 102], [92, 159], [94, 157], [6, 75], [188, 255], [5, 76], [133, 205], [2, 90]]target = [[55, 20], [47, 52], [85, 42], [136, 147], [97, 56], [158, 99], [145, 143], [17, 9], [199, 198], [34, 26], [184, 165], [39, 28]]
+Epoch 0: |                                                                                                                                                     | 1559/? [1:39:10<00:00,  0.26it/s, v_num=5b6z]train step 1560; scene = [['f6a87eade96cceb1']]; loss = 0.063170
+Epoch 0: |                                                                                                                                                     | 1560/? [1:39:14<00:00,  0.26it/s, v_num=5b6z]context = [[9, 17, 23, 26, 33, 40, 46, 62, 63, 68, 69, 71], [117, 126, 137, 140, 146, 163, 169, 173, 178, 186, 201, 205]]target = [[25, 36, 53, 31, 59, 20, 68, 24, 66, 13, 33, 67], [128, 179, 169, 126, 188, 187, 122, 127, 144, 165, 150, 119]]
+Epoch 0: |                                                                                                                                                     | 1569/? [1:39:47<00:00,  0.26it/s, v_num=5b6z]train step 1570; scene = [['ab1c5358d3bb05db']]; loss = 0.040961
+Epoch 0: |                                                                                                                                                     | 1570/? [1:39:51<00:00,  0.26it/s, v_num=5b6z]context = [[6, 8, 14, 38, 44, 69, 83, 89], [31, 47, 49, 50, 68, 77, 89, 91], [32, 46, 50, 61, 66, 84, 89, 100]]target = [[69, 79, 13, 31, 70, 24, 8, 27], [53, 42, 44, 68, 90, 65, 87, 49], [85, 91, 86, 84, 67, 83, 72, 92]]
+Epoch 0: |                                                                                                                                                     | 1579/? [1:40:25<00:00,  0.26it/s, v_num=5b6z]train step 1580; scene = [['572357b6b69cb9ec'], ['326dd7b41ce515ac']]; loss = 0.044385
+Epoch 0: |                                                                                                                                                     | 1580/? [1:40:28<00:00,  0.26it/s, v_num=5b6z]context = [[26, 32, 46, 57, 64, 105], [32, 38, 66, 84, 85, 95], [75, 87, 98, 106, 108, 130], [7, 9, 12, 20, 48, 60]]target = [[83, 29, 52, 46, 45, 39], [41, 73, 88, 76, 54, 82], [78, 99, 126, 113, 86, 105], [29, 45, 35, 34, 32, 16]]
+Epoch 0: |                                                                                                                                                     | 1589/? [1:41:03<00:00,  0.26it/s, v_num=5b6z]train step 1590; scene = [['4558408811e71246'], ['4a3ff4b7939ec268']]; loss = 0.065950
+Epoch 0: |                                                                                                                                                     | 1590/? [1:41:07<00:00,  0.26it/s, v_num=5b6z]context = [[187, 198, 222, 224, 227, 240], [196, 213, 228, 234, 257, 263], [15, 24, 48, 51, 62, 85], [64, 73, 82, 83, 109, 117]]target = [[231, 221, 188, 222, 220, 232], [257, 228, 261, 254, 244, 210], [45, 49, 23, 62, 65, 84], [74, 94, 93, 85, 66, 104]]
+Epoch 0: |                                                                                                                                                     | 1599/? [1:41:41<00:00,  0.26it/s, v_num=5b6z]train step 1600; scene = [['1849a41079039a3a']]; loss = 0.054742
+Epoch 0: |                                                                                                                                                     | 1600/? [1:41:45<00:00,  0.26it/s, v_num=5b6z]context = [[7, 10, 13, 17, 18, 30, 32, 34, 35, 38, 42, 49, 64, 66, 71, 74, 77, 84, 89, 91, 93, 95, 103, 104]]target = [[52, 78, 50, 26, 46, 43, 79, 25, 89, 70, 58, 18, 103, 98, 16, 74, 20, 13, 85, 96, 71, 76, 91, 34]]
+[2026-02-24 20:57:11,485][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1609/? [1:42:25<00:00,  0.26it/s, v_num=5b6z]train step 1610; scene = [['6405acc04d687493'], ['a802441c288b598f'], ['e2a52c2d94f21d4d']]; loss = 0.064816
+Epoch 0: |                                                                                                                                                     | 1610/? [1:42:28<00:00,  0.26it/s, v_num=5b6z]context = [[110, 111, 116, 122, 133, 134, 135, 143, 145, 148, 155, 167, 173, 177, 178, 180, 181, 183, 192, 193, 197, 199, 203, 207]]target = [[137, 115, 169, 197, 176, 116, 146, 159, 144, 119, 129, 161, 181, 160, 151, 196, 131, 204, 201, 157, 138, 123, 180, 117]]
+Epoch 0: |                                                                                                                                                     | 1619/? [1:43:03<00:00,  0.26it/s, v_num=5b6z]train step 1620; scene = [['a270004198eddd9b'], ['cc306ca65c5dbc24'], ['1a5022549d2718ae'], ['cdde30f3352829cb']]; loss = 0.060088
+Epoch 0: |                                                                                                                                                     | 1620/? [1:43:07<00:00,  0.26it/s, v_num=5b6z]context = [[72, 79, 103, 136], [6, 15, 31, 60], [1, 48, 52, 80], [95, 135, 145, 162], [136, 166, 176, 206], [0, 7, 45, 61]]target = [[90, 120, 96, 123], [15, 25, 11, 32], [5, 49, 65, 22], [116, 131, 128, 139], [144, 141, 183, 190], [34, 48, 35, 12]]
+Epoch 0: |                                                                                                                                                     | 1629/? [1:43:41<00:00,  0.26it/s, v_num=5b6z]train step 1630; scene = [['22c6139e04c88c0c']]; loss = 0.043561
+Epoch 0: |                                                                                                                                                     | 1630/? [1:43:45<00:00,  0.26it/s, v_num=5b6z]context = [[5, 7, 10, 22, 36, 52], [7, 11, 21, 27, 46, 72], [0, 50, 64, 79, 88, 90], [4, 24, 28, 39, 44, 49]]target = [[41, 40, 13, 50, 28, 25], [39, 20, 47, 50, 25, 55], [75, 69, 56, 70, 5, 13], [27, 35, 24, 13, 14, 40]]
+Epoch 0: |                                                                                                                                                     | 1639/? [1:44:19<00:00,  0.26it/s, v_num=5b6z]train step 1640; scene = [['62ce0e4e9490317d']]; loss = 0.034076
+Epoch 0: |                                                                                                                                                     | 1640/? [1:44:23<00:00,  0.26it/s, v_num=5b6z]context = [[4, 6, 8, 15, 23, 24, 26, 32, 35, 39, 50, 55], [0, 10, 22, 26, 27, 29, 35, 42, 60, 66, 73, 77]]target = [[41, 43, 30, 44, 33, 16, 39, 51, 12, 54, 53, 29], [4, 16, 21, 43, 47, 46, 44, 53, 45, 24, 51, 50]]
+Epoch 0: |                                                                                                                                                     | 1649/? [1:44:58<00:00,  0.26it/s, v_num=5b6z]train step 1650; scene = [['bad5f833677cd32f'], ['2cf78f93fd569aa1']]; loss = 0.069634
+Epoch 0: |                                                                                                                                                     | 1650/? [1:45:02<00:00,  0.26it/s, v_num=5b6z]context = [[9, 71], [34, 102], [46, 110], [10, 61], [137, 186], [1, 52], [128, 195], [10, 100], [198, 258], [16, 67], [12, 66], [5, 66]]target = [[26, 42], [83, 46], [67, 48], [14, 57], [154, 155], [49, 29], [155, 140], [41, 56], [236, 235], [32, 61], [50, 32], [35, 61]]
+Epoch 0: |                                                                                                                                                     | 1659/? [1:45:36<00:00,  0.26it/s, v_num=5b6z]train step 1660; scene = [['1f3f484a027f93d9'], ['12e85fb6e140ee85'], ['197edde42c1eaac3'], ['59058fb21817aa6b']]; loss = 0.055708
+Epoch 0: |                                                                                                                                                     | 1660/? [1:45:40<00:00,  0.26it/s, v_num=5b6z]context = [[24, 30, 39, 55, 81, 105], [182, 183, 195, 207, 215, 233], [25, 65, 77, 78, 93, 105], [20, 40, 46, 77, 82, 103]]target = [[48, 70, 100, 51, 59, 97], [197, 190, 194, 208, 198, 218], [44, 91, 86, 70, 101, 30], [65, 21, 102, 25, 30, 54]]
+Epoch 0: |                                                                                                                                                     | 1669/? [1:46:14<00:00,  0.26it/s, v_num=5b6z]train step 1670; scene = [['25aff15f6c54558b']]; loss = 0.039602
+Epoch 0: |                                                                                                                                                     | 1670/? [1:46:18<00:00,  0.26it/s, v_num=5b6z]context = [[154, 158, 162, 163, 165, 191, 194, 196, 201, 203, 207, 211], [137, 138, 151, 153, 161, 165, 166, 167, 185, 186, 197, 203]]target = [[203, 173, 193, 192, 207, 165, 187, 157, 174, 190, 182, 176], [165, 166, 189, 176, 201, 140, 139, 163, 184, 169, 186, 157]]
+Epoch 0: |                                                                                                                                                     | 1679/? [1:46:53<00:00,  0.26it/s, v_num=5b6z]train step 1680; scene = [['c513a3c2f59aa548'], ['493538f3442cb9fd']]; loss = 0.044710
+Epoch 0: |                                                                                                                                                     | 1680/? [1:46:56<00:00,  0.26it/s, v_num=5b6z]context = [[8, 9, 12, 14, 16, 17, 18, 20, 23, 25, 29, 36, 38, 46, 52, 63, 64, 78, 82, 90, 95, 97, 102, 105]]target = [[19, 24, 49, 13, 77, 40, 51, 54, 85, 10, 74, 88, 50, 94, 78, 71, 68, 87, 44, 95, 63, 81, 82, 96]]
+Epoch 0: |                                                                                                                                                     | 1689/? [1:47:31<00:00,  0.26it/s, v_num=5b6z]train step 1690; scene = [['da5de01cf7e41541']]; loss = 0.059239
+Epoch 0: |                                                                                                                                                     | 1690/? [1:47:34<00:00,  0.26it/s, v_num=5b6z]context = [[166, 168, 170, 180, 181, 183, 184, 191, 219, 234, 239, 240], [52, 57, 62, 79, 87, 99, 104, 107, 111, 115, 118, 124]]target = [[226, 177, 191, 171, 214, 212, 213, 234, 229, 186, 238, 219], [60, 111, 110, 109, 72, 56, 96, 101, 61, 66, 118, 69]]
+Epoch 0: |                                                                                                                                                     | 1699/? [1:48:08<00:00,  0.26it/s, v_num=5b6z]train step 1700; scene = [['3e763e3c28e87eed']]; loss = 0.025197
+Epoch 0: |                                                                                                                                                     | 1700/? [1:48:12<00:00,  0.26it/s, v_num=5b6z]context = [[45, 50, 54, 60, 92, 105, 108, 110], [0, 8, 30, 31, 41, 42, 43, 45], [22, 28, 31, 33, 41, 59, 72, 99]]target = [[71, 52, 86, 64, 72, 69, 73, 59], [14, 36, 2, 1, 35, 4, 16, 20], [74, 61, 93, 94, 31, 42, 26, 28]]
+Epoch 0: |                                                                                                                                                     | 1709/? [1:48:45<00:00,  0.26it/s, v_num=5b6z]train step 1710; scene = [['3727bb4b44708f89']]; loss = 0.090334
+Epoch 0: |                                                                                                                                                     | 1710/? [1:48:48<00:00,  0.26it/s, v_num=5b6z]context = [[59, 62, 71, 76, 105, 110, 112, 114], [1, 2, 12, 23, 30, 34, 44, 48], [75, 78, 86, 87, 90, 114, 117, 122]]target = [[104, 100, 89, 101, 94, 102, 96, 84], [4, 24, 32, 21, 39, 26, 25, 20], [89, 94, 76, 90, 85, 96, 83, 77]]
+Epoch 0: |                                                                                                                                                     | 1719/? [1:49:23<00:00,  0.26it/s, v_num=5b6z]train step 1720; scene = [['d938c73738634cbf'], ['d7f3b4e12f3f3af6'], ['1a17332c1d690519']]; loss = 0.039246
+Epoch 0: |                                                                                                                                                     | 1720/? [1:49:27<00:00,  0.26it/s, v_num=5b6z]context = [[26, 28, 33, 44, 45, 49, 55, 57, 58, 59, 67, 72, 75, 79, 82, 94, 98, 100, 113, 115, 117, 118, 122, 123]]target = [[107, 48, 44, 86, 57, 93, 73, 51, 58, 88, 78, 59, 33, 99, 100, 87, 62, 38, 36, 84, 119, 122, 91, 74]]
+Epoch 0: |                                                                                                                                                     | 1729/? [1:50:01<00:00,  0.26it/s, v_num=5b6z]train step 1730; scene = [['2d9ea631e6423141'], ['b0c6597c77c51a8c']]; loss = 0.058166
+Epoch 0: |                                                                                                                                                     | 1730/? [1:50:05<00:00,  0.26it/s, v_num=5b6z]context = [[87, 105, 125, 127, 131, 138, 143, 150], [1, 14, 30, 33, 45, 48, 57, 66], [8, 18, 21, 26, 59, 92, 93, 94]]target = [[89, 146, 124, 104, 103, 96, 110, 143], [29, 33, 7, 12, 20, 27, 53, 51], [57, 81, 93, 74, 29, 30, 33, 60]]
+Epoch 0: |                                                                                                                                                     | 1739/? [1:50:39<00:00,  0.26it/s, v_num=5b6z]train step 1740; scene = [['56863fb499e2ff9a'], ['f5a3eca31fcdabf8'], ['946597adeb926d39'], ['23eecb3bfb301179'], ['4d999bce04f7516a'], ['b1f32a7b0a8e25ce'], ['72ac5af57c88795c'], ['cc45417d1c5dab8d']]; loss = 0.082100
+Epoch 0: |                                                                                                                                                     | 1740/? [1:50:43<00:00,  0.26it/s, v_num=5b6z]context = [[1, 4, 8, 12, 19, 26, 28, 29, 39, 42, 47, 50, 55, 66, 67, 68, 70, 76, 83, 85, 91, 92, 97, 98]]target = [[72, 40, 83, 3, 21, 13, 14, 61, 46, 53, 36, 57, 6, 10, 12, 37, 7, 22, 19, 78, 56, 25, 54, 80]]
+Epoch 0: |                                                                                                                                                     | 1749/? [1:51:17<00:00,  0.26it/s, v_num=5b6z]train step 1750; scene = [['7351b1a8a7405871'], ['edacb8db81943446']]; loss = 0.055478
+Epoch 0: |                                                                                                                                                     | 1750/? [1:51:21<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1750; scene = ['3b273cb40c55db95'];
+target intrinsic:  tensor(1.0504, device='cuda:0') tensor(1.0506, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9450, device='cuda:0') tensor(0.9430, device='cuda:0')
+[2026-02-24 21:06:44,763][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1750/? [1:51:22<00:00,  0.26it/s, v_num=5b6z]context = [[6, 12, 18, 21, 26, 27, 30, 33, 34, 35, 50, 55], [2, 5, 11, 19, 32, 36, 39, 44, 45, 46, 47, 56]]target = [[43, 26, 45, 30, 52, 34, 39, 12, 21, 11, 37, 19], [8, 19, 54, 25, 33, 4, 10, 55, 46, 34, 52, 3]]
+Epoch 0: |                                                                                                                                                     | 1759/? [1:51:55<00:00,  0.26it/s, v_num=5b6z]train step 1760; scene = [['616d045c5963ecb2'], ['4bbb9fadc993969a'], ['c8b833961ffd8ac5'], ['321addf5693a66aa']]; loss = 0.051873
+Epoch 0: |                                                                                                                                                     | 1760/? [1:51:59<00:00,  0.26it/s, v_num=5b6z]context = [[111, 116, 130, 132, 142, 146, 147, 158, 160, 164, 176, 181], [19, 20, 21, 24, 28, 31, 47, 58, 61, 70, 71, 77]]target = [[125, 122, 167, 145, 144, 152, 130, 121, 153, 165, 113, 139], [70, 73, 42, 59, 66, 23, 39, 72, 22, 60, 26, 25]]
+Epoch 0: |                                                                                                                                                     | 1769/? [1:52:33<00:00,  0.26it/s, v_num=5b6z]train step 1770; scene = [['8eb7ab71f603c9d4'], ['cc966a9b2af2232f']]; loss = 0.065768
+Epoch 0: |                                                                                                                                                     | 1770/? [1:52:36<00:00,  0.26it/s, v_num=5b6z]context = [[102, 104, 107, 108, 112, 122, 123, 126, 129, 133, 136, 142, 146, 147, 150, 161, 162, 163, 164, 184, 189, 191, 193, 199]]target = [[120, 136, 118, 149, 133, 122, 193, 137, 187, 174, 159, 156, 194, 170, 121, 189, 103, 178, 197, 196, 129, 195, 115, 181]]
+Epoch 0: |                                                                                                                                                     | 1779/? [1:53:11<00:00,  0.26it/s, v_num=5b6z]train step 1780; scene = [['bb28a1c09df8a484']]; loss = 0.042836
+Epoch 0: |                                                                                                                                                     | 1780/? [1:53:15<00:00,  0.26it/s, v_num=5b6z]context = [[11, 70], [25, 75], [78, 152], [51, 119], [24, 81], [136, 193], [20, 96], [30, 109], [7, 87], [58, 129], [66, 145], [27, 74]]target = [[17, 37], [64, 73], [135, 104], [69, 118], [42, 58], [145, 168], [30, 83], [99, 43], [42, 16], [61, 111], [134, 98], [55, 71]]
+Epoch 0: |                                                                                                                                                     | 1789/? [1:53:48<00:00,  0.26it/s, v_num=5b6z]train step 1790; scene = [['48c2b63b452555af'], ['1aa5fdc9ff855ee2']]; loss = 0.065579
+Epoch 0: |                                                                                                                                                     | 1790/? [1:53:52<00:00,  0.26it/s, v_num=5b6z]context = [[32, 36, 37, 42, 43, 52, 63, 67, 68, 73, 74, 90], [27, 40, 53, 55, 57, 63, 68, 79, 80, 81, 100, 102]]target = [[70, 67, 89, 83, 52, 35, 74, 73, 84, 44, 54, 33], [44, 60, 41, 98, 100, 51, 46, 99, 47, 92, 37, 31]]
+Epoch 0: |                                                                                                                                                     | 1799/? [1:54:27<00:00,  0.26it/s, v_num=5b6z]train step 1800; scene = [['b41ece74dd5aa87d'], ['e620345f5468d2e3']]; loss = 0.028817
+Epoch 0: |                                                                                                                                                     | 1800/? [1:54:31<00:00,  0.26it/s, v_num=5b6z]context = [[2, 47], [107, 163], [22, 107], [67, 146], [82, 130], [87, 151], [59, 133], [191, 242], [121, 203], [113, 167], [106, 159], [27, 94]]target = [[37, 42], [161, 109], [92, 85], [142, 139], [116, 104], [106, 134], [116, 87], [201, 206], [142, 129], [159, 155], [116, 134], [57, 36]]
+[2026-02-24 21:09:57,624][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1809/? [1:55:04<00:00,  0.26it/s, v_num=5b6z]train step 1810; scene = [['9e1b58ec292d6690'], ['63f6e847b2e2a2e7'], ['aa570297256d99b6'], ['02195bc2cdded89e'], ['8839e65c6f6bb394'], ['c6bd8d3b3367444f']]; loss = 0.071091
+Epoch 0: |                                                                                                                                                     | 1810/? [1:55:07<00:00,  0.26it/s, v_num=5b6z]context = [[3, 14, 18, 21, 23, 31, 32, 34, 36, 63, 70, 78], [24, 32, 33, 47, 62, 69, 70, 73, 80, 85, 86, 88]]target = [[5, 7, 41, 28, 48, 67, 12, 68, 47, 26, 10, 31], [25, 68, 40, 79, 82, 58, 26, 28, 36, 48, 66, 29]]
+Epoch 0: |                                                                                                                                                     | 1819/? [1:55:42<00:00,  0.26it/s, v_num=5b6z]train step 1820; scene = [['8ad3918394be8d98']]; loss = 0.061743
+Epoch 0: |                                                                                                                                                     | 1820/? [1:55:46<00:00,  0.26it/s, v_num=5b6z]context = [[0, 4, 7, 9, 13, 15, 17, 35, 37, 61, 69, 72], [5, 23, 24, 25, 26, 31, 50, 56, 60, 68, 74, 76]]target = [[45, 49, 51, 46, 6, 14, 52, 22, 65, 69, 10, 37], [9, 37, 72, 6, 28, 14, 64, 47, 15, 16, 55, 33]]
+Epoch 0: |                                                                                                                                                     | 1829/? [1:56:21<00:00,  0.26it/s, v_num=5b6z]train step 1830; scene = [['af4565fb713ed79f']]; loss = 0.031953
+Epoch 0: |                                                                                                                                                     | 1830/? [1:56:24<00:00,  0.26it/s, v_num=5b6z]context = [[133, 138, 168, 169, 171, 180, 190, 192, 196, 203, 204, 213], [81, 82, 84, 87, 92, 93, 95, 97, 105, 110, 118, 131]]target = [[147, 196, 136, 153, 201, 185, 205, 161, 169, 149, 210, 189], [114, 103, 99, 119, 107, 110, 88, 124, 86, 122, 126, 120]]
+Epoch 0: |                                                                                                                                                     | 1839/? [1:56:59<00:00,  0.26it/s, v_num=5b6z]train step 1840; scene = [['c690f7b93b2b75c5'], ['f0dbc908f65693ae'], ['ee81b7d704303a2c'], ['a2a3a75cafb630f5']]; loss = 0.071337
+Epoch 0: |                                                                                                                                                     | 1840/? [1:57:03<00:00,  0.26it/s, v_num=5b6z]context = [[95, 106, 107, 115, 116, 118, 120, 126, 128, 130, 133, 139, 148, 149, 150, 153, 159, 163, 173, 178, 182, 183, 184, 192]]target = [[167, 153, 132, 96, 110, 126, 111, 129, 162, 144, 185, 176, 106, 97, 156, 108, 177, 172, 148, 157, 133, 137, 149, 141]]
+Epoch 0: |                                                                                                                                                     | 1849/? [1:57:37<00:00,  0.26it/s, v_num=5b6z]train step 1850; scene = [['a1fa4bf737381a53'], ['2d7243929156069e'], ['201287bd653fa906'], ['2d7243929156069e'], ['d25d6a9b1356b134'], ['4a0f95a3db913b56']]; loss = 0.071632
+Epoch 0: |                                                                                                                                                     | 1850/? [1:57:40<00:00,  0.26it/s, v_num=5b6z]context = [[30, 32, 34, 52, 72, 85, 90, 108], [9, 25, 30, 33, 41, 46, 48, 84], [51, 72, 76, 82, 83, 99, 109, 115]]target = [[46, 41, 107, 87, 40, 64, 57, 53], [35, 56, 83, 38, 62, 21, 53, 37], [109, 111, 114, 53, 86, 67, 91, 104]]
+Epoch 0: |                                                                                                                                                     | 1859/? [1:58:13<00:00,  0.26it/s, v_num=5b6z]train step 1860; scene = [['082779aebfbf4a46']]; loss = 0.033247
+Epoch 0: |                                                                                                                                                     | 1860/? [1:58:17<00:00,  0.26it/s, v_num=5b6z]context = [[3, 17, 39, 49], [112, 121, 139, 157], [128, 136, 147, 178], [2, 10, 33, 87], [25, 50, 81, 98], [28, 48, 69, 113]]target = [[8, 40, 43, 26], [113, 124, 154, 133], [156, 164, 137, 168], [62, 24, 86, 11], [68, 70, 78, 73], [73, 100, 109, 29]]
+Epoch 0: |                                                                                                                                                     | 1869/? [1:58:50<00:00,  0.26it/s, v_num=5b6z]train step 1870; scene = [['01c143c81a4d2145'], ['9b0b82db99ff360a'], ['80f5da17b4119bbf'], ['8c5dce5d79b3d2aa'], ['d3a0a89d951a6101'], ['da1f9f2859b59142'], ['20b791440b7ec0b5'], ['00703cbf7531ef11'], ['4aef5d4b9287e08a'], ['27ed8e7077af6540'], ['04e7f97215df7078'], ['08735c801ab7efb5']]; loss = 0.095811
+Epoch 0: |                                                                                                                                                     | 1870/? [1:58:54<00:00,  0.26it/s, v_num=5b6z]context = [[41, 44, 47, 48, 50, 51, 56, 67, 72, 81, 82, 83, 84, 85, 89, 90, 97, 105, 107, 110, 115, 125, 135, 138]]target = [[79, 108, 90, 99, 85, 65, 59, 69, 73, 126, 131, 100, 114, 52, 87, 47, 123, 54, 43, 137, 94, 88, 70, 75]]
+Epoch 0: |                                                                                                                                                     | 1879/? [1:59:29<00:00,  0.26it/s, v_num=5b6z]train step 1880; scene = [['42b2dabdc5ea4d93']]; loss = 0.033271
+Epoch 0: |                                                                                                                                                     | 1880/? [1:59:33<00:00,  0.26it/s, v_num=5b6z]context = [[17, 29, 31, 34, 36, 39, 48, 59, 64, 66, 67, 79], [42, 47, 51, 56, 57, 61, 64, 70, 74, 83, 92, 93]]target = [[53, 20, 25, 74, 39, 30, 21, 59, 50, 66, 40, 44], [43, 84, 45, 72, 82, 85, 56, 73, 92, 76, 75, 46]]
+Epoch 0: |                                                                                                                                                     | 1889/? [2:00:07<00:00,  0.26it/s, v_num=5b6z]train step 1890; scene = [['ed4855892fd5fa4a'], ['9683be82b6c1e851']]; loss = 0.061055
+Epoch 0: |                                                                                                                                                     | 1890/? [2:00:11<00:00,  0.26it/s, v_num=5b6z]context = [[22, 37, 39, 49, 50, 52, 54, 56, 59, 61, 72, 74, 84, 88, 93, 94, 97, 102, 104, 105, 108, 109, 112, 119]]target = [[59, 53, 102, 34, 27, 44, 100, 45, 32, 65, 36, 42, 87, 82, 85, 26, 80, 60, 90, 113, 86, 78, 40, 68]]
+Epoch 0: |                                                                                                                                                     | 1899/? [2:00:46<00:00,  0.26it/s, v_num=5b6z]train step 1900; scene = [['1844e80dafb62927'], ['b43f32da9a7800ee'], ['11094b4b71a9ab00'], ['0f5daf32c2e8fefd'], ['f8e7bd9e403fc04a'], ['b11afc47f2feb21c'], ['5994966dd0897ead'], ['6d2aae2a7dd35e14'], ['e545aecc18cfa501'], ['82d896f5142ee6dd'], ['140b10a4f6bb5aa5'], ['2ec8edfc07c8841f']]; loss = 0.090232
+Epoch 0: |                                                                                                                                                     | 1900/? [2:00:50<00:00,  0.26it/s, v_num=5b6z]context = [[1, 20, 26, 32, 34, 53], [83, 87, 95, 106, 125, 139], [0, 13, 29, 39, 46, 77], [90, 122, 126, 130, 159, 167]]target = [[36, 12, 39, 17, 4, 7], [92, 86, 107, 112, 135, 137], [32, 69, 36, 28, 50, 38], [132, 161, 105, 165, 142, 102]]
+Epoch 0: |                                                                                                                                                     | 1909/? [2:01:25<00:00,  0.26it/s, v_num=5b6z]train step 1910; scene = [['50f7971dda42084f'], ['159a472b24f1c395'], ['04a191933c5b05ec']]; loss = 0.075593
+Epoch 0: |                                                                                                                                                     | 1910/? [2:01:29<00:00,  0.26it/s, v_num=5b6z]context = [[74, 78, 86, 94, 96, 101, 104, 107, 108, 109, 115, 117, 119, 131, 133, 135, 137, 144, 151, 152, 155, 158, 161, 171]]target = [[108, 142, 163, 121, 87, 111, 149, 96, 120, 145, 97, 122, 147, 101, 123, 95, 94, 132, 86, 89, 99, 140, 104, 136]]
+Epoch 0: |                                                                                                                                                     | 1919/? [2:02:02<00:00,  0.26it/s, v_num=5b6z]train step 1920; scene = [['bf4bbf858718317d'], ['39ed8d2efe760b94'], ['d9644f4985e51a1f']]; loss = 0.047545
+Epoch 0: |                                                                                                                                                     | 1920/? [2:02:06<00:00,  0.26it/s, v_num=5b6z]context = [[9, 79], [15, 63], [28, 73], [101, 167], [5, 70], [53, 102], [2, 75], [86, 135], [1, 58], [5, 84], [1, 76], [135, 208]]target = [[43, 20], [62, 45], [70, 64], [114, 144], [60, 52], [93, 101], [13, 68], [120, 89], [30, 23], [43, 16], [57, 14], [194, 205]]
+Epoch 0: |                                                                                                                                                     | 1929/? [2:02:39<00:00,  0.26it/s, v_num=5b6z]train step 1930; scene = [['6efa62598fececd0'], ['b25a0f4ffca51d79'], ['90dbfac63e6b89be'], ['08a366317a388734']]; loss = 0.076830
+Epoch 0: |                                                                                                                                                     | 1930/? [2:02:42<00:00,  0.26it/s, v_num=5b6z]context = [[3, 19, 23, 28, 32, 40, 48, 50, 53, 60, 63, 64, 66, 70, 74, 78, 80, 82, 87, 93, 96, 97, 98, 100]]target = [[63, 24, 82, 55, 68, 93, 95, 7, 53, 65, 76, 35, 83, 11, 22, 34, 52, 16, 61, 92, 26, 47, 32, 46]]
+Epoch 0: |                                                                                                                                                     | 1939/? [2:03:17<00:00,  0.26it/s, v_num=5b6z]train step 1940; scene = [['1236365ec263ad76']]; loss = 0.041582
+Epoch 0: |                                                                                                                                                     | 1940/? [2:03:21<00:00,  0.26it/s, v_num=5b6z]context = [[10, 11, 16, 23, 26, 33, 34, 43, 44, 48, 54, 58, 70, 73, 75, 78, 82, 86, 89, 91, 102, 105, 106, 107]]target = [[88, 68, 24, 39, 22, 100, 19, 44, 57, 94, 41, 37, 53, 69, 96, 63, 81, 104, 50, 64, 101, 15, 40, 26]]
+Epoch 0: |                                                                                                                                                     | 1949/? [2:03:55<00:00,  0.26it/s, v_num=5b6z]train step 1950; scene = [['2bc8b64aafc5870c'], ['1202c32d91ad3ee3'], ['300571576edc008c']]; loss = 0.104781
+Epoch 0: |                                                                                                                                                     | 1950/? [2:03:59<00:00,  0.26it/s, v_num=5b6z]context = [[11, 14, 25, 42, 47, 49, 58, 66, 68, 69, 79, 81], [39, 55, 56, 60, 63, 70, 74, 86, 93, 96, 99, 104]]target = [[25, 55, 33, 13, 26, 62, 22, 52, 32, 67, 30, 40], [95, 42, 43, 41, 59, 96, 102, 88, 48, 66, 100, 40]]
+Epoch 0: |                                                                                                                                                     | 1959/? [2:04:33<00:00,  0.26it/s, v_num=5b6z]train step 1960; scene = [['3ded3e3c1fe76ee3']]; loss = 0.042926
+Epoch 0: |                                                                                                                                                     | 1960/? [2:04:37<00:00,  0.26it/s, v_num=5b6z]context = [[109, 112, 117, 120, 130, 137, 142, 144, 145, 150, 151, 152, 155, 159, 160, 162, 166, 171, 175, 179, 194, 195, 196, 206]]target = [[168, 171, 125, 169, 131, 150, 188, 139, 145, 152, 126, 184, 205, 142, 119, 196, 148, 155, 185, 197, 154, 203, 143, 147]]
+Epoch 0: |                                                                                                                                                     | 1969/? [2:05:12<00:00,  0.26it/s, v_num=5b6z]train step 1970; scene = [['3b21f48e23e4917f']]; loss = 0.040997
+Epoch 0: |                                                                                                                                                     | 1970/? [2:05:16<00:00,  0.26it/s, v_num=5b6z]context = [[13, 15, 19, 21, 30, 33, 39, 49, 54, 57, 72, 85], [18, 26, 35, 36, 39, 46, 50, 60, 70, 71, 74, 80]]target = [[39, 30, 38, 35, 26, 78, 75, 56, 18, 44, 77, 71], [22, 24, 70, 65, 49, 19, 62, 59, 71, 73, 51, 39]]
+Epoch 0: |                                                                                                                                                     | 1979/? [2:05:50<00:00,  0.26it/s, v_num=5b6z]train step 1980; scene = [['723db63d24c84d1d']]; loss = 0.041961
+Epoch 0: |                                                                                                                                                     | 1980/? [2:05:54<00:00,  0.26it/s, v_num=5b6z]context = [[114, 115, 120, 122, 123, 125, 126, 140, 141, 145, 147, 161, 170, 179, 180, 181, 193, 195, 196, 197, 199, 203, 207, 211]]target = [[201, 148, 171, 190, 135, 143, 189, 141, 205, 151, 187, 206, 183, 144, 200, 147, 181, 184, 120, 207, 152, 128, 170, 204]]
+Epoch 0: |                                                                                                                                                     | 1989/? [2:06:29<00:00,  0.26it/s, v_num=5b6z]train step 1990; scene = [['12eb36bba5c89eeb'], ['40c5311e3b3accef'], ['06d2876e8c40a3b6'], ['2da17464ef895b63'], ['21951a6ae1c4b225'], ['ab78b3eb64029b73'], ['f6ef16edbf87f358'], ['66e4f3268dafe823']]; loss = 0.085061
+Epoch 0: |                                                                                                                                                     | 1990/? [2:06:31<00:00,  0.26it/s, v_num=5b6z]context = [[6, 9, 11, 12, 24, 34, 39, 44, 45, 53, 55, 59, 61, 66, 70, 71, 76, 78, 79, 82, 85, 87, 92, 103]]target = [[32, 63, 22, 65, 39, 36, 57, 91, 64, 13, 66, 60, 14, 71, 26, 41, 23, 101, 29, 93, 92, 90, 98, 73]]
+Epoch 0: |                                                                                                                                                     | 1999/? [2:07:06<00:00,  0.26it/s, v_num=5b6z]train step 2000; scene = [['1ac1373478877088']]; loss = 0.053479
+Epoch 0: |                                                                                                                                                     | 2000/? [2:07:10<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 2000; scene = ['be75142d4652fe3e'];
+target intrinsic:  tensor(0.9402, device='cuda:0') tensor(0.9404, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9333, device='cuda:0') tensor(0.9351, device='cuda:0')
+[2026-02-24 21:22:33,649][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2000/? [2:07:11<00:00,  0.26it/s, v_num=5b6z]context = [[16, 17, 18, 27, 35, 40, 41, 47, 48, 51, 52, 53, 54, 56, 79, 80, 87, 89, 92, 96, 100, 101, 106, 113]]target = [[102, 53, 81, 57, 21, 101, 52, 64, 43, 58, 59, 94, 46, 108, 32, 79, 36, 104, 26, 106, 31, 29, 60, 41]]
+[2026-02-24 21:22:37,986][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2009/? [2:07:49<00:00,  0.26it/s, v_num=5b6z]train step 2010; scene = [['e2f2a27bfce53270']]; loss = 0.040696
+Epoch 0: |                                                                                                                                                     | 2010/? [2:07:53<00:00,  0.26it/s, v_num=5b6z]context = [[32, 34, 44, 61, 73, 79], [66, 91, 112, 119, 124, 128], [52, 79, 88, 91, 97, 100], [65, 83, 97, 105, 110, 128]]target = [[68, 72, 41, 73, 74, 40], [74, 67, 110, 72, 105, 98], [93, 66, 74, 77, 83, 65], [81, 72, 117, 80, 83, 114]]
+Epoch 0: |                                                                                                                                                     | 2019/? [2:08:27<00:00,  0.26it/s, v_num=5b6z]train step 2020; scene = [['71f80e4a8d79031c'], ['aec60a6dd8090690'], ['475ee0875ccb635a']]; loss = 0.058211
+Epoch 0: |                                                                                                                                                     | 2020/? [2:08:31<00:00,  0.26it/s, v_num=5b6z]context = [[1, 4, 12, 16, 21, 23, 29, 40, 49, 54, 58, 60, 62, 74, 75, 79, 81, 82, 85, 91, 92, 94, 97, 98]]target = [[52, 69, 55, 22, 50, 24, 41, 46, 80, 58, 32, 92, 54, 81, 38, 23, 3, 29, 97, 40, 25, 39, 78, 73]]
+Epoch 0: |                                                                                                                                                     | 2029/? [2:09:06<00:00,  0.26it/s, v_num=5b6z]train step 2030; scene = [['60b6914bafa3c3f2']]; loss = 0.060865
+Epoch 0: |                                                                                                                                                     | 2030/? [2:09:10<00:00,  0.26it/s, v_num=5b6z]context = [[45, 87, 105], [54, 118, 143], [79, 120, 143], [186, 243, 262], [14, 67, 94], [27, 68, 76], [13, 28, 58], [1, 69, 80]]target = [[68, 84, 80], [134, 122, 96], [131, 97, 136], [244, 198, 209], [34, 77, 22], [70, 43, 37], [43, 32, 33], [15, 54, 19]]
+Epoch 0: |                                                                                                                                                     | 2039/? [2:09:44<00:00,  0.26it/s, v_num=5b6z]train step 2040; scene = [['8e94de9c1bc0732c'], ['0bb8a7807f7095fd']]; loss = 0.037678
+Epoch 0: |                                                                                                                                                     | 2040/? [2:09:47<00:00,  0.26it/s, v_num=5b6z]context = [[25, 79, 84, 91], [12, 32, 41, 66], [10, 17, 31, 64], [69, 71, 109, 114], [31, 49, 67, 84], [14, 27, 46, 84]]target = [[48, 80, 51, 88], [16, 41, 21, 33], [63, 61, 13, 38], [73, 107, 79, 90], [41, 35, 60, 56], [29, 80, 26, 25]]
+Epoch 0: |                                                                                                                                                     | 2049/? [2:10:22<00:00,  0.26it/s, v_num=5b6z]train step 2050; scene = [['c4e12df63403eadf'], ['8359c9726f078a38'], ['e0b75e74fdeffde9']]; loss = 0.072039
+Epoch 0: |                                                                                                                                                     | 2050/? [2:10:25<00:00,  0.26it/s, v_num=5b6z]context = [[152, 157, 160, 167, 191, 197], [71, 81, 105, 126, 127, 139], [10, 12, 47, 54, 66, 83], [85, 88, 106, 117, 127, 136]]target = [[167, 195, 187, 184, 157, 176], [120, 107, 92, 124, 115, 110], [14, 13, 80, 40, 24, 47], [93, 96, 97, 132, 102, 131]]
+Epoch 0: |                                                                                                                                                     | 2059/? [2:10:58<00:00,  0.26it/s, v_num=5b6z]train step 2060; scene = [['22062ed897320134'], ['08e4f6a5098b0d3a'], ['d3057752d15cc3ed']]; loss = 0.099077
+Epoch 0: |                                                                                                                                                     | 2060/? [2:11:02<00:00,  0.26it/s, v_num=5b6z]context = [[164, 195, 232], [18, 42, 81], [5, 55, 93], [175, 199, 220], [6, 42, 65], [2, 7, 60], [149, 190, 195], [33, 42, 82]]target = [[209, 178, 225], [53, 24, 75], [32, 35, 24], [201, 183, 204], [50, 59, 38], [5, 37, 42], [187, 161, 189], [44, 78, 62]]
+Epoch 0: |                                                                                                                                                     | 2069/? [2:11:37<00:00,  0.26it/s, v_num=5b6z]train step 2070; scene = [['85ef0eb4a42e3425'], ['703430ad773c95bc'], ['4fd9d45647d536e5']]; loss = 0.135498
+Epoch 0: |                                                                                                                                                     | 2070/? [2:11:41<00:00,  0.26it/s, v_num=5b6z]context = [[78, 84, 92, 97, 108, 109, 115, 124, 138, 150, 158, 159], [42, 43, 53, 57, 69, 71, 77, 83, 84, 86, 88, 92]]target = [[127, 113, 118, 102, 153, 144, 132, 139, 125, 129, 84, 79], [57, 81, 79, 87, 72, 61, 48, 88, 54, 84, 44, 65]]
+Epoch 0: |                                                                                                                                                     | 2079/? [2:12:15<00:00,  0.26it/s, v_num=5b6z]train step 2080; scene = [['968b857dda7e955a'], ['8e713ad26d00feac']]; loss = 0.042663
+Epoch 0: |                                                                                                                                                     | 2080/? [2:12:18<00:00,  0.26it/s, v_num=5b6z]context = [[33, 57, 88, 112], [20, 29, 43, 76], [20, 56, 96, 110], [3, 13, 50, 53], [44, 50, 64, 108], [65, 92, 120, 152]]target = [[101, 99, 69, 88], [37, 25, 36, 30], [99, 72, 91, 75], [14, 32, 43, 39], [105, 93, 56, 106], [69, 98, 70, 122]]
+Epoch 0: |                                                                                                                                                     | 2089/? [2:12:53<00:00,  0.26it/s, v_num=5b6z]train step 2090; scene = [['0565bd311bf73bbb'], ['eceeb7a49f302da9'], ['2073f379b98e47e8'], ['c5536f755d325407'], ['06ed257e33ae67f5'], ['b12f64d6002ec745']]; loss = 0.064503
+Epoch 0: |                                                                                                                                                     | 2090/? [2:12:57<00:00,  0.26it/s, v_num=5b6z]context = [[14, 20, 22, 23, 24, 26, 36, 41, 74, 81, 83, 87], [2, 10, 11, 16, 25, 26, 34, 35, 41, 49, 62, 63]]target = [[16, 23, 39, 54, 53, 59, 48, 70, 20, 63, 50, 36], [60, 32, 15, 7, 5, 51, 40, 52, 10, 56, 22, 41]]
+Epoch 0: |                                                                                                                                                     | 2099/? [2:13:32<00:00,  0.26it/s, v_num=5b6z]train step 2100; scene = [['3101309753e2f063'], ['d55fbb3dcd24f08e']]; loss = 0.035979
+Epoch 0: |                                                                                                                                                     | 2100/? [2:13:36<00:00,  0.26it/s, v_num=5b6z]context = [[145, 173, 174, 182, 187, 197, 204, 211], [65, 79, 86, 88, 93, 112, 118, 122], [98, 107, 122, 145, 156, 159, 162, 163]]target = [[208, 167, 187, 146, 173, 180, 166, 172], [108, 72, 116, 104, 92, 115, 83, 68], [160, 128, 158, 103, 134, 125, 101, 109]]
+Epoch 0: |                                                                                                                                                     | 2109/? [2:14:09<00:00,  0.26it/s, v_num=5b6z]train step 2110; scene = [['4fd151a48542df52'], ['c53e0b350f04f159'], ['ec1dcf652eae675d'], ['f001500643d191d6']]; loss = 0.125602
+Epoch 0: |                                                                                                                                                     | 2110/? [2:14:13<00:00,  0.26it/s, v_num=5b6z]context = [[42, 45, 48, 60, 81, 84, 88, 90, 102, 104, 121, 130], [22, 29, 34, 38, 39, 44, 49, 50, 53, 56, 59, 74]]target = [[108, 62, 127, 115, 43, 114, 52, 51, 100, 126, 61, 87], [64, 41, 45, 69, 25, 27, 65, 43, 68, 71, 38, 35]]
+Epoch 0: |                                                                                                                                                     | 2119/? [2:14:47<00:00,  0.26it/s, v_num=5b6z]train step 2120; scene = [['4f515c197a061a67'], ['0c7171edef36d44d'], ['ec5f9801aa83c8aa']]; loss = 0.067993
+Epoch 0: |                                                                                                                                                     | 2120/? [2:14:51<00:00,  0.26it/s, v_num=5b6z]context = [[124, 126, 135, 171], [32, 69, 83, 84], [30, 85, 88, 105], [27, 33, 42, 117], [2, 33, 53, 56], [24, 47, 86, 102]]target = [[165, 167, 132, 145], [80, 75, 52, 61], [73, 92, 100, 67], [87, 74, 112, 34], [45, 41, 49, 35], [80, 73, 95, 30]]
+Epoch 0: |                                                                                                                                                     | 2129/? [2:15:25<00:00,  0.26it/s, v_num=5b6z]train step 2130; scene = [['2dd8ae2b71457753'], ['3670e4d9d26e7534']]; loss = 0.050730
+Epoch 0: |                                                                                                                                                     | 2130/? [2:15:28<00:00,  0.26it/s, v_num=5b6z]context = [[24, 25, 27, 28, 32, 34, 36, 39, 48, 50, 55, 61, 62, 63, 64, 67, 77, 87, 102, 106, 110, 116, 117, 121]]target = [[45, 42, 59, 107, 40, 101, 31, 75, 70, 54, 81, 53, 83, 73, 64, 100, 51, 37, 47, 117, 60, 43, 48, 90]]
+Epoch 0: |                                                                                                                                                     | 2139/? [2:16:03<00:00,  0.26it/s, v_num=5b6z]train step 2140; scene = [['158ecf8d21e5af57'], ['4dfd4268f80ef274'], ['04c4eef824be2a53']]; loss = 0.051098
+Epoch 0: |                                                                                                                                                     | 2140/? [2:16:07<00:00,  0.26it/s, v_num=5b6z]context = [[131, 152, 157, 167, 188, 191], [139, 146, 149, 153, 171, 223], [126, 136, 156, 187, 188, 202], [32, 37, 41, 56, 82, 87]]target = [[189, 157, 139, 148, 184, 180], [151, 183, 209, 179, 185, 140], [159, 166, 155, 170, 182, 131], [78, 79, 34, 85, 50, 59]]
+Epoch 0: |                                                                                                                                                     | 2149/? [2:16:41<00:00,  0.26it/s, v_num=5b6z]train step 2150; scene = [['db62795c284bf764'], ['916b86f95631b480']]; loss = 0.066782
+Epoch 0: |                                                                                                                                                     | 2150/? [2:16:45<00:00,  0.26it/s, v_num=5b6z]context = [[4, 6, 16, 19, 21, 30, 35, 38, 41, 48, 51, 52, 54, 62, 63, 67, 69, 77, 80, 81, 86, 93, 95, 101]]target = [[56, 16, 36, 50, 25, 34, 65, 87, 40, 5, 42, 54, 18, 32, 68, 76, 98, 52, 37, 85, 41, 79, 13, 91]]
+Epoch 0: |                                                                                                                                                     | 2159/? [2:17:20<00:00,  0.26it/s, v_num=5b6z]train step 2160; scene = [['72077f7ff39fc73e'], ['9ae93c878c7dbe8a'], ['4bac79c3a17ed149'], ['208bd0af6bc8937b'], ['84e6b90f0c2567d8'], ['f4ba6d204cb14df7'], ['758770f2884e9a79'], ['36df585860d0ad88'], ['69a6e3951e138ca8'], ['5d986af113fbac56'], ['d46599d6e4a2b451'], ['ee020a8773034321']]; loss = 0.086925
+Epoch 0: |                                                                                                                                                     | 2160/? [2:17:24<00:00,  0.26it/s, v_num=5b6z]context = [[6, 8, 10, 50, 66, 87], [50, 69, 72, 95, 105, 111], [126, 165, 170, 172, 176, 211], [5, 9, 19, 42, 47, 50]]target = [[9, 68, 72, 19, 24, 67], [80, 85, 64, 70, 104, 91], [184, 172, 136, 189, 179, 175], [48, 14, 45, 22, 40, 16]]
+Epoch 0: |                                                                                                                                                     | 2169/? [2:17:58<00:00,  0.26it/s, v_num=5b6z]train step 2170; scene = [['e1d9afaa8899ee32'], ['1060ca933281b55c'], ['fbf253fca4a29e87'], ['5ff87250a0eb913b'], ['d02e6b104723d39a'], ['42b208082fce3bc2'], ['ee6e5709a57be759'], ['a191c34eb75bbaec'], ['0cbafecfcb0f7727'], ['e6461cee8a9474d5'], ['fda8bac8ddac590f'], ['465fa8314b741006']]; loss = 0.119943
+Epoch 0: |                                                                                                                                                     | 2170/? [2:18:02<00:00,  0.26it/s, v_num=5b6z]context = [[5, 20, 23, 37, 49, 52, 59, 76], [11, 28, 31, 35, 41, 43, 44, 57], [5, 15, 30, 36, 47, 59, 61, 62]]target = [[53, 50, 40, 18, 39, 33, 23, 17], [27, 38, 40, 28, 55, 45, 26, 25], [56, 9, 61, 23, 59, 30, 45, 35]]
+Epoch 0: |                                                                                                                                                     | 2179/? [2:18:35<00:00,  0.26it/s, v_num=5b6z]train step 2180; scene = [['91599919681fac69'], ['f9cece9ebde532d0']]; loss = 0.054116
+Epoch 0: |                                                                                                                                                     | 2180/? [2:18:39<00:00,  0.26it/s, v_num=5b6z]context = [[46, 63, 73, 75, 76, 83, 99, 102], [17, 31, 32, 39, 47, 52, 58, 68], [83, 91, 110, 113, 116, 117, 121, 133]]target = [[72, 92, 67, 101, 68, 55, 49, 74], [30, 31, 53, 26, 41, 35, 20, 62], [93, 131, 116, 96, 125, 107, 127, 109]]
+Epoch 0: |                                                                                                                                                     | 2189/? [2:19:14<00:00,  0.26it/s, v_num=5b6z]train step 2190; scene = [['4d48befa72535f0a']]; loss = 0.056361
+Epoch 0: |                                                                                                                                                     | 2190/? [2:19:18<00:00,  0.26it/s, v_num=5b6z]context = [[113, 123, 125, 126, 132, 133, 135, 138, 149, 150, 154, 160, 167, 172, 175, 191, 194, 197, 198, 200, 201, 205, 206, 210]]target = [[120, 130, 147, 187, 142, 115, 202, 171, 117, 163, 191, 170, 139, 169, 181, 155, 150, 141, 192, 146, 188, 174, 205, 179]]
+Epoch 0: |                                                                                                                                                     | 2199/? [2:19:52<00:00,  0.26it/s, v_num=5b6z]train step 2200; scene = [['76ce9bd95ed81200'], ['91ac7d7027dcc46d'], ['3c3e1619744887ca'], ['003d2563b3c1023e']]; loss = 0.051504
+Epoch 0: |                                                                                                                                                     | 2200/? [2:19:56<00:00,  0.26it/s, v_num=5b6z]context = [[11, 13, 17, 19, 42, 43, 50, 51, 58, 60, 61, 63, 65, 71, 74, 77, 82, 85, 89, 95, 99, 101, 103, 108]]target = [[61, 31, 48, 27, 88, 78, 23, 53, 24, 93, 63, 73, 20, 32, 55, 35, 29, 75, 21, 76, 49, 42, 28, 40]]
+[2026-02-24 21:35:22,840][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2209/? [2:20:37<00:00,  0.26it/s, v_num=5b6z]train step 2210; scene = [['40873337bb9786ab'], ['4f354a704f70f53b'], ['57cea8a71cab48cf'], ['19c383ad9f21d0d3'], ['a98ec507d448dea2'], ['b663a474fd1e4ca4']]; loss = 0.083505
+Epoch 0: |                                                                                                                                                     | 2210/? [2:20:40<00:00,  0.26it/s, v_num=5b6z]context = [[9, 12, 24, 32, 37, 38, 44, 46, 51, 65, 71, 75], [87, 89, 90, 92, 104, 114, 118, 119, 120, 122, 143, 144]]target = [[61, 31, 40, 37, 52, 13, 14, 68, 57, 25, 70, 54], [95, 130, 128, 134, 91, 114, 133, 102, 129, 98, 138, 116]]
+Epoch 0: |                                                                                                                                                     | 2219/? [2:21:14<00:00,  0.26it/s, v_num=5b6z]train step 2220; scene = [['d85f0322f215aa54'], ['1318e656455d56d2'], ['e5f41c9021b0b7c1']]; loss = 0.058452
+Epoch 0: |                                                                                                                                                     | 2220/? [2:21:18<00:00,  0.26it/s, v_num=5b6z]context = [[12, 17, 27, 29, 36, 37, 44, 50, 52, 54, 55, 65, 75, 80, 81, 83, 84, 85, 86, 89, 91, 94, 95, 109]]target = [[83, 45, 76, 92, 104, 41, 30, 59, 79, 28, 86, 38, 85, 64, 66, 51, 106, 15, 89, 16, 46, 108, 99, 33]]
+Epoch 0: |                                                                                                                                                     | 2229/? [2:21:53<00:00,  0.26it/s, v_num=5b6z]train step 2230; scene = [['9af18a1c4c45a179']]; loss = 0.067387
+Epoch 0: |                                                                                                                                                     | 2230/? [2:21:57<00:00,  0.26it/s, v_num=5b6z]context = [[40, 43, 44, 61, 67, 70, 74, 77, 79, 85, 87, 92, 93, 98, 100, 103, 108, 115, 117, 123, 128, 134, 135, 137]]target = [[119, 88, 102, 62, 79, 131, 53, 122, 46, 128, 130, 59, 54, 76, 71, 57, 107, 126, 93, 65, 72, 115, 69, 66]]
+Epoch 0: |                                                                                                                                                     | 2239/? [2:22:31<00:00,  0.26it/s, v_num=5b6z]train step 2240; scene = [['e454026f5348630e'], ['d6a1f3e13c45df99'], ['4dd9c5fab7e6ec75'], ['352d2bdc1900b5e0']]; loss = 0.046533
+Epoch 0: |                                                                                                                                                     | 2240/? [2:22:34<00:00,  0.26it/s, v_num=5b6z]context = [[5, 53], [14, 60], [21, 91], [60, 113], [2, 85], [31, 99], [27, 81], [3, 54], [15, 101], [11, 69], [0, 45], [193, 272]]target = [[39, 37], [57, 23], [71, 41], [71, 104], [30, 23], [39, 89], [58, 57], [31, 40], [83, 33], [61, 53], [28, 1], [235, 232]]
+Epoch 0: |                                                                                                                                                     | 2249/? [2:23:09<00:00,  0.26it/s, v_num=5b6z]train step 2250; scene = [['5afa3097bb38b159'], ['c0f67af5cd34e8d8']]; loss = 0.050247
+Epoch 0: |                                                                                                                                                     | 2250/? [2:23:13<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 2250; scene = ['651a7f83ed093001'];
+target intrinsic:  tensor(0.8796, device='cuda:0') tensor(0.8798, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9039, device='cuda:0') tensor(0.9064, device='cuda:0')
+[2026-02-24 21:38:36,652][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2250/? [2:23:14<00:00,  0.26it/s, v_num=5b6z]context = [[0, 21, 34, 39, 59, 63, 83, 84], [23, 26, 33, 34, 53, 58, 59, 68], [118, 137, 141, 150, 154, 170, 188, 192]]target = [[14, 79, 44, 19, 83, 18, 29, 23], [48, 24, 60, 52, 35, 61, 29, 64], [136, 187, 135, 171, 137, 131, 147, 125]]
+Epoch 0: |                                                                                                                                                     | 2259/? [2:23:47<00:00,  0.26it/s, v_num=5b6z]train step 2260; scene = [['39cafa06c9c431dc']]; loss = 0.063293
+Epoch 0: |                                                                                                                                                     | 2260/? [2:23:51<00:00,  0.26it/s, v_num=5b6z]context = [[46, 51, 80, 81, 83, 97, 105, 114], [2, 9, 10, 12, 28, 29, 51, 70], [38, 41, 42, 46, 69, 72, 118, 124]]target = [[60, 58, 103, 89, 80, 51, 99, 109], [25, 68, 19, 67, 43, 14, 52, 15], [91, 55, 98, 96, 74, 116, 83, 113]]
+Epoch 0: |                                                                                                                                                     | 2269/? [2:24:26<00:00,  0.26it/s, v_num=5b6z]train step 2270; scene = [['10a2d0db5c8c0962']]; loss = 0.039546
+Epoch 0: |                                                                                                                                                     | 2270/? [2:24:29<00:00,  0.26it/s, v_num=5b6z]context = [[33, 39, 52, 59, 85, 90, 93, 112], [16, 18, 22, 27, 40, 46, 55, 70], [2, 32, 40, 41, 42, 46, 51, 71]]target = [[40, 55, 82, 107, 111, 35, 102, 36], [25, 31, 19, 58, 55, 64, 62, 28], [48, 13, 68, 46, 30, 12, 42, 40]]
+Epoch 0: |                                                                                                                                                     | 2279/? [2:25:04<00:00,  0.26it/s, v_num=5b6z]train step 2280; scene = [['bcbd422c5506815b'], ['dd1a25142bac8d29'], ['ba3e38a071a523f0'], ['2248f7eaf18cca06'], ['d0edef000b9ca743'], ['d9aacea21e110a52']]; loss = 0.067320
+Epoch 0: |                                                                                                                                                     | 2280/? [2:25:08<00:00,  0.26it/s, v_num=5b6z]context = [[103, 107, 133, 140, 142, 146, 151, 159, 168, 169, 172, 177], [2, 4, 9, 11, 12, 17, 28, 38, 40, 46, 56, 59]]target = [[132, 133, 104, 118, 122, 149, 173, 175, 125, 120, 168, 164], [39, 49, 57, 47, 31, 13, 28, 52, 21, 16, 55, 29]]
+Epoch 0: |                                                                                                                                                     | 2289/? [2:25:42<00:00,  0.26it/s, v_num=5b6z]train step 2290; scene = [['fa6cb78e1503d22c'], ['f592013dce557cce']]; loss = 0.049504
+Epoch 0: |                                                                                                                                                     | 2290/? [2:25:46<00:00,  0.26it/s, v_num=5b6z]context = [[30, 35, 41, 49, 50, 59, 65, 78, 89, 91, 113, 116], [126, 127, 128, 134, 152, 160, 162, 166, 171, 172, 176, 177]]target = [[51, 61, 81, 100, 48, 110, 54, 112, 60, 113, 86, 55], [140, 169, 144, 127, 174, 150, 129, 143, 163, 132, 173, 152]]
+Epoch 0: |                                                                                                                                                     | 2299/? [2:26:20<00:00,  0.26it/s, v_num=5b6z]train step 2300; scene = [['e23f945703f61bfc']]; loss = 0.052567
+Epoch 0: |                                                                                                                                                     | 2300/? [2:26:24<00:00,  0.26it/s, v_num=5b6z]context = [[73, 94, 106, 123, 128, 130], [6, 22, 33, 42, 45, 51], [165, 166, 167, 196, 205, 228], [98, 109, 112, 134, 135, 152]]target = [[128, 88, 106, 120, 129, 82], [23, 48, 38, 26, 11, 14], [208, 193, 188, 218, 183, 210], [134, 143, 122, 104, 117, 137]]
+Epoch 0: |                                                                                                                                                     | 2309/? [2:26:59<00:00,  0.26it/s, v_num=5b6z]train step 2310; scene = [['347f4e7f9f627a12'], ['d1d992e581136ac6'], ['59e7800dec3fa9f5'], ['cf6a4349d0ffdfcf']]; loss = 0.047780
+Epoch 0: |                                                                                                                                                     | 2310/? [2:27:02<00:00,  0.26it/s, v_num=5b6z]context = [[0, 5, 8, 13, 15, 43, 47, 51, 61, 68, 72, 74], [8, 13, 17, 27, 29, 38, 49, 53, 56, 60, 63, 65]]target = [[5, 42, 38, 57, 31, 53, 36, 35, 59, 55, 2, 11], [23, 48, 28, 17, 61, 47, 12, 64, 36, 9, 30, 57]]
+Epoch 0: |                                                                                                                                                     | 2319/? [2:27:36<00:00,  0.26it/s, v_num=5b6z]train step 2320; scene = [['60c37b519a01205d'], ['b0ee123a8cfc8e62'], ['c5b4562390525d10'], ['8b950107c02ffaa9']]; loss = 0.067367
+Epoch 0: |                                                                                                                                                     | 2320/? [2:27:40<00:00,  0.26it/s, v_num=5b6z]context = [[8, 11, 13, 15, 17, 25, 29, 39, 40, 45, 48, 58], [4, 18, 21, 36, 47, 50, 60, 63, 71, 76, 79, 82]]target = [[56, 37, 21, 47, 26, 51, 36, 12, 49, 27, 24, 16], [62, 43, 80, 40, 30, 39, 37, 19, 6, 13, 35, 10]]
+Epoch 0: |                                                                                                                                                     | 2329/? [2:28:12<00:00,  0.26it/s, v_num=5b6z]train step 2330; scene = [['5c9b898102b16eae'], ['583ab14553881ee8'], ['e204af0947f704ad'], ['39edfd183c4f5b5b'], ['a85c79f15c396d71'], ['3d7a200dab472990']]; loss = 0.062477
+Epoch 0: |                                                                                                                                                     | 2330/? [2:28:16<00:00,  0.26it/s, v_num=5b6z]context = [[9, 23, 91], [5, 23, 61], [94, 164, 165], [149, 165, 237], [73, 137, 159], [1, 13, 48], [5, 55, 59], [25, 55, 75]]target = [[33, 46, 62], [42, 29, 22], [125, 112, 119], [187, 202, 170], [139, 87, 131], [7, 2, 23], [57, 39, 50], [62, 34, 56]]
+Epoch 0: |                                                                                                                                                     | 2339/? [2:28:51<00:00,  0.26it/s, v_num=5b6z]train step 2340; scene = [['c4005922f59686ae']]; loss = 0.047818
+Epoch 0: |                                                                                                                                                     | 2340/? [2:28:55<00:00,  0.26it/s, v_num=5b6z]context = [[9, 13, 14, 15, 29, 42, 47, 48, 52, 54, 58, 60, 61, 68, 76, 77, 81, 88, 89, 97, 98, 99, 105, 106]]target = [[72, 73, 28, 65, 82, 67, 96, 56, 75, 94, 16, 101, 63, 80, 83, 33, 41, 61, 93, 100, 14, 86, 32, 11]]
+Epoch 0: |                                                                                                                                                     | 2349/? [2:29:30<00:00,  0.26it/s, v_num=5b6z]train step 2350; scene = [['009e573e59c8c393'], ['7c48a30f23ea42c3'], ['739123afdcc19a64'], ['ddc11c891f471dd0'], ['38b7b864ae7b21e9'], ['997656bdb430ad43']]; loss = 0.053913
+Epoch 0: |                                                                                                                                                     | 2350/? [2:29:34<00:00,  0.26it/s, v_num=5b6z]context = [[115, 120, 127, 130, 134, 143, 147, 148, 149, 150, 159, 163, 173, 175, 184, 186, 187, 188, 190, 195, 199, 202, 205, 212]]target = [[148, 147, 199, 184, 190, 152, 134, 189, 127, 120, 201, 163, 186, 165, 171, 193, 117, 210, 183, 211, 195, 156, 204, 145]]
+Epoch 0: |                                                                                                                                                     | 2359/? [2:30:07<00:00,  0.26it/s, v_num=5b6z]train step 2360; scene = [['9a447c89080e9b56'], ['9eb8e7f262b10c23'], ['c2b8b3e74c64553a'], ['b6f9cfe435a0fde7']]; loss = 0.052806
+Epoch 0: |                                                                                                                                                     | 2360/? [2:30:11<00:00,  0.26it/s, v_num=5b6z]context = [[1, 4, 8, 16, 17, 28, 40, 45, 51, 53, 60, 62], [1, 4, 5, 11, 12, 13, 14, 15, 17, 29, 40, 82]]target = [[8, 52, 12, 39, 55, 43, 33, 16, 53, 18, 25, 2], [71, 5, 21, 6, 62, 55, 45, 49, 37, 14, 27, 56]]
+Epoch 0: |                                                                                                                                                     | 2369/? [2:30:45<00:00,  0.26it/s, v_num=5b6z]train step 2370; scene = [['090ced9a667843cb']]; loss = 0.042516
+Epoch 0: |                                                                                                                                                     | 2370/? [2:30:49<00:00,  0.26it/s, v_num=5b6z]context = [[13, 19, 35, 37, 43, 61, 92, 94], [0, 8, 19, 21, 24, 25, 27, 45], [1, 5, 38, 42, 45, 47, 75, 82]]target = [[46, 77, 90, 47, 83, 21, 76, 54], [3, 11, 9, 33, 20, 41, 42, 4], [72, 28, 74, 38, 52, 40, 35, 34]]
+Epoch 0: |                                                                                                                                                     | 2379/? [2:31:23<00:00,  0.26it/s, v_num=5b6z]train step 2380; scene = [['8ef9ff3189c85eee'], ['7e8630a890a85545']]; loss = 0.052316
+Epoch 0: |                                                                                                                                                     | 2380/? [2:31:27<00:00,  0.26it/s, v_num=5b6z]context = [[48, 55, 58, 63, 67, 68, 71, 73, 75, 76, 79, 80, 86, 93, 107, 108, 109, 118, 119, 120, 129, 133, 135, 145]]target = [[133, 59, 121, 87, 96, 110, 67, 101, 49, 142, 60, 134, 109, 132, 98, 122, 103, 100, 80, 57, 65, 86, 128, 143]]
+Epoch 0: |                                                                                                                                                     | 2389/? [2:32:01<00:00,  0.26it/s, v_num=5b6z]train step 2390; scene = [['343a98bd8cfda2de'], ['c09d7898ef37ba32']]; loss = 0.055488
+Epoch 0: |                                                                                                                                                     | 2390/? [2:32:05<00:00,  0.26it/s, v_num=5b6z]context = [[19, 20, 31, 32, 41, 46, 49, 50, 51, 53, 59, 60, 62, 63, 74, 77, 82, 84, 85, 91, 93, 97, 98, 116]]target = [[56, 31, 77, 102, 39, 54, 21, 57, 66, 75, 100, 115, 26, 53, 34, 113, 97, 90, 25, 82, 108, 60, 67, 38]]
+Epoch 0: |                                                                                                                                                     | 2399/? [2:32:39<00:00,  0.26it/s, v_num=5b6z]train step 2400; scene = [['8d758914077e5926']]; loss = 0.051131
+Epoch 0: |                                                                                                                                                     | 2400/? [2:32:43<00:00,  0.26it/s, v_num=5b6z]context = [[114, 117, 119, 120, 131, 136, 141, 148, 149, 151, 163, 168], [44, 47, 52, 76, 77, 78, 85, 88, 89, 90, 97, 99]]target = [[141, 153, 135, 128, 126, 119, 140, 147, 161, 120, 155, 127], [82, 75, 49, 62, 84, 54, 78, 70, 55, 72, 96, 92]]
+[2026-02-24 21:48:09,908][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2409/? [2:33:19<00:00,  0.26it/s, v_num=5b6z]train step 2410; scene = [['d35650707515c2cf']]; loss = 0.050326
+Epoch 0: |                                                                                                                                                     | 2410/? [2:33:22<00:00,  0.26it/s, v_num=5b6z]context = [[4, 8, 16, 30, 45, 53, 70, 77], [123, 139, 144, 149, 166, 184, 187, 207], [38, 44, 51, 66, 68, 90, 94, 98]]target = [[7, 17, 15, 70, 47, 48, 35, 72], [190, 168, 157, 135, 200, 160, 161, 128], [84, 75, 51, 67, 64, 46, 95, 49]]
+Epoch 0: |                                                                                                                                                     | 2419/? [2:33:57<00:00,  0.26it/s, v_num=5b6z]train step 2420; scene = [['ea5def9e91076788']]; loss = 0.038031
+Epoch 0: |                                                                                                                                                     | 2420/? [2:34:01<00:00,  0.26it/s, v_num=5b6z]context = [[93, 107, 110, 121, 127, 148], [165, 169, 170, 208, 209, 214], [0, 6, 7, 23, 45, 83], [12, 33, 38, 45, 81, 91]]target = [[131, 99, 135, 127, 142, 120], [167, 177, 176, 213, 171, 191], [22, 45, 64, 29, 67, 51], [47, 31, 65, 87, 60, 85]]
+Epoch 0: |                                                                                                                                                     | 2429/? [2:34:35<00:00,  0.26it/s, v_num=5b6z]train step 2430; scene = [['45fe4bbb2526ca6e'], ['0db2c3f2775880de'], ['4a0f95a3db913b56']]; loss = 0.035424
+Epoch 0: |                                                                                                                                                     | 2430/? [2:34:39<00:00,  0.26it/s, v_num=5b6z]context = [[22, 66, 68, 80, 92, 99], [26, 44, 52, 76, 80, 86], [3, 38, 40, 69, 77, 89], [43, 83, 90, 109, 121, 126]]target = [[93, 85, 61, 37, 46, 63], [48, 57, 64, 35, 27, 84], [43, 41, 48, 23, 25, 67], [111, 55, 61, 105, 114, 116]]
+Epoch 0: |                                                                                                                                                     | 2439/? [2:35:13<00:00,  0.26it/s, v_num=5b6z]train step 2440; scene = [['512f1df8266be984'], ['50c8a233bd82a613']]; loss = 0.068842
+Epoch 0: |                                                                                                                                                     | 2440/? [2:35:17<00:00,  0.26it/s, v_num=5b6z]context = [[96, 106, 114, 124, 131, 158], [48, 54, 56, 57, 87, 102], [2, 8, 16, 43, 50, 59], [22, 27, 37, 53, 69, 77]]target = [[112, 148, 103, 120, 104, 132], [72, 96, 86, 90, 52, 98], [46, 13, 56, 3, 57, 55], [58, 57, 71, 67, 28, 45]]
+Epoch 0: |                                                                                                                                                     | 2449/? [2:35:50<00:00,  0.26it/s, v_num=5b6z]train step 2450; scene = [['ac584e6fc77676a4'], ['19a3235d680e11c4'], ['61b9def6cdf00024'], ['fc86266e2fcb72fd']]; loss = 0.048387
+Epoch 0: |                                                                                                                                                     | 2450/? [2:35:54<00:00,  0.26it/s, v_num=5b6z]context = [[16, 41, 71], [43, 46, 97], [10, 18, 90], [3, 16, 53], [143, 187, 198], [25, 36, 70], [98, 136, 179], [19, 65, 78]]target = [[45, 18, 43], [95, 96, 51], [32, 21, 70], [5, 8, 27], [146, 154, 151], [28, 31, 48], [133, 137, 151], [33, 53, 64]]
+Epoch 0: |                                                                                                                                                     | 2459/? [2:36:29<00:00,  0.26it/s, v_num=5b6z]train step 2460; scene = [['dc5f38f005c3ebd6']]; loss = 0.077075
+Epoch 0: |                                                                                                                                                     | 2460/? [2:36:32<00:00,  0.26it/s, v_num=5b6z]context = [[4, 5, 11, 12, 16, 17, 19, 22, 23, 30, 39, 44, 49, 50, 51, 61, 77, 82, 87, 89, 95, 96, 97, 101]]target = [[53, 36, 13, 57, 86, 80, 97, 72, 84, 27, 61, 94, 35, 51, 54, 67, 43, 66, 62, 89, 7, 92, 10, 47]]
+Epoch 0: |                                                                                                                                                     | 2469/? [2:37:07<00:00,  0.26it/s, v_num=5b6z]train step 2470; scene = [['c7620995ebe9c4e6'], ['06241ebed1658f34'], ['1f214117250f089a'], ['5e31e0691d426537'], ['11c4a7d67bc2629e'], ['715e8695976cdb61'], ['6888f7ca14081419'], ['8fe341dcd0880bd5']]; loss = 0.071062
+Epoch 0: |                                                                                                                                                     | 2470/? [2:37:11<00:00,  0.26it/s, v_num=5b6z]context = [[17, 28, 38, 42, 68, 86], [2, 4, 28, 37, 51, 58], [7, 21, 24, 54, 64, 70], [41, 42, 61, 71, 81, 89]]target = [[55, 30, 52, 28, 44, 76], [22, 36, 3, 57, 34, 9], [16, 35, 34, 61, 56, 43], [56, 87, 85, 73, 68, 50]]
+Epoch 0: |                                                                                                                                                     | 2479/? [2:37:45<00:00,  0.26it/s, v_num=5b6z]train step 2480; scene = [['cf98d3219d144500']]; loss = 0.068008
+Epoch 0: |                                                                                                                                                     | 2480/? [2:37:49<00:00,  0.26it/s, v_num=5b6z]context = [[151, 158, 171, 173, 176, 205, 206, 230], [74, 88, 94, 98, 119, 130, 146, 163], [5, 8, 33, 42, 52, 53, 57, 59]]target = [[177, 223, 212, 220, 205, 207, 225, 222], [102, 132, 115, 112, 145, 156, 119, 100], [12, 23, 49, 20, 55, 32, 10, 39]]
+Epoch 0: |                                                                                                                                                     | 2489/? [2:38:23<00:00,  0.26it/s, v_num=5b6z]train step 2490; scene = [['ebff5d05f1bb086f'], ['880427ff150f7b4d'], ['9dd0efd4b4626604'], ['f52c70025f40e56d']]; loss = 0.057226
+Epoch 0: |                                                                                                                                                     | 2490/? [2:38:27<00:00,  0.26it/s, v_num=5b6z]context = [[19, 27, 31, 35, 36, 42, 57, 60, 63, 64, 65, 70], [21, 26, 31, 32, 50, 56, 62, 67, 74, 81, 90, 93]]target = [[51, 33, 64, 68, 53, 29, 54, 27, 26, 69, 35, 48], [59, 60, 83, 76, 86, 68, 61, 36, 71, 22, 69, 82]]
+Epoch 0: |                                                                                                                                                     | 2499/? [2:39:00<00:00,  0.26it/s, v_num=5b6z]train step 2500; scene = [['7dc5c394263df267']]; loss = 0.057663
+Epoch 0: |                                                                                                                                                     | 2500/? [2:39:03<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 2500; scene = ['97ef4323919c5e8a'];
+target intrinsic:  tensor(0.8889, device='cuda:0') tensor(0.8892, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(1.0102, device='cuda:0') tensor(1.0074, device='cuda:0')
+[2026-02-24 21:54:26,886][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2500/? [2:39:04<00:00,  0.26it/s, v_num=5b6z]context = [[17, 21, 30, 44, 52, 84], [24, 34, 39, 45, 53, 84], [159, 163, 179, 192, 194, 207], [115, 116, 136, 154, 172, 188]]target = [[31, 20, 44, 37, 73, 59], [47, 48, 59, 36, 52, 44], [198, 165, 170, 193, 188, 175], [173, 128, 152, 148, 158, 172]]
+Epoch 0: |                                                                                                                                                     | 2509/? [2:39:38<00:00,  0.26it/s, v_num=5b6z]train step 2510; scene = [['e806eb496df1dfa0'], ['3caa89905a59e7e6'], ['3fc603b4c4531c11'], ['08da23838ee6e23b'], ['89a0f0553f7f2bdb'], ['5080347450bbcb08']]; loss = 0.054380
+Epoch 0: |                                                                                                                                                     | 2510/? [2:39:42<00:00,  0.26it/s, v_num=5b6z]context = [[7, 16, 17, 33, 34, 42, 43, 45, 48, 49, 52, 60], [16, 27, 29, 31, 38, 42, 58, 64, 71, 89, 93, 104]]target = [[59, 11, 58, 34, 22, 51, 55, 46, 9, 12, 31, 56], [90, 95, 100, 19, 82, 44, 96, 91, 97, 28, 88, 77]]
+Epoch 0: |                                                                                                                                                     | 2519/? [2:40:17<00:00,  0.26it/s, v_num=5b6z]train step 2520; scene = [['0a94493d6f4cd4be'], ['1c3a8fcf1547dcf7'], ['c34950273c4d538b'], ['397652e91be52496'], ['d9a133a4747493d2'], ['46dc3900c302593a'], ['e281213868014707'], ['83c959de5ae5b86c']]; loss = 0.076510
+Epoch 0: |                                                                                                                                                     | 2520/? [2:40:21<00:00,  0.26it/s, v_num=5b6z]context = [[22, 26, 34, 62, 68, 89, 103, 111], [16, 23, 45, 57, 69, 77, 101, 105], [9, 21, 23, 24, 56, 60, 66, 68]]target = [[29, 88, 85, 106, 103, 79, 65, 23], [61, 100, 46, 104, 23, 66, 18, 40], [25, 38, 15, 16, 64, 45, 34, 30]]
+Epoch 0: |                                                                                                                                                     | 2529/? [2:40:55<00:00,  0.26it/s, v_num=5b6z]train step 2530; scene = [['871ec8d951a81c62']]; loss = 0.049166
+Epoch 0: |                                                                                                                                                     | 2530/? [2:40:59<00:00,  0.26it/s, v_num=5b6z]context = [[39, 41, 46, 48, 52, 53, 55, 66, 68, 70, 83, 92, 93, 97, 103, 106, 110, 111, 118, 124, 125, 131, 133, 136]]target = [[125, 130, 53, 128, 124, 75, 61, 112, 64, 70, 99, 120, 57, 47, 69, 60, 43, 85, 79, 63, 45, 135, 104, 129]]
+Epoch 0: |                                                                                                                                                     | 2539/? [2:41:31<00:00,  0.26it/s, v_num=5b6z]train step 2540; scene = [['8e78ce55e3180547']]; loss = 0.047604
+Epoch 0: |                                                                                                                                                     | 2540/? [2:41:35<00:00,  0.26it/s, v_num=5b6z]context = [[19, 22, 26, 27, 29, 35, 40, 42, 54, 59, 60, 61, 64, 74, 77, 86, 92, 102, 106, 109, 110, 111, 115, 116]]target = [[89, 69, 46, 77, 65, 57, 110, 49, 70, 97, 38, 58, 31, 40, 101, 74, 20, 88, 92, 90, 80, 28, 94, 37]]
+Epoch 0: |                                                                                                                                                     | 2549/? [2:42:09<00:00,  0.26it/s, v_num=5b6z]train step 2550; scene = [['6dfaec91f745fdd9'], ['3f72021b93b6224a'], ['36be66d194d57ec8']]; loss = 0.070550
+Epoch 0: |                                                                                                                                                     | 2550/? [2:42:13<00:00,  0.26it/s, v_num=5b6z]context = [[47, 51, 57, 66, 73, 85, 93, 96, 97, 107, 113, 121], [26, 41, 43, 44, 54, 64, 69, 75, 78, 82, 85, 86]]target = [[83, 71, 53, 103, 110, 111, 55, 70, 94, 72, 74, 106], [32, 47, 85, 36, 49, 51, 41, 62, 35, 84, 31, 28]]
+Epoch 0: |                                                                                                                                                     | 2559/? [2:42:46<00:00,  0.26it/s, v_num=5b6z]train step 2560; scene = [['7c6d160c26de6887'], ['5ef6fe9ef5309457'], ['66de8729fe760dd8'], ['faba60084d22aa27']]; loss = 0.049004
+Epoch 0: |                                                                                                                                                     | 2560/? [2:42:50<00:00,  0.26it/s, v_num=5b6z]context = [[44, 89, 97], [157, 230, 235], [6, 8, 64], [19, 33, 76], [24, 39, 91], [14, 17, 65], [34, 59, 83], [19, 28, 77]]target = [[93, 77, 73], [195, 166, 202], [62, 61, 20], [61, 65, 30], [82, 90, 72], [32, 60, 21], [67, 74, 43], [39, 27, 37]]
+Epoch 0: |                                                                                                                                                     | 2569/? [2:43:24<00:00,  0.26it/s, v_num=5b6z]train step 2570; scene = [['678d8464781a3de2'], ['9ab12a31f2a3b9fb']]; loss = 0.048516
+Epoch 0: |                                                                                                                                                     | 2570/? [2:43:28<00:00,  0.26it/s, v_num=5b6z]context = [[41, 92], [2, 64], [133, 190], [7, 81], [6, 63], [33, 93], [65, 146], [18, 69], [15, 63], [13, 78], [4, 79], [66, 126]]target = [[90, 71], [12, 37], [175, 173], [56, 25], [7, 56], [61, 69], [132, 93], [28, 45], [17, 31], [21, 61], [69, 76], [120, 115]]
+Epoch 0: |                                                                                                                                                     | 2579/? [2:44:02<00:00,  0.26it/s, v_num=5b6z]train step 2580; scene = [['c40b23830f7437d5'], ['a0c7cabb66c795a4'], ['df63cd7eb6e92486']]; loss = 0.061765
+Epoch 0: |                                                                                                                                                     | 2580/? [2:44:05<00:00,  0.26it/s, v_num=5b6z]context = [[110, 115, 117, 128, 136, 137, 138, 141, 143, 145, 150, 168], [7, 10, 23, 30, 35, 48, 57, 59, 63, 64, 65, 69]]target = [[128, 149, 154, 126, 167, 132, 162, 151, 112, 134, 133, 131], [18, 25, 50, 30, 12, 19, 47, 39, 54, 35, 37, 56]]
+Epoch 0: |                                                                                                                                                     | 2589/? [2:44:39<00:00,  0.26it/s, v_num=5b6z]train step 2590; scene = [['65cdcbcb16f0ebe5'], ['4b12a530a5ab03d0']]; loss = 0.038755
+Epoch 0: |                                                                                                                                                     | 2590/? [2:44:43<00:00,  0.26it/s, v_num=5b6z]context = [[32, 42, 45, 59, 89, 100], [65, 89, 96, 104, 111, 139], [177, 197, 217, 235, 253, 264], [81, 94, 117, 131, 132, 142]]target = [[41, 75, 46, 34, 67, 60], [125, 90, 94, 114, 92, 95], [196, 223, 250, 229, 245, 259], [124, 91, 129, 94, 85, 131]]
+Epoch 0: |                                                                                                                                                     | 2599/? [2:45:18<00:00,  0.26it/s, v_num=5b6z]train step 2600; scene = [['cb3bac70297d52c0']]; loss = 0.042353
+Epoch 0: |                                                                                                                                                     | 2600/? [2:45:22<00:00,  0.26it/s, v_num=5b6z]context = [[3, 4, 7, 15, 37, 44, 46, 49], [71, 83, 85, 107, 113, 118, 123, 135], [4, 6, 26, 29, 32, 45, 51, 52]]target = [[38, 11, 27, 26, 16, 48, 39, 30], [81, 90, 94, 125, 75, 97, 110, 98], [22, 27, 29, 35, 7, 9, 49, 21]]
+[2026-02-24 22:00:48,561][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2609/? [2:45:56<00:00,  0.26it/s, v_num=5b6z]train step 2610; scene = [['262404ef111a096b'], ['5ac9dbde9cf64eed'], ['c32ee937dbec7221'], ['201e44ad015728c7'], ['1fb562b09fc361ea'], ['e454026f5348630e'], ['d1772c09b4b6d95f'], ['8e7cc5c956a4cf95']]; loss = 0.062483
+Epoch 0: |                                                                                                                                                     | 2610/? [2:45:59<00:00,  0.26it/s, v_num=5b6z]context = [[70, 73, 74, 76, 77, 79, 82, 88, 109, 110, 115, 124, 125, 136, 141, 148, 150, 151, 155, 158, 159, 160, 161, 167]]target = [[156, 121, 159, 150, 128, 138, 162, 81, 102, 71, 103, 74, 137, 144, 108, 166, 72, 161, 129, 110, 165, 136, 106, 99]]
+Epoch 0: |                                                                                                                                                     | 2619/? [2:46:33<00:00,  0.26it/s, v_num=5b6z]train step 2620; scene = [['4dd74d4d53abb812'], ['9029d8d8f6f0e98a'], ['360253c55aef692f'], ['b07a54eda9cdaccb'], ['ca553e70eff3aa6f'], ['70b45bbd1147fbf0']]; loss = 0.043530
+Epoch 0: |                                                                                                                                                     | 2620/? [2:46:37<00:00,  0.26it/s, v_num=5b6z]context = [[27, 55, 94], [5, 48, 83], [56, 94, 118], [121, 137, 181], [0, 61, 78], [36, 80, 108], [2, 18, 55], [8, 60, 69]]target = [[62, 80, 57], [21, 23, 34], [84, 71, 87], [139, 178, 159], [59, 19, 76], [104, 40, 51], [43, 20, 8], [47, 68, 60]]
+Epoch 0: |                                                                                                                                                     | 2629/? [2:47:11<00:00,  0.26it/s, v_num=5b6z]train step 2630; scene = [['c13076a68aeeb481'], ['0bafd80c9e2ae41b']]; loss = 0.054449
+Epoch 0: |                                                                                                                                                     | 2630/? [2:47:15<00:00,  0.26it/s, v_num=5b6z]context = [[135, 142, 145, 147, 149, 150, 158, 162, 163, 169, 172, 203], [22, 29, 55, 60, 62, 63, 64, 68, 69, 70, 71, 72]]target = [[190, 143, 159, 194, 198, 175, 197, 146, 166, 181, 174, 187], [47, 36, 59, 40, 28, 46, 65, 24, 29, 26, 69, 51]]
+Epoch 0: |                                                                                                                                                     | 2639/? [2:47:50<00:00,  0.26it/s, v_num=5b6z]train step 2640; scene = [['eae08bd3d5c892e7'], ['60acba38edb03894']]; loss = 0.034908
+Epoch 0: |                                                                                                                                                     | 2640/? [2:47:53<00:00,  0.26it/s, v_num=5b6z]context = [[166, 176, 251], [156, 173, 220], [69, 137, 148], [197, 204, 277], [28, 29, 100], [44, 111, 118], [0, 56, 61], [4, 68, 81]]target = [[170, 244, 200], [193, 173, 188], [75, 86, 116], [253, 268, 251], [49, 30, 48], [106, 54, 50], [24, 3, 29], [24, 58, 37]]
+Epoch 0: |                                                                                                                                                     | 2649/? [2:48:28<00:00,  0.26it/s, v_num=5b6z]train step 2650; scene = [['c520ae3405948a0a'], ['94340815d8708eea']]; loss = 0.032884
+Epoch 0: |                                                                                                                                                     | 2650/? [2:48:32<00:00,  0.26it/s, v_num=5b6z]context = [[6, 10, 15, 16, 20, 26, 31, 32, 37, 40, 42, 47, 55, 57, 66, 69, 70, 80, 82, 85, 88, 93, 101, 103]]target = [[72, 23, 99, 11, 66, 101, 81, 102, 30, 39, 54, 38, 17, 84, 69, 32, 22, 18, 26, 41, 79, 40, 57, 46]]
+Epoch 0: |                                                                                                                                                     | 2659/? [2:49:07<00:00,  0.26it/s, v_num=5b6z]train step 2660; scene = [['6666ae9375b1656e'], ['64619aa7bd88899d']]; loss = 0.073450
+Epoch 0: |                                                                                                                                                     | 2660/? [2:49:10<00:00,  0.26it/s, v_num=5b6z]context = [[13, 14, 27, 29, 31, 32, 39, 48, 50, 52, 54, 57, 58, 61, 67, 73, 79, 90, 93, 95, 105, 108, 109, 110]]target = [[83, 48, 82, 104, 25, 53, 89, 72, 93, 51, 41, 86, 24, 49, 32, 98, 78, 73, 88, 45, 44, 99, 58, 79]]
+Epoch 0: |                                                                                                                                                     | 2669/? [2:49:44<00:00,  0.26it/s, v_num=5b6z]train step 2670; scene = [['f0712581f6277ffc'], ['5c4442779124ec3d']]; loss = 0.049745
+Epoch 0: |                                                                                                                                                     | 2670/? [2:49:48<00:00,  0.26it/s, v_num=5b6z]context = [[0, 1, 4, 5, 6, 8, 9, 11, 13, 29, 32, 34, 35, 38, 43, 47, 50, 58, 61, 67, 86, 94, 96, 97]]target = [[38, 81, 77, 72, 58, 43, 36, 5, 7, 23, 27, 62, 13, 30, 52, 55, 19, 54, 84, 76, 49, 65, 73, 91]]
+Epoch 0: |                                                                                                                                                     | 2679/? [2:50:22<00:00,  0.26it/s, v_num=5b6z]train step 2680; scene = [['f276b95e48af6e36'], ['e5db691627ea5357']]; loss = 0.070302
+Epoch 0: |                                                                                                                                                     | 2680/? [2:50:26<00:00,  0.26it/s, v_num=5b6z]context = [[108, 120, 121, 130, 145, 149, 158, 169], [9, 17, 36, 43, 44, 69, 77, 78], [107, 120, 131, 140, 141, 142, 147, 152]]target = [[139, 144, 124, 138, 161, 143, 120, 110], [22, 74, 33, 18, 69, 50, 38, 23], [145, 132, 140, 147, 126, 119, 148, 128]]
+Epoch 0: |                                                                                                                                                     | 2689/? [2:51:01<00:00,  0.26it/s, v_num=5b6z]train step 2690; scene = [['2059d1d6c79bc51b'], ['6b1950140a598578'], ['a20b4125ede06429']]; loss = 0.097303
+Epoch 0: |                                                                                                                                                     | 2690/? [2:51:05<00:00,  0.26it/s, v_num=5b6z]context = [[3, 15, 21, 22, 26, 27, 30, 34, 45, 64, 67, 73], [6, 7, 10, 12, 24, 39, 43, 45, 47, 55, 58, 64]]target = [[49, 18, 47, 23, 50, 70, 62, 17, 42, 46, 27, 68], [59, 12, 7, 26, 50, 58, 57, 39, 23, 49, 55, 24]]
+Epoch 0: |                                                                                                                                                     | 2699/? [2:51:39<00:00,  0.26it/s, v_num=5b6z]train step 2700; scene = [['319bd9ea90f25ea3'], ['12879245713d8124'], ['9bb2a6670058b7b2']]; loss = 0.063345
+Epoch 0: |                                                                                                                                                     | 2700/? [2:51:43<00:00,  0.26it/s, v_num=5b6z]context = [[39, 40, 41, 44, 48, 49, 52, 58, 63, 67, 82, 83, 85, 87, 88, 93, 102, 104, 108, 109, 113, 118, 131, 136]]target = [[43, 106, 81, 41, 105, 135, 98, 78, 103, 44, 47, 61, 45, 52, 124, 48, 91, 59, 58, 46, 85, 109, 74, 56]]
+Epoch 0: |                                                                                                                                                     | 2709/? [2:52:15<00:00,  0.26it/s, v_num=5b6z]train step 2710; scene = [['753238098c2307cc']]; loss = 0.087537
+Epoch 0: |                                                                                                                                                     | 2710/? [2:52:19<00:00,  0.26it/s, v_num=5b6z]context = [[6, 7, 8, 9, 17, 19, 29, 32, 39, 48, 54, 55], [110, 116, 118, 121, 122, 128, 134, 138, 144, 149, 170, 178]]target = [[9, 51, 23, 19, 49, 54, 38, 7, 37, 53, 17, 16], [147, 156, 124, 177, 118, 162, 160, 143, 140, 119, 154, 133]]
+Epoch 0: |                                                                                                                                                     | 2719/? [2:52:54<00:00,  0.26it/s, v_num=5b6z]train step 2720; scene = [['7194b8d204f4a0b6'], ['5c0ddb9de8c16f05'], ['1a87a846ba692048']]; loss = 0.049094
+Epoch 0: |                                                                                                                                                     | 2720/? [2:52:57<00:00,  0.26it/s, v_num=5b6z]context = [[73, 88, 89, 91, 93, 96, 104, 111, 114, 115, 118, 120, 121, 125, 133, 136, 137, 143, 144, 145, 151, 161, 165, 170]]target = [[94, 126, 164, 115, 101, 103, 78, 120, 75, 153, 88, 86, 74, 106, 116, 141, 119, 144, 140, 165, 146, 167, 158, 139]]
+Epoch 0: |                                                                                                                                                     | 2729/? [2:53:32<00:00,  0.26it/s, v_num=5b6z]train step 2730; scene = [['b7a4f7a6d35961d4'], ['b7003ac834dc298b'], ['05596054e7569f2b'], ['b8aed6b43cd738c9']]; loss = 0.062017
+Epoch 0: |                                                                                                                                                     | 2730/? [2:53:36<00:00,  0.26it/s, v_num=5b6z]context = [[117, 186, 194, 196], [34, 38, 40, 79], [10, 13, 37, 77], [102, 103, 113, 167], [9, 42, 52, 80], [136, 174, 192, 218]]target = [[187, 180, 158, 134], [57, 76, 69, 70], [49, 37, 70, 57], [158, 143, 120, 136], [15, 34, 75, 65], [138, 214, 217, 143]]
+Epoch 0: |                                                                                                                                                     | 2739/? [2:54:09<00:00,  0.26it/s, v_num=5b6z]train step 2740; scene = [['d9641b3f6ca0b13d'], ['61ef38380e8172c7'], ['3ec462170f378b3f'], ['bc0b4df5aef0a622'], ['6bef29b74b93e80a'], ['8ef643c4e1cb9baf'], ['aa8259399a115c5f'], ['fa35453daaa8e408']]; loss = 0.065948
+Epoch 0: |                                                                                                                                                     | 2740/? [2:54:13<00:00,  0.26it/s, v_num=5b6z]context = [[0, 1, 7, 8, 9, 10, 11, 18, 22, 31, 32, 39, 49, 53, 59, 66, 69, 72, 74, 78, 81, 82, 93, 97]]target = [[74, 55, 50, 21, 47, 85, 2, 42, 41, 40, 39, 93, 49, 45, 23, 66, 73, 38, 89, 36, 87, 30, 48, 9]]
+Epoch 0: |                                                                                                                                                     | 2749/? [2:54:47<00:00,  0.26it/s, v_num=5b6z]train step 2750; scene = [['2225123ef31a93e4']]; loss = 0.107691
+Epoch 0: |                                                                                                                                                     | 2750/? [2:54:51<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 2750; scene = ['3e07add8413f8157'];
+target intrinsic:  tensor(0.8521, device='cuda:0') tensor(0.8523, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8733, device='cuda:0') tensor(0.8735, device='cuda:0')
+[2026-02-24 22:10:14,366][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2750/? [2:54:52<00:00,  0.26it/s, v_num=5b6z]context = [[12, 14, 17, 18, 20, 21, 32, 34, 38, 40, 54, 57, 65, 67, 72, 77, 82, 85, 92, 96, 98, 102, 105, 109]]target = [[101, 39, 104, 51, 15, 33, 16, 24, 94, 58, 72, 48, 20, 17, 22, 37, 107, 63, 40, 79, 69, 27, 28, 60]]
+Epoch 0: |                                                                                                                                                     | 2759/? [2:55:25<00:00,  0.26it/s, v_num=5b6z]train step 2760; scene = [['03a4631365a95cb2'], ['9f622f50132c1efe'], ['4ec29bd9dfd30c2e'], ['314d09d3997725a7'], ['c437043ac9ced6c1'], ['accfed2fa6ff849d'], ['57a1c0730778cddd'], ['bea0e295ee56d42c'], ['75e5ba7e7fb64fc0'], ['71968459b3168e4f'], ['a08c8a37e64ce67d'], ['bf27ccdada0c2373']]; loss = 0.088617
+Epoch 0: |                                                                                                                                                     | 2760/? [2:55:29<00:00,  0.26it/s, v_num=5b6z]context = [[41, 48, 51, 52, 59, 61, 68, 69, 72, 73, 75, 103], [9, 10, 23, 35, 37, 38, 41, 43, 48, 69, 76, 84]]target = [[56, 74, 53, 76, 101, 97, 93, 42, 89, 83, 91, 79], [14, 17, 28, 50, 58, 32, 54, 30, 61, 34, 83, 15]]
+Epoch 0: |                                                                                                                                                     | 2769/? [2:56:04<00:00,  0.26it/s, v_num=5b6z]train step 2770; scene = [['539e0d3713447384'], ['c49e7882e04566c0'], ['69c9010e3d4209bc']]; loss = 0.041709
+Epoch 0: |                                                                                                                                                     | 2770/? [2:56:08<00:00,  0.26it/s, v_num=5b6z]context = [[0, 9, 19, 31, 34, 68], [7, 20, 31, 36, 41, 53], [1, 15, 29, 56, 58, 74], [0, 19, 21, 35, 43, 50]]target = [[10, 2, 39, 12, 11, 26], [21, 31, 11, 28, 27, 52], [16, 13, 66, 17, 65, 6], [39, 21, 41, 27, 40, 25]]
+Epoch 0: |                                                                                                                                                     | 2779/? [2:56:42<00:00,  0.26it/s, v_num=5b6z]train step 2780; scene = [['ede896927ea91dd6'], ['7badba95b5be610e']]; loss = 0.048774
+Epoch 0: |                                                                                                                                                     | 2780/? [2:56:46<00:00,  0.26it/s, v_num=5b6z]context = [[0, 7, 26, 27, 51, 57], [5, 15, 18, 23, 55, 81], [27, 32, 44, 76, 87, 98], [8, 16, 18, 21, 24, 58]]target = [[48, 19, 22, 26, 32, 17], [57, 65, 27, 75, 51, 22], [85, 50, 79, 82, 30, 89], [41, 25, 12, 52, 32, 47]]
+Epoch 0: |                                                                                                                                                     | 2789/? [2:57:20<00:00,  0.26it/s, v_num=5b6z]train step 2790; scene = [['1e7d7ef1404597f0'], ['ea02d0f42c603c21']]; loss = 0.033389
+Epoch 0: |                                                                                                                                                     | 2790/? [2:57:24<00:00,  0.26it/s, v_num=5b6z]context = [[99, 100, 101, 107, 108, 110, 111, 113, 116, 128, 131, 139, 145, 149, 155, 162, 167, 170, 172, 179, 183, 191, 194, 196]]target = [[122, 186, 110, 138, 149, 133, 161, 112, 141, 184, 189, 178, 100, 131, 188, 155, 174, 195, 185, 171, 137, 160, 176, 170]]
+Epoch 0: |                                                                                                                                                     | 2799/? [2:57:59<00:00,  0.26it/s, v_num=5b6z]train step 2800; scene = [['c8e92789f25baec1'], ['b477406d6064f1a3']]; loss = 0.033460
+Epoch 0: |                                                                                                                                                     | 2800/? [2:58:03<00:00,  0.26it/s, v_num=5b6z]context = [[2, 4, 5, 18, 22, 24, 31, 37, 39, 45, 54, 57, 60, 64, 67, 73, 82, 83, 84, 86, 87, 97, 98, 99]]target = [[23, 55, 39, 36, 80, 52, 28, 29, 89, 63, 56, 61, 98, 78, 40, 9, 77, 82, 69, 47, 43, 68, 81, 44]]
+[2026-02-24 22:13:29,419][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2809/? [2:58:41<00:00,  0.26it/s, v_num=5b6z]train step 2810; scene = [['31907a6ebec6ef2e']]; loss = 0.033759
+Epoch 0: |                                                                                                                                                     | 2810/? [2:58:45<00:00,  0.26it/s, v_num=5b6z]context = [[96, 97, 145, 156, 165, 173], [147, 150, 159, 169, 173, 196], [21, 33, 34, 46, 56, 75], [7, 17, 47, 50, 51, 58]]target = [[144, 115, 146, 171, 129, 101], [194, 177, 170, 186, 155, 193], [34, 60, 52, 63, 32, 22], [37, 45, 50, 22, 27, 35]]
+Epoch 0: |                                                                                                                                                     | 2819/? [2:59:20<00:00,  0.26it/s, v_num=5b6z]train step 2820; scene = [['b60f8abb905bb1e1']]; loss = 0.067870
+Epoch 0: |                                                                                                                                                     | 2820/? [2:59:24<00:00,  0.26it/s, v_num=5b6z]context = [[11, 22, 72], [43, 82, 129], [0, 56, 58], [40, 93, 105], [0, 28, 57], [3, 75, 76], [51, 109, 110], [26, 85, 112]]target = [[21, 31, 45], [57, 82, 123], [45, 30, 31], [94, 95, 98], [27, 53, 38], [43, 50, 49], [102, 80, 69], [106, 107, 98]]
+Epoch 0: |                                                                                                                                                     | 2829/? [2:59:59<00:00,  0.26it/s, v_num=5b6z]train step 2830; scene = [['ad1ad6648a30592c'], ['448e12a47b369490'], ['d3dddf816450d1d0'], ['d7fd0247a853f44b'], ['27cf7ba27eed1eae'], ['fd73259e7c2801e8'], ['9869d33078eac049'], ['06779b91ed31eaf1']]; loss = 0.067531
+Epoch 0: |                                                                                                                                                     | 2830/? [3:00:03<00:00,  0.26it/s, v_num=5b6z]context = [[15, 17, 19, 20, 21, 26, 36, 40, 42, 44, 47, 49, 50, 57, 67, 77, 78, 87, 89, 90, 102, 104, 105, 112]]target = [[103, 47, 82, 105, 98, 89, 52, 35, 33, 55, 65, 60, 107, 87, 38, 34, 20, 40, 81, 102, 88, 53, 93, 16]]
+Epoch 0: |                                                                                                                                                     | 2839/? [3:00:37<00:00,  0.26it/s, v_num=5b6z]train step 2840; scene = [['a911c16d8056a34d']]; loss = 0.032519
+Epoch 0: |                                                                                                                                                     | 2840/? [3:00:41<00:00,  0.26it/s, v_num=5b6z]context = [[12, 15, 17, 19, 47, 48, 51, 58], [32, 38, 40, 56, 57, 72, 75, 78], [75, 90, 96, 97, 109, 115, 129, 139]]target = [[53, 44, 45, 42, 50, 17, 36, 54], [71, 72, 37, 48, 54, 39, 74, 52], [104, 134, 124, 132, 129, 127, 137, 121]]
+Epoch 0: |                                                                                                                                                     | 2849/? [3:01:15<00:00,  0.26it/s, v_num=5b6z]train step 2850; scene = [['9753a8291766b5da'], ['3a8f14b855f7f4ee'], ['ba3f2372b7959e95']]; loss = 0.047547
+Epoch 0: |                                                                                                                                                     | 2850/? [3:01:19<00:00,  0.26it/s, v_num=5b6z]context = [[134, 142, 143, 146, 150, 151, 154, 155, 177, 179, 184, 208], [70, 78, 83, 88, 97, 106, 108, 112, 115, 126, 133, 139]]target = [[160, 140, 154, 139, 135, 166, 141, 167, 155, 136, 164, 199], [72, 104, 100, 117, 114, 103, 112, 123, 133, 115, 110, 81]]
+Epoch 0: |                                                                                                                                                     | 2859/? [3:01:53<00:00,  0.26it/s, v_num=5b6z]train step 2860; scene = [['8f4f366720645ea0'], ['fef48b769b17f0ed']]; loss = 0.032263
+Epoch 0: |                                                                                                                                                     | 2860/? [3:01:56<00:00,  0.26it/s, v_num=5b6z]context = [[115, 116, 123, 135, 137, 144, 146, 152, 155, 161, 162, 168, 176, 177, 179, 184, 187, 190, 198, 199, 200, 202, 209, 212]]target = [[195, 155, 181, 142, 154, 202, 144, 161, 136, 132, 137, 192, 148, 189, 188, 174, 178, 133, 156, 134, 124, 197, 199, 166]]
+Epoch 0: |                                                                                                                                                     | 2869/? [3:02:31<00:00,  0.26it/s, v_num=5b6z]train step 2870; scene = [['dff9f50bbc0d0d5d'], ['3baf335bfb54b9c0'], ['c74adb4381e65e99']]; loss = 0.054328
+Epoch 0: |                                                                                                                                                     | 2870/? [3:02:35<00:00,  0.26it/s, v_num=5b6z]context = [[30, 41, 45, 47, 51, 57, 66, 68, 77, 83, 85, 90, 95, 96, 103, 105, 110, 111, 116, 120, 122, 123, 126, 127]]target = [[118, 32, 75, 92, 109, 89, 43, 36, 72, 115, 96, 111, 76, 68, 90, 70, 69, 66, 38, 77, 81, 45, 34, 42]]
+Epoch 0: |                                                                                                                                                     | 2879/? [3:03:10<00:00,  0.26it/s, v_num=5b6z]train step 2880; scene = [['d6bd4b0784843fdd'], ['bd1bc8c20a660a96'], ['f8f0491a2268ee5e'], ['90bcc3b4b3d551a8'], ['57ccd86b6a8979b3'], ['9d66dba4551f79f8'], ['a30f898a6e455745'], ['73973f88f24769be']]; loss = 0.054677
+Epoch 0: |                                                                                                                                                     | 2880/? [3:03:13<00:00,  0.26it/s, v_num=5b6z]context = [[2, 4, 5, 6, 10, 14, 18, 21, 29, 36, 41, 54], [21, 24, 33, 45, 50, 54, 58, 64, 75, 78, 87, 97]]target = [[11, 35, 12, 39, 53, 18, 48, 26, 24, 29, 9, 50], [42, 27, 84, 60, 87, 49, 94, 52, 68, 54, 44, 76]]
+Epoch 0: |                                                                                                                                                     | 2889/? [3:03:48<00:00,  0.26it/s, v_num=5b6z]train step 2890; scene = [['294bfa0dd8a9eada'], ['a069f40a4a017b66']]; loss = 0.058696
+Epoch 0: |                                                                                                                                                     | 2890/? [3:03:52<00:00,  0.26it/s, v_num=5b6z]context = [[3, 8, 10, 16, 25, 28, 29, 35, 36, 37, 43, 55, 60, 69, 72, 75, 82, 84, 87, 90, 91, 94, 95, 100]]target = [[34, 96, 98, 26, 86, 19, 46, 11, 4, 54, 8, 29, 87, 56, 9, 25, 82, 49, 6, 37, 32, 43, 21, 38]]
+Epoch 0: |                                                                                                                                                     | 2899/? [3:04:25<00:00,  0.26it/s, v_num=5b6z]train step 2900; scene = [['cf6618aadac4ddd9'], ['f7e27052900e847e'], ['0b173d0c5951a7f5'], ['3551ff5b8a497fb7']]; loss = 0.038795
+Epoch 0: |                                                                                                                                                     | 2900/? [3:04:29<00:00,  0.26it/s, v_num=5b6z]context = [[38, 95, 96], [42, 65, 102], [13, 22, 77], [198, 232, 261], [13, 56, 82], [7, 21, 79], [0, 49, 80], [83, 129, 157]]target = [[85, 41, 90], [74, 61, 87], [54, 41, 58], [242, 233, 231], [20, 63, 48], [10, 63, 69], [44, 7, 71], [89, 154, 86]]
+Epoch 0: |                                                                                                                                                     | 2909/? [3:05:04<00:00,  0.26it/s, v_num=5b6z]train step 2910; scene = [['b5924605972475e2'], ['64554b0854be0a81']]; loss = 0.033421
+Epoch 0: |                                                                                                                                                     | 2910/? [3:05:08<00:00,  0.26it/s, v_num=5b6z]context = [[0, 15, 38, 47], [103, 127, 147, 175], [174, 186, 243, 250], [14, 30, 57, 62], [88, 115, 122, 134], [64, 65, 75, 127]]target = [[11, 38, 21, 27], [165, 125, 140, 149], [233, 239, 215, 225], [41, 57, 27, 47], [94, 121, 104, 109], [65, 69, 83, 107]]
+Epoch 0: |                                                                                                                                                     | 2919/? [3:05:42<00:00,  0.26it/s, v_num=5b6z]train step 2920; scene = [['d53f3d87d749e474']]; loss = 0.043357
+Epoch 0: |                                                                                                                                                     | 2920/? [3:05:46<00:00,  0.26it/s, v_num=5b6z]context = [[94, 95, 106, 107, 112, 114, 118, 121, 124, 126, 128, 132, 137, 138, 147, 148, 153, 156, 173, 175, 176, 184, 188, 191]]target = [[164, 111, 98, 129, 114, 99, 141, 146, 144, 153, 96, 149, 109, 122, 151, 187, 136, 167, 132, 176, 95, 148, 126, 181]]
+Epoch 0: |                                                                                                                                                     | 2929/? [3:06:19<00:00,  0.26it/s, v_num=5b6z]train step 2930; scene = [['e61532beb3fa8b63'], ['c0803f4d1cb0eeea']]; loss = 0.030850
+Epoch 0: |                                                                                                                                                     | 2930/? [3:06:23<00:00,  0.26it/s, v_num=5b6z]context = [[16, 17, 22, 25, 31, 34, 37, 38, 41, 46, 54, 55, 57, 66, 75, 82, 84, 85, 92, 101, 104, 107, 111, 113]]target = [[109, 112, 73, 65, 111, 102, 34, 28, 20, 49, 46, 85, 17, 81, 50, 58, 22, 89, 52, 19, 63, 87, 59, 25]]
+Epoch 0: |                                                                                                                                                     | 2939/? [3:06:56<00:00,  0.26it/s, v_num=5b6z]train step 2940; scene = [['73aee8654106974f']]; loss = 0.036441
+Epoch 0: |                                                                                                                                                     | 2940/? [3:07:00<00:00,  0.26it/s, v_num=5b6z]context = [[80, 83, 85, 91, 93, 101, 104, 109, 110, 118, 125, 131, 133, 138, 140, 142, 147, 160, 163, 164, 167, 168, 173, 177]]target = [[172, 125, 159, 124, 146, 83, 138, 151, 164, 144, 84, 121, 153, 91, 105, 107, 109, 163, 131, 114, 89, 94, 98, 108]]
+Epoch 0: |                                                                                                                                                     | 2949/? [3:07:35<00:00,  0.26it/s, v_num=5b6z]train step 2950; scene = [['79d5ea6cd3f0fdb2'], ['85fb2f64303a1388'], ['9e3c241e9f50165d'], ['6a94bfa75e7988c8'], ['270022f0b06e71d5'], ['02f1d3b1d43877df'], ['6f4cc17690dcdd2e'], ['70c5a81e8b7868cc'], ['63982f095d5089b0'], ['f7498ea452fed198'], ['d07aa4d1691ccf58'], ['5474e6cd7ccd6d1a']]; loss = 0.069565
+Epoch 0: |                                                                                                                                                     | 2950/? [3:07:38<00:00,  0.26it/s, v_num=5b6z]context = [[1, 20, 21, 24, 25, 26, 27, 30, 36, 39, 50, 53], [60, 74, 75, 77, 82, 90, 99, 106, 111, 113, 114, 133]]target = [[12, 37, 4, 3, 45, 18, 13, 11, 51, 27, 44, 24], [74, 66, 127, 95, 119, 112, 130, 78, 75, 77, 114, 97]]
+Epoch 0: |                                                                                                                                                     | 2959/? [3:08:13<00:00,  0.26it/s, v_num=5b6z]train step 2960; scene = [['0e5f43adc84b0435']]; loss = 0.076283
+Epoch 0: |                                                                                                                                                     | 2960/? [3:08:16<00:00,  0.26it/s, v_num=5b6z]context = [[73, 79, 98, 99, 100, 103, 107, 126], [40, 42, 46, 63, 67, 92, 93, 118], [3, 7, 12, 36, 38, 39, 41, 59]]target = [[113, 108, 119, 82, 81, 84, 86, 124], [90, 106, 43, 103, 69, 87, 113, 111], [18, 41, 24, 30, 7, 20, 53, 54]]
+Epoch 0: |                                                                                                                                                     | 2969/? [3:08:51<00:00,  0.26it/s, v_num=5b6z]train step 2970; scene = [['d3784c9108c25d42']]; loss = 0.033211
+Epoch 0: |                                                                                                                                                     | 2970/? [3:08:55<00:00,  0.26it/s, v_num=5b6z]context = [[10, 11, 17, 29, 30, 31, 33, 39, 48, 60, 63, 64, 69, 70, 72, 87, 91, 92, 93, 95, 99, 100, 104, 107]]target = [[63, 97, 21, 50, 94, 32, 27, 70, 60, 78, 65, 45, 53, 85, 56, 33, 38, 80, 98, 12, 44, 95, 43, 34]]
+Epoch 0: |                                                                                                                                                     | 2979/? [3:09:29<00:00,  0.26it/s, v_num=5b6z]train step 2980; scene = [['058bac6c226fc1a9'], ['7d900c809e896e32']]; loss = 0.038093
+Epoch 0: |                                                                                                                                                     | 2980/? [3:09:33<00:00,  0.26it/s, v_num=5b6z]context = [[64, 69, 70, 71, 83, 85, 86, 93, 94, 96, 111, 112, 113, 117, 123, 138, 139, 140, 141, 148, 149, 153, 159, 161]]target = [[97, 127, 104, 150, 132, 129, 78, 128, 116, 125, 134, 130, 149, 99, 137, 79, 88, 140, 94, 98, 142, 153, 155, 73]]
+Epoch 0: |                                                                                                                                                     | 2989/? [3:10:07<00:00,  0.26it/s, v_num=5b6z]train step 2990; scene = [['51f581faf9000425'], ['89bc971b5dfbd294'], ['511207c07d553599']]; loss = 0.042557
+Epoch 0: |                                                                                                                                                     | 2990/? [3:10:10<00:00,  0.26it/s, v_num=5b6z]context = [[27, 42, 44, 84, 95, 99], [168, 178, 219, 221, 222, 224], [50, 56, 89, 100, 116, 131], [117, 131, 156, 157, 160, 171]]target = [[41, 88, 49, 83, 97, 46], [194, 197, 187, 203, 199, 177], [130, 88, 129, 93, 52, 103], [130, 162, 157, 127, 156, 133]]
+Epoch 0: |                                                                                                                                                     | 2999/? [3:10:43<00:00,  0.26it/s, v_num=5b6z]train step 3000; scene = [['d77284dc9b9d1031']]; loss = 0.135510
+Epoch 0: |                                                                                                                                                     | 3000/? [3:10:47<00:00,  0.26it/s, v_num=5b6z]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 3000; scene = ['a76028640ffa1ef9'];
+target intrinsic:  tensor(0.8569, device='cuda:0') tensor(0.8571, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8813, device='cuda:0') tensor(0.8807, device='cuda:0')
+[2026-02-24 22:26:21,789][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 3000/? [3:11:00<00:00,  0.26it/s, v_num=5b6z]context = [[176, 195, 196, 203, 214, 228, 231, 237], [75, 81, 89, 91, 113, 117, 152, 155], [54, 61, 76, 79, 90, 95, 104, 105]]target = [[207, 201, 216, 186, 203, 231, 198, 179], [144, 150, 79, 87, 147, 123, 115, 138], [86, 82, 99, 72, 100, 74, 66, 62]]
+[2026-02-24 22:26:25,372][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 3001/? [3:11:04<00:00,  0.26it/s, v_num=5b6z]
+`Trainer.fit` stopped: `max_steps=3001` reached.
+Peak VRAM: 103.417 GB (allocated), 137.797 GB (reserved)
+Total elapsed: 3.20 hours
+Saved memory info to: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_OURS/peak_vram_memory.json
diff --git a/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/requirements.txt b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..7fbf9096f92b53f8bb2a7e5467c79ecbe64faca5
--- /dev/null
+++ b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/requirements.txt
@@ -0,0 +1,172 @@
+wheel==0.45.1
+pytz==2025.2
+easydict==1.13
+antlr4-python3-runtime==4.9.3
+wadler_lindig==0.1.7
+urllib3==2.5.0
+tzdata==2025.2
+typing-inspection==0.4.1
+tabulate==0.9.0
+smmap==5.0.2
+kornia_rs==0.1.9
+setuptools==78.1.1
+safetensors==0.5.3
+PyYAML==6.0.2
+PySocks==1.7.1
+pyparsing==3.2.5
+pydantic_core==2.33.2
+pycparser==2.23
+protobuf==6.32.1
+propcache==0.3.2
+proglog==0.1.12
+fsspec==2024.6.1
+platformdirs==4.4.0
+pip==25.2
+pillow==10.4.0
+frozenlist==1.7.0
+packaging==24.2
+opt_einsum==3.4.0
+numpy==1.26.4
+ninja==1.13.0
+fonttools==4.60.0
+networkx==3.4.2
+multidict==6.6.4
+mdurl==0.1.2
+MarkupSafe==3.0.2
+kiwisolver==1.4.9
+imageio-ffmpeg==0.6.0
+idna==3.7
+hf-xet==1.1.10
+gmpy2==2.2.1
+einops==0.8.1
+filelock==3.17.0
+decorator==4.4.2
+dacite==1.9.2
+cycler==0.12.1
+colorama==0.4.6
+click==8.3.0
+nvidia-nvtx-cu12==12.8.90
+charset-normalizer==3.3.2
+certifi==2025.8.3
+beartype==0.19.0
+attrs==25.3.0
+async-timeout==5.0.1
+annotated-types==0.7.0
+aiohappyeyeballs==2.6.1
+yarl==1.20.1
+tifffile==2025.5.10
+sentry-sdk==2.39.0
+scipy==1.15.3
+pydantic==2.11.9
+pandas==2.3.2
+opencv-python==4.11.0.86
+omegaconf==2.3.0
+markdown-it-py==4.0.0
+lightning-utilities==0.14.3
+lazy_loader==0.4
+jaxtyping==0.2.37
+imageio==2.37.0
+gitdb==4.0.12
+contourpy==1.3.2
+colorspacious==1.1.2
+cffi==1.17.1
+aiosignal==1.4.0
+scikit-video==1.1.11
+scikit-image==0.25.2
+rich==14.1.0
+moviepy==1.0.3
+matplotlib==3.10.6
+hydra-core==1.3.2
+nvidia-nccl-cu12==2.27.3
+huggingface-hub==0.35.1
+GitPython==3.1.45
+brotlicffi==1.0.9.2
+aiohttp==3.12.15
+torchmetrics==1.8.2
+opt-einsum-fx==0.1.4
+kornia==0.8.1
+pytorch-lightning==2.5.1
+lpips==0.1.4
+e3nn==0.6.0
+lightning==2.5.1
+nvidia-cusparselt-cu12==0.7.1
+triton==3.4.0
+nvidia-nvjitlink-cu12==12.8.93
+nvidia-curand-cu12==10.3.9.90
+nvidia-cufile-cu12==1.13.1.3
+nvidia-cuda-runtime-cu12==12.8.90
+nvidia-cuda-nvrtc-cu12==12.8.93
+nvidia-cuda-cupti-cu12==12.8.90
+nvidia-cublas-cu12==12.8.4.1
+nvidia-cusparse-cu12==12.5.8.93
+nvidia-cufft-cu12==11.3.3.83
+nvidia-cudnn-cu12==9.10.2.21
+nvidia-cusolver-cu12==11.7.3.90
+torch==2.8.0+cu128
+torchvision==0.23.0+cu128
+torchaudio==2.8.0+cu128
+torch_scatter==2.1.2+pt28cu128
+gsplat==1.5.3
+wandb==0.25.0
+cuda-bindings==13.0.3
+cuda-pathfinder==1.3.3
+Jinja2==3.1.6
+mpmath==1.3.0
+nvidia-cublas==13.1.0.3
+nvidia-cuda-cupti==13.0.85
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-runtime==13.0.96
+nvidia-cudnn-cu13==9.15.1.9
+nvidia-cufft==12.0.0.61
+nvidia-cufile==1.15.1.6
+nvidia-curand==10.4.0.35
+nvidia-cusolver==12.0.4.66
+nvidia-cusparse==12.6.3.3
+nvidia-cusparselt-cu13==0.8.0
+nvidia-nccl-cu13==2.28.9
+nvidia-nvjitlink==13.0.88
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvtx==13.0.85
+requests==2.32.5
+sentencepiece==0.2.1
+sympy==1.14.0
+torchcodec==0.10.0
+torchdata==0.10.0
+torchtext==0.6.0
+anyio==4.12.0
+asttokens==3.0.1
+comm==0.2.3
+debugpy==1.8.19
+executing==2.2.1
+h11==0.16.0
+httpcore==1.0.9
+httpx==0.28.1
+ipykernel==7.1.0
+ipython==9.8.0
+ipython_pygments_lexers==1.1.1
+ipywidgets==8.1.8
+jedi==0.19.2
+jupyter_client==8.7.0
+jupyter_core==5.9.1
+jupyterlab_widgets==3.0.16
+matplotlib-inline==0.2.1
+nest-asyncio==1.6.0
+parso==0.8.5
+pexpect==4.9.0
+prompt_toolkit==3.0.52
+psutil==7.2.1
+ptyprocess==0.7.0
+pure_eval==0.2.3
+Pygments==2.19.2
+python-dateutil==2.9.0.post0
+pyzmq==27.1.0
+shellingham==1.5.4
+six==1.17.0
+stack-data==0.6.3
+tornado==6.5.4
+tqdm==4.67.1
+traitlets==5.14.3
+typer-slim==0.21.0
+typing_extensions==4.15.0
+wcwidth==0.2.14
+widgetsnbextension==4.0.15
diff --git a/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/wandb-metadata.json b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/wandb-metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..260839682fea099b54c274530ee72a7ebd36b41d
--- /dev/null
+++ b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/wandb-metadata.json
@@ -0,0 +1,92 @@
+{
+  "os":  "Linux-6.8.0-90-generic-x86_64-with-glibc2.39",
+  "python":  "CPython 3.12.12",
+  "startedAt":  "2026-02-24T19:15:08.304921Z",
+  "args":  [
+    "+experiment=re10k_ablation_24v",
+    "wandb.mode=online",
+    "wandb.name=ABLATION_0225_OURS"
+  ],
+  "program":  "-m src.main",
+  "git":  {
+    "remote":  "git@github.com:K-nowing/CVPR2026.git",
+    "commit":  "2512754c6c27ca5150bf17fbcbdde3f192fd53cc"
+  },
+  "email":  "dna9041@korea.ac.kr",
+  "root":  "/workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_OURS",
+  "host":  "27d18dedec6d",
+  "executable":  "/venv/main/bin/python",
+  "cpu_count":  128,
+  "cpu_count_logical":  256,
+  "gpu":  "NVIDIA H200",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "1170378588160",
+      "used":  "612674392064"
+    }
+  },
+  "memory":  {
+    "total":  "1622948257792"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-2649ab80-a3a6-5a1c-0fa5-12bc11bd75e9"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-e92921d9-c681-246f-af93-637e0dc938ca"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-ffe12ffc-9bb7-82de-5692-1ec0ee2e68d8"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-499e5acd-b6ab-2010-c51b-ee9b5aa65825"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-3b2522d9-1c72-e49b-2c30-96165680b74a"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-a9a280c5-b2f9-dc1e-a8a9-7326a74001ff"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-07d0167b-a6a1-1900-2d27-7c6c11598409"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-8362a999-20d1-c27b-5d18-032d23f859ab"
+    }
+  ],
+  "cudaVersion":  "13.1",
+  "writerId":  "lma14qrq4ffkxha58hrfyhtyvrmlfx2i"
+}
\ No newline at end of file
diff --git a/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/wandb-summary.json b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/wandb-summary.json
new file mode 100644
index 0000000000000000000000000000000000000000..73c8caa8ae9fe5dd4543ddf07eb2e86eeecaae23
--- /dev/null
+++ b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/files/wandb-summary.json
@@ -0,0 +1 @@
+{"val/psnr":21.833919525146484,"val/gaussian_num_ratio":0.3997955322265625,"lr-AdamW/pg2-momentum":0.9,"loss/aux_1/mse":0.023314595222473145,"lr-AdamW/pg1-momentum":0.9,"_runtime":11484,"val/lpips":0.15058016777038574,"loss/aux_2/lpips":0.010988885536789894,"loss/aux_2/mse":0.021367380395531654,"active_mask_imgs":{"format":"png","count":1,"filenames":["media/images/active_mask_imgs_198_c40793305ed32fbebf33.png"],"captions":["a76028640ffa1ef9"],"_type":"images/separated","width":536,"height":800},"_wandb":{"runtime":11484},"loss/aux_1/error_score":0.26384127140045166,"lr-AdamW/pg1":2.003594834351718e-05,"lr-AdamW/pg2":2e-05,"val/ssim":0.8224983215332031,"epoch":0,"comparison":{"format":"png","count":1,"filenames":["media/images/comparison_197_7a08eead29b131fa3472.png"],"captions":["a76028640ffa1ef9"],"_type":"images/separated","width":1064,"height":1098},"loss/aux_0/lpips":0.011281725019216537,"loss/aux_0/error_score":0.38854989409446716,"train/scene_scale":1.007591724395752,"_timestamp":1.7719719874732008e+09,"loss/final_3dgs/mse":0.017216090112924576,"train/psnr_probabilistic":18.861385345458984,"train/comparison":{"captions":[["0621c7675fab1418"]],"_type":"images/separated","width":1328,"height":2154,"format":"png","count":1,"filenames":["media/images/train/comparison_202_cd1b6f4b037275c862f9.png"]},"loss/final_3dgs/lpips":0.010056810453534126,"trainer/global_step":3001,"_step":202,"loss/camera":0.0006347345770336688,"train/error_scores":{"count":1,"filenames":["media/images/train/error_scores_201_ff4206b6e1e67b9747a0.png"],"captions":[["0621c7675fab1418"]],"_type":"images/separated","width":1328,"height":2120,"format":"png"},"loss/total":0.13550999760627747,"info/global_step":3000,"loss/aux_0/mse":0.01684199832379818,"loss/scene_scale_reg":0.00027991057140752673,"loss/aux_1/lpips":0.011291351169347763,"error_scores":{"height":536,"format":"png","count":1,"filenames":["media/images/error_scores_199_8210df2265e5ec17f86c.png"],"captions":["a76028640ffa1ef9"],"_type":"images/separated","width":800}}
\ No newline at end of file
diff --git a/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug-core.log b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug-core.log
new file mode 100644
index 0000000000000000000000000000000000000000..1bb3851a3da538c902bc1bcbaf593458a1b1bf8e
--- /dev/null
+++ b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug-core.log
@@ -0,0 +1,15 @@
+{"time":"2026-02-24T19:15:08.401067312Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpolh0ef_f/port-90349.txt","pid":90349,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-02-24T19:15:08.401700012Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":90349}
+{"time":"2026-02-24T19:15:08.401675802Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-90349-93084-214034883/socket","Net":"unix"}}
+{"time":"2026-02-24T19:15:08.582167376Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-02-24T19:15:08.591392017Z","level":"INFO","msg":"handleInformInit: received","streamId":"0b125b6z","id":"1(@)"}
+{"time":"2026-02-24T19:15:09.222880558Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"0b125b6z","id":"1(@)"}
+{"time":"2026-02-24T19:15:15.586457954Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"oxc0a3k6ggl8"}
+{"time":"2026-02-24T22:26:34.518248714Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
+{"time":"2026-02-24T22:26:34.518338446Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
+{"time":"2026-02-24T22:26:34.518373096Z","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-02-24T22:26:34.518414607Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
+{"time":"2026-02-24T22:26:34.519469613Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-90349-93084-214034883/socket","Net":"unix"}}
+{"time":"2026-02-24T22:26:35.605950781Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
+{"time":"2026-02-24T22:26:35.605994712Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
+{"time":"2026-02-24T22:26:35.606019742Z","level":"INFO","msg":"server is closed"}
diff --git a/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug-internal.log b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..14163630e92832b08bd4b3f84b08b6cb465a87f7
--- /dev/null
+++ b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug-internal.log
@@ -0,0 +1,11 @@
+{"time":"2026-02-24T19:15:08.591653472Z","level":"INFO","msg":"stream: starting","core version":"0.25.0"}
+{"time":"2026-02-24T19:15:09.22244861Z","level":"INFO","msg":"stream: created new stream","id":"0b125b6z"}
+{"time":"2026-02-24T19:15:09.222653934Z","level":"INFO","msg":"handler: started","stream_id":"0b125b6z"}
+{"time":"2026-02-24T19:15:09.222865877Z","level":"INFO","msg":"stream: started","id":"0b125b6z"}
+{"time":"2026-02-24T19:15:09.222943579Z","level":"INFO","msg":"writer: started","stream_id":"0b125b6z"}
+{"time":"2026-02-24T19:15:09.222946409Z","level":"INFO","msg":"sender: started","stream_id":"0b125b6z"}
+{"time":"2026-02-24T22:26:34.518352356Z","level":"INFO","msg":"stream: closing","id":"0b125b6z"}
+{"time":"2026-02-24T22:26:35.362766174Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-02-24T22:26:35.604459738Z","level":"INFO","msg":"handler: closed","stream_id":"0b125b6z"}
+{"time":"2026-02-24T22:26:35.604786383Z","level":"INFO","msg":"sender: closed","stream_id":"0b125b6z"}
+{"time":"2026-02-24T22:26:35.604815153Z","level":"INFO","msg":"stream: closed","id":"0b125b6z"}
diff --git a/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug.log b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..d6cac35968a95acc41624d3c77cf62af7e1e3185
--- /dev/null
+++ b/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug.log
@@ -0,0 +1,21 @@
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_setup.py:_flush():81] Current SDK version is 0.25.0
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_setup.py:_flush():81] Configure stats pid to 90349
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug.log
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_OURS/wandb/run-20260224_191508-0b125b6z/logs/debug-internal.log
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_init.py:init():844] calling init triggers
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
+config: {'model': {'encoder': {'name': 'dcsplat', 'input_image_shape': [518, 518], 'head_mode': 'pcd', 'num_level': 3, 'gs_param_dim': 256, 'align_corners': False, 'use_voxelize': True}, 'decoder': {'name': 'splatting_cuda', 'background_color': [0.0, 0.0, 0.0], 'make_scale_invariant': False}, 'density_control': {'name': 'density_control_module', 'mean_dim': 32, 'gs_param_dim': 256, 'refinement_layer_num': 1, 'num_level': 3, 'grad_mode': 'absgrad', 'use_mean_features': True, 'refinement_type': 'voxelize', 'refinement_hidden_dim': 32, 'aggregation_mode': 'mean', 'num_heads': 1, 'score_mode': 'absgrad', 'latent_dim': 128, 'num_latents': 64, 'num_self_attn_per_block': 2, 'voxel_size': 0.001, 'aux_refine': False, 'refine_error': False, 'use_refine_module': True, 'voxelize_activate': True, 'use_depth': False}}, 'render_loss': {'mse': {'weight': 1.0}, 'lpips': {'weight': 0.05, 'apply_after_step': 0}}, 'density_control_loss': {'error_score': {'weight': 0.01, 'log_scale': False, 'grad_scale': 10000.0, 'mode': 'original'}}, 'direct_loss': {'l1': {'weight': 0.8}, 'ssim': {'weight': 0.2}}, 'wandb': {'project': 'DCSplat', 'entity': 'scene-representation-group', 'name': 'ABLATION_0225_OURS', 'mode': 'online', 'tags': ['re10k', '256x256']}, 'mode': 'train', 'data_loader': {'train': {'num_workers': 16, 'persistent_workers': True, 'batch_size': 16, 'seed': 1234}, 'test': {'num_workers': 4, 'persistent_workers': False, 'batch_size': 1, 'seed': 2345}, 'val': {'num_workers': 1, 'persistent_workers': True, 'batch_size': 1, 'seed': 3456}}, 'optimizer': {'lr': 0.0002, 'warm_up_steps': 25, 'backbone_lr_multiplier': 0.1, 'backbone_trainable': 'T+H', 'accumulate': 1}, 'checkpointing': {'load': None, 'every_n_train_steps': 1500, 'save_top_k': 2, 'save_weights_only': False}, 'train': {'extended_visualization': False, 'print_log_every_n_steps': 10, 'camera_loss': 10.0, 'one_sample_validation': None, 'align_corners': False, 'intrinsic_scaling': False, 'verbose': False, 'beta_dist_param': [0.5, 4.0], 'use_refine_aux': False, 'train_target_set': True, 'train_gs_num': 1, 'ext_scale_detach': False, 'cam_scale_mode': 'sum', 'scene_scale_reg_loss': 0.01, 'train_aux': True, 'vggt_cam_loss': True, 'vggt_distil': False, 'context_view_train': False}, 'test': {'output_path': 'test/ablation/re10k', 'align_pose': False, 'pose_align_steps': 100, 'rot_opt_lr': 0.005, 'trans_opt_lr': 0.005, 'compute_scores': True, 'save_image': False, 'save_video': False, 'save_active_mask_image': False, 'save_error_score_image': False, 'save_compare': False, 'pred_intrinsic': False, 'error_threshold': 0.4, 'error_threshold_list': [0.2, 0.4, 0.6, 0.8, 1.0], 'threshold_mode': 'ratio', 'nvs_view_N_list': [3, 6, 16, 32, 64]}, 'seed': 111123, 'trainer': {'max_steps': 3001, 'val_check_interval': 250, 'gradient_clip_val': 0.5, 'num_nodes': 1}, 'dataset': {'re10k': {'make_baseline_1': True, 'relative_pose': True, 'augment': True, 'background_color': [0.0, 0.0, 0.0], 'overfit_to_scene': None, 'skip_bad_shape': True, 'view_sampler': {'name': 'bounded', 'num_target_views': 4, 'num_context_views': 2, 'min_distance_between_context_views': 45, 'max_distance_between_context_views': 90, 'min_distance_to_context_views': 0, 'warm_up_steps': 1000, 'initial_min_distance_between_context_views': 25, 'initial_max_distance_between_context_views': 25, 'same_target_gap': False, 'num_target_set': 3}, 'name': 're10k', 'roots': ['datasets/re10k'], 'input_image_shape': [256, 256], 'original_image_shape': [360, 640], 'cameras_are_circular': False, 'baseline_min': 0.001, 'baseline_max': 10000000000.0, 'max_fov': 100.0, 'dynamic_context_views': True, 'max_context_views_per_gpu': 24}}, '_wandb': {}}
+2026-02-24 19:15:08,307 INFO    MainThread:90349 [wandb_init.py:init():892] starting backend
+2026-02-24 19:15:08,582 INFO    MainThread:90349 [wandb_init.py:init():895] sending inform_init request
+2026-02-24 19:15:08,588 INFO    MainThread:90349 [wandb_init.py:init():903] backend started and connected
+2026-02-24 19:15:08,591 INFO    MainThread:90349 [wandb_init.py:init():973] updated telemetry
+2026-02-24 19:15:08,598 INFO    MainThread:90349 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
+2026-02-24 19:15:10,455 INFO    MainThread:90349 [wandb_init.py:init():1042] starting run threads in backend
+2026-02-24 19:15:10,580 INFO    MainThread:90349 [wandb_run.py:_console_start():2524] atexit reg
+2026-02-24 19:15:10,580 INFO    MainThread:90349 [wandb_run.py:_redirect():2373] redirect: wrap_raw
+2026-02-24 19:15:10,580 INFO    MainThread:90349 [wandb_run.py:_redirect():2442] Wrapping output streams.
+2026-02-24 19:15:10,582 INFO    MainThread:90349 [wandb_run.py:_redirect():2465] Redirects installed.
+2026-02-24 19:15:10,584 INFO    MainThread:90349 [wandb_init.py:init():1082] run started, returning control to user process
+2026-02-24 22:26:34,518 INFO    wandb-AsyncioManager-main:90349 [service_client.py:_forward_responses():134] Reached EOF.
+2026-02-24 22:26:34,518 INFO    wandb-AsyncioManager-main:90349 [mailbox.py:close():155] Closing mailbox, abandoning 1 handles.
diff --git a/ABLATION_0225_noRefineModule/.hydra/config.yaml b/ABLATION_0225_noRefineModule/.hydra/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..d40ea32c3bb82917c930d7f18220af5caee4d1d5
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/.hydra/config.yaml
@@ -0,0 +1,185 @@
+model:
+  encoder:
+    name: dcsplat
+    input_image_shape:
+    - 518
+    - 518
+    head_mode: pcd
+    num_level: 3
+    gs_param_dim: 256
+    align_corners: false
+    use_voxelize: true
+  decoder:
+    name: splatting_cuda
+    background_color:
+    - 0.0
+    - 0.0
+    - 0.0
+    make_scale_invariant: false
+  density_control:
+    name: density_control_module
+    mean_dim: 32
+    gs_param_dim: 256
+    refinement_layer_num: 1
+    num_level: 3
+    grad_mode: absgrad
+    use_mean_features: true
+    refinement_type: voxelize
+    refinement_hidden_dim: 32
+    aggregation_mode: mean
+    num_heads: 1
+    score_mode: absgrad
+    latent_dim: 128
+    num_latents: 64
+    num_self_attn_per_block: 2
+    voxel_size: 0.001
+    aux_refine: false
+    refine_error: false
+    use_refine_module: false
+    voxelize_activate: true
+    use_depth: false
+render_loss:
+  mse:
+    weight: 1.0
+  lpips:
+    weight: 0.05
+    apply_after_step: 0
+density_control_loss:
+  error_score:
+    weight: 0.01
+    log_scale: false
+    grad_scale: 10000.0
+    mode: original
+direct_loss:
+  l1:
+    weight: 0.8
+  ssim:
+    weight: 0.2
+wandb:
+  project: DCSplat
+  entity: scene-representation-group
+  name: ABLATION_0225_noRefineModule
+  mode: online
+  tags:
+  - re10k
+  - 256x256
+mode: train
+data_loader:
+  train:
+    num_workers: 16
+    persistent_workers: true
+    batch_size: 16
+    seed: 1234
+  test:
+    num_workers: 4
+    persistent_workers: false
+    batch_size: 1
+    seed: 2345
+  val:
+    num_workers: 1
+    persistent_workers: true
+    batch_size: 1
+    seed: 3456
+optimizer:
+  lr: 0.0002
+  warm_up_steps: 25
+  backbone_lr_multiplier: 0.1
+  backbone_trainable: T+H
+  accumulate: 1
+checkpointing:
+  load: null
+  every_n_train_steps: 1500
+  save_top_k: 2
+  save_weights_only: false
+train:
+  extended_visualization: false
+  print_log_every_n_steps: 10
+  camera_loss: 10.0
+  one_sample_validation: null
+  align_corners: false
+  intrinsic_scaling: false
+  verbose: false
+  beta_dist_param:
+  - 0.5
+  - 4.0
+  use_refine_aux: false
+  train_target_set: true
+  train_gs_num: 1
+  ext_scale_detach: false
+  cam_scale_mode: sum
+  scene_scale_reg_loss: 0.01
+  train_aux: true
+  vggt_cam_loss: true
+  vggt_distil: false
+  context_view_train: false
+test:
+  output_path: test/ablation/re10k
+  align_pose: false
+  pose_align_steps: 100
+  rot_opt_lr: 0.005
+  trans_opt_lr: 0.005
+  compute_scores: true
+  save_image: false
+  save_video: false
+  save_active_mask_image: false
+  save_error_score_image: false
+  save_compare: false
+  pred_intrinsic: false
+  error_threshold: 0.4
+  error_threshold_list:
+  - 0.2
+  - 0.4
+  - 0.6
+  - 0.8
+  - 1.0
+  threshold_mode: ratio
+  nvs_view_N_list:
+  - 3
+  - 6
+  - 16
+  - 32
+  - 64
+seed: 111123
+trainer:
+  max_steps: 3001
+  val_check_interval: 250
+  gradient_clip_val: 0.5
+  num_nodes: 1
+dataset:
+  re10k:
+    make_baseline_1: true
+    relative_pose: true
+    augment: true
+    background_color:
+    - 0.0
+    - 0.0
+    - 0.0
+    overfit_to_scene: null
+    skip_bad_shape: true
+    view_sampler:
+      name: bounded
+      num_target_views: 4
+      num_context_views: 2
+      min_distance_between_context_views: 45
+      max_distance_between_context_views: 90
+      min_distance_to_context_views: 0
+      warm_up_steps: 1000
+      initial_min_distance_between_context_views: 25
+      initial_max_distance_between_context_views: 25
+      same_target_gap: false
+      num_target_set: 3
+    name: re10k
+    roots:
+    - datasets/re10k
+    input_image_shape:
+    - 256
+    - 256
+    original_image_shape:
+    - 360
+    - 640
+    cameras_are_circular: false
+    baseline_min: 0.001
+    baseline_max: 10000000000.0
+    max_fov: 100.0
+    dynamic_context_views: true
+    max_context_views_per_gpu: 24
diff --git a/ABLATION_0225_noRefineModule/.hydra/hydra.yaml b/ABLATION_0225_noRefineModule/.hydra/hydra.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..f475f79f5ec70e008366fee3703f8eaf53359964
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/.hydra/hydra.yaml
@@ -0,0 +1,165 @@
+hydra:
+  run:
+    dir: outputs/ablation/re10k/${wandb.name}
+  sweep:
+    dir: multirun/${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:
+    - +experiment=re10k_ablation_24v
+    - wandb.mode=online
+    - wandb.name=ABLATION_0225_noRefineModule
+    - model.density_control.use_refine_module=false
+  job:
+    name: main
+    chdir: null
+    override_dirname: +experiment=re10k_ablation_24v,model.density_control.use_refine_module=false,wandb.mode=online,wandb.name=ABLATION_0225_noRefineModule
+    id: ???
+    num: ???
+    config_name: main
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.3.2
+    version_base: '1.3'
+    cwd: /workspace/code/CVPR2026
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/code/CVPR2026/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_noRefineModule
+    choices:
+      experiment: re10k_ablation_24v
+      dataset@dataset.re10k: re10k
+      dataset/view_sampler_dataset_specific_config@dataset.re10k.view_sampler: bounded_re10k
+      dataset/view_sampler@dataset.re10k.view_sampler: bounded
+      model/density_control: density_control_module
+      model/decoder: splatting_cuda
+      model/encoder: dcsplat
+      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
diff --git a/ABLATION_0225_noRefineModule/.hydra/overrides.yaml b/ABLATION_0225_noRefineModule/.hydra/overrides.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..d99b1c96c27191227b2ef2c4c4fb5cfb66d48b7f
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/.hydra/overrides.yaml
@@ -0,0 +1,4 @@
+- +experiment=re10k_ablation_24v
+- wandb.mode=online
+- wandb.name=ABLATION_0225_noRefineModule
+- model.density_control.use_refine_module=false
diff --git a/ABLATION_0225_noRefineModule/main.log b/ABLATION_0225_noRefineModule/main.log
new file mode 100644
index 0000000000000000000000000000000000000000..6287e741684ff8fee255ee3b6f82b487a753c872
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/main.log
@@ -0,0 +1,128 @@
+[2026-02-25 07:31:34,037][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 07:31:40,112][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 07:31:40,112][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 07:32:30,542][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/data_connector.py:425: The 'val_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=31` in the `DataLoader` to improve performance.
+
+[2026-02-25 07:32:30,543][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 07:32:33,093][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:32:33,103][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/utilities/data.py:79: Trying to infer the `batch_size` from an ambiguous collection. The batch size we found is 1. To avoid any miscalculations, use `self.log(..., batch_size=batch_size)`.
+
+[2026-02-25 07:32:33,104][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 07:32:33,104][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 07:32:34,792][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/functional.py:554: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /pytorch/aten/src/ATen/native/TensorShape.cpp:4322.)
+  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
+
+[2026-02-25 07:32:35,076][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/psnr', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 07:32:35,077][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/lpips', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 07:32:35,077][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/ssim', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 07:32:35,078][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/gaussian_num_ratio', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 07:32:35,078][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('info/global_step', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 07:32:44,871][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [57, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [57, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 07:32:44,967][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:34:17,416][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 07:45:01,533][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:48:10,917][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:57:27,231][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:03:33,811][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:09:48,816][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:19:01,130][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:22:10,768][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:34:25,661][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:34:29,312][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:46:50,776][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:49:55,355][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:59:12,245][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:05:35,984][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:11:48,010][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:21:06,680][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:24:15,287][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:36:29,623][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:36:33,850][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:48:56,864][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:52:05,306][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:01:25,665][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:07:31,359][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:13:42,512][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:22:55,254][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:26:05,189][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:38:39,652][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:38:43,134][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_noRefineModule/peak_vram_memory.json b/ABLATION_0225_noRefineModule/peak_vram_memory.json
new file mode 100644
index 0000000000000000000000000000000000000000..5664fd2bf9604fb14b1c1adfa1d73c8a3dadf075
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/peak_vram_memory.json
@@ -0,0 +1,6 @@
+{
+  "peak_memory_allocated_gb": 96.07,
+  "peak_memory_reserved_gb": 136.279,
+  "total_elapsed_hours": 3.12,
+  "mode": "train"
+}
\ No newline at end of file
diff --git a/ABLATION_0225_noRefineModule/train_ddp_process_3.log b/ABLATION_0225_noRefineModule/train_ddp_process_3.log
new file mode 100644
index 0000000000000000000000000000000000000000..054ca0733848115bbf5b8dfe0f966961cce62d6c
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/train_ddp_process_3.log
@@ -0,0 +1,66 @@
+[2026-02-25 07:31:50,767][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 07:32:08,454][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 07:32:08,455][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 07:32:30,542][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 07:32:44,868][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [57, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [57, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 07:32:45,002][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:34:17,440][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 07:45:01,533][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:57:27,232][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:09:48,815][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:22:10,768][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:34:29,312][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:46:50,775][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:59:12,243][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:11:48,007][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:24:15,287][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:36:33,848][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:48:56,863][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:01:25,665][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:13:42,514][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:26:05,189][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:38:43,134][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_noRefineModule/train_ddp_process_4.log b/ABLATION_0225_noRefineModule/train_ddp_process_4.log
new file mode 100644
index 0000000000000000000000000000000000000000..6f1667b76384f9f8fd5f473a576783af2f137f12
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/train_ddp_process_4.log
@@ -0,0 +1,66 @@
+[2026-02-25 07:31:50,601][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 07:32:19,908][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 07:32:19,908][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 07:32:30,542][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 07:32:44,872][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [57, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [57, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 07:32:45,084][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:34:17,446][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 07:45:01,534][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:57:27,231][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:09:48,816][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:22:10,768][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:34:29,312][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:46:50,775][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:59:12,243][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:11:48,007][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:24:15,287][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:36:33,848][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:48:56,863][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:01:25,665][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:13:42,512][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:26:05,190][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:38:43,134][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_noRefineModule/train_ddp_process_7.log b/ABLATION_0225_noRefineModule/train_ddp_process_7.log
new file mode 100644
index 0000000000000000000000000000000000000000..ff5a1d3f99bbeb665a67abe38cb7a401f137fa8c
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/train_ddp_process_7.log
@@ -0,0 +1,66 @@
+[2026-02-25 07:31:50,806][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 07:32:14,953][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 07:32:14,956][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 07:32:30,542][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 07:32:44,356][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [57, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [57, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 07:32:44,996][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:34:17,417][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 07:45:01,533][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:57:27,231][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:09:48,816][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:22:10,770][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:34:29,312][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:46:50,775][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 08:59:12,244][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:11:48,007][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:24:15,287][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:36:33,848][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 09:48:56,863][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:01:25,665][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:13:42,512][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:26:05,189][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:38:43,142][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_noRefineModule/wandb/debug-internal.log b/ABLATION_0225_noRefineModule/wandb/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..77a6a5015e15fecf0a7233dd1c15eced819110aa
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/wandb/debug-internal.log
@@ -0,0 +1,11 @@
+{"time":"2026-02-25T07:32:27.611867617Z","level":"INFO","msg":"stream: starting","core version":"0.25.0"}
+{"time":"2026-02-25T07:32:28.03755666Z","level":"INFO","msg":"stream: created new stream","id":"2f0bcys0"}
+{"time":"2026-02-25T07:32:28.037863635Z","level":"INFO","msg":"handler: started","stream_id":"2f0bcys0"}
+{"time":"2026-02-25T07:32:28.037970207Z","level":"INFO","msg":"stream: started","id":"2f0bcys0"}
+{"time":"2026-02-25T07:32:28.038020847Z","level":"INFO","msg":"writer: started","stream_id":"2f0bcys0"}
+{"time":"2026-02-25T07:32:28.038027757Z","level":"INFO","msg":"sender: started","stream_id":"2f0bcys0"}
+{"time":"2026-02-25T10:38:52.520830581Z","level":"INFO","msg":"stream: closing","id":"2f0bcys0"}
+{"time":"2026-02-25T10:38:53.390340772Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-02-25T10:38:53.699950002Z","level":"INFO","msg":"handler: closed","stream_id":"2f0bcys0"}
+{"time":"2026-02-25T10:38:53.700227926Z","level":"INFO","msg":"sender: closed","stream_id":"2f0bcys0"}
+{"time":"2026-02-25T10:38:53.700251656Z","level":"INFO","msg":"stream: closed","id":"2f0bcys0"}
diff --git a/ABLATION_0225_noRefineModule/wandb/debug.log b/ABLATION_0225_noRefineModule/wandb/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..2dab590a5f31eeb0e6d17e20ae047b827a5d4d4d
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/wandb/debug.log
@@ -0,0 +1,21 @@
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_setup.py:_flush():81] Current SDK version is 0.25.0
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_setup.py:_flush():81] Configure stats pid to 137621
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug.log
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug-internal.log
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_init.py:init():844] calling init triggers
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
+config: {'model': {'encoder': {'name': 'dcsplat', 'input_image_shape': [518, 518], 'head_mode': 'pcd', 'num_level': 3, 'gs_param_dim': 256, 'align_corners': False, 'use_voxelize': True}, 'decoder': {'name': 'splatting_cuda', 'background_color': [0.0, 0.0, 0.0], 'make_scale_invariant': False}, 'density_control': {'name': 'density_control_module', 'mean_dim': 32, 'gs_param_dim': 256, 'refinement_layer_num': 1, 'num_level': 3, 'grad_mode': 'absgrad', 'use_mean_features': True, 'refinement_type': 'voxelize', 'refinement_hidden_dim': 32, 'aggregation_mode': 'mean', 'num_heads': 1, 'score_mode': 'absgrad', 'latent_dim': 128, 'num_latents': 64, 'num_self_attn_per_block': 2, 'voxel_size': 0.001, 'aux_refine': False, 'refine_error': False, 'use_refine_module': False, 'voxelize_activate': True, 'use_depth': False}}, 'render_loss': {'mse': {'weight': 1.0}, 'lpips': {'weight': 0.05, 'apply_after_step': 0}}, 'density_control_loss': {'error_score': {'weight': 0.01, 'log_scale': False, 'grad_scale': 10000.0, 'mode': 'original'}}, 'direct_loss': {'l1': {'weight': 0.8}, 'ssim': {'weight': 0.2}}, 'wandb': {'project': 'DCSplat', 'entity': 'scene-representation-group', 'name': 'ABLATION_0225_noRefineModule', 'mode': 'online', 'tags': ['re10k', '256x256']}, 'mode': 'train', 'data_loader': {'train': {'num_workers': 16, 'persistent_workers': True, 'batch_size': 16, 'seed': 1234}, 'test': {'num_workers': 4, 'persistent_workers': False, 'batch_size': 1, 'seed': 2345}, 'val': {'num_workers': 1, 'persistent_workers': True, 'batch_size': 1, 'seed': 3456}}, 'optimizer': {'lr': 0.0002, 'warm_up_steps': 25, 'backbone_lr_multiplier': 0.1, 'backbone_trainable': 'T+H', 'accumulate': 1}, 'checkpointing': {'load': None, 'every_n_train_steps': 1500, 'save_top_k': 2, 'save_weights_only': False}, 'train': {'extended_visualization': False, 'print_log_every_n_steps': 10, 'camera_loss': 10.0, 'one_sample_validation': None, 'align_corners': False, 'intrinsic_scaling': False, 'verbose': False, 'beta_dist_param': [0.5, 4.0], 'use_refine_aux': False, 'train_target_set': True, 'train_gs_num': 1, 'ext_scale_detach': False, 'cam_scale_mode': 'sum', 'scene_scale_reg_loss': 0.01, 'train_aux': True, 'vggt_cam_loss': True, 'vggt_distil': False, 'context_view_train': False}, 'test': {'output_path': 'test/ablation/re10k', 'align_pose': False, 'pose_align_steps': 100, 'rot_opt_lr': 0.005, 'trans_opt_lr': 0.005, 'compute_scores': True, 'save_image': False, 'save_video': False, 'save_active_mask_image': False, 'save_error_score_image': False, 'save_compare': False, 'pred_intrinsic': False, 'error_threshold': 0.4, 'error_threshold_list': [0.2, 0.4, 0.6, 0.8, 1.0], 'threshold_mode': 'ratio', 'nvs_view_N_list': [3, 6, 16, 32, 64]}, 'seed': 111123, 'trainer': {'max_steps': 3001, 'val_check_interval': 250, 'gradient_clip_val': 0.5, 'num_nodes': 1}, 'dataset': {'re10k': {'make_baseline_1': True, 'relative_pose': True, 'augment': True, 'background_color': [0.0, 0.0, 0.0], 'overfit_to_scene': None, 'skip_bad_shape': True, 'view_sampler': {'name': 'bounded', 'num_target_views': 4, 'num_context_views': 2, 'min_distance_between_context_views': 45, 'max_distance_between_context_views': 90, 'min_distance_to_context_views': 0, 'warm_up_steps': 1000, 'initial_min_distance_between_context_views': 25, 'initial_max_distance_between_context_views': 25, 'same_target_gap': False, 'num_target_set': 3}, 'name': 're10k', 'roots': ['datasets/re10k'], 'input_image_shape': [256, 256], 'original_image_shape': [360, 640], 'cameras_are_circular': False, 'baseline_min': 0.001, 'baseline_max': 10000000000.0, 'max_fov': 100.0, 'dynamic_context_views': True, 'max_context_views_per_gpu': 24}}, '_wandb': {}}
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_init.py:init():892] starting backend
+2026-02-25 07:32:27,602 INFO    MainThread:137621 [wandb_init.py:init():895] sending inform_init request
+2026-02-25 07:32:27,609 INFO    MainThread:137621 [wandb_init.py:init():903] backend started and connected
+2026-02-25 07:32:27,613 INFO    MainThread:137621 [wandb_init.py:init():973] updated telemetry
+2026-02-25 07:32:27,622 INFO    MainThread:137621 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
+2026-02-25 07:32:28,628 INFO    MainThread:137621 [wandb_init.py:init():1042] starting run threads in backend
+2026-02-25 07:32:28,738 INFO    MainThread:137621 [wandb_run.py:_console_start():2524] atexit reg
+2026-02-25 07:32:28,738 INFO    MainThread:137621 [wandb_run.py:_redirect():2373] redirect: wrap_raw
+2026-02-25 07:32:28,738 INFO    MainThread:137621 [wandb_run.py:_redirect():2442] Wrapping output streams.
+2026-02-25 07:32:28,738 INFO    MainThread:137621 [wandb_run.py:_redirect():2465] Redirects installed.
+2026-02-25 07:32:28,740 INFO    MainThread:137621 [wandb_init.py:init():1082] run started, returning control to user process
+2026-02-25 10:38:52,520 INFO    wandb-AsyncioManager-main:137621 [service_client.py:_forward_responses():134] Reached EOF.
+2026-02-25 10:38:52,520 INFO    wandb-AsyncioManager-main:137621 [mailbox.py:close():155] Closing mailbox, abandoning 1 handles.
diff --git a/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/config.yaml b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..b79b9c1694a77dce96de2796de5377d8752736b0
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/config.yaml
@@ -0,0 +1,307 @@
+_wandb:
+    value:
+        cli_version: 0.25.0
+        e:
+            z1winms0ab80rmcbaynf075otkwpygrq:
+                args:
+                    - +experiment=re10k_ablation_24v
+                    - wandb.mode=online
+                    - wandb.name=ABLATION_0225_noRefineModule
+                    - model.density_control.use_refine_module=false
+                cpu_count: 128
+                cpu_count_logical: 256
+                cudaVersion: "13.1"
+                disk:
+                    /:
+                        total: "1170378588160"
+                        used: "708558733312"
+                email: dna9041@korea.ac.kr
+                executable: /venv/main/bin/python
+                git:
+                    commit: 2512754c6c27ca5150bf17fbcbdde3f192fd53cc
+                    remote: git@github.com:K-nowing/CVPR2026.git
+                gpu: NVIDIA H200
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-2649ab80-a3a6-5a1c-0fa5-12bc11bd75e9
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-e92921d9-c681-246f-af93-637e0dc938ca
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-ffe12ffc-9bb7-82de-5692-1ec0ee2e68d8
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-499e5acd-b6ab-2010-c51b-ee9b5aa65825
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-3b2522d9-1c72-e49b-2c30-96165680b74a
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-a9a280c5-b2f9-dc1e-a8a9-7326a74001ff
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-07d0167b-a6a1-1900-2d27-7c6c11598409
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-8362a999-20d1-c27b-5d18-032d23f859ab
+                host: 27d18dedec6d
+                memory:
+                    total: "1622948257792"
+                os: Linux-6.8.0-90-generic-x86_64-with-glibc2.39
+                program: -m src.main
+                python: CPython 3.12.12
+                root: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_noRefineModule
+                startedAt: "2026-02-25T07:32:27.352870Z"
+                writerId: z1winms0ab80rmcbaynf075otkwpygrq
+        m:
+            - "1": trainer/global_step
+              "6":
+                - 3
+              "7": []
+            - "2": '*'
+              "5": 1
+              "6":
+                - 1
+              "7": []
+        python_version: 3.12.12
+        t:
+            "1":
+                - 1
+                - 41
+                - 49
+                - 50
+                - 106
+            "2":
+                - 1
+                - 41
+                - 49
+                - 50
+                - 106
+            "3":
+                - 7
+                - 13
+                - 15
+                - 16
+                - 66
+            "4": 3.12.12
+            "5": 0.25.0
+            "12": 0.25.0
+            "13": linux-x86_64
+checkpointing:
+    value:
+        every_n_train_steps: 1500
+        load: null
+        save_top_k: 2
+        save_weights_only: false
+data_loader:
+    value:
+        test:
+            batch_size: 1
+            num_workers: 4
+            persistent_workers: false
+            seed: 2345
+        train:
+            batch_size: 16
+            num_workers: 16
+            persistent_workers: true
+            seed: 1234
+        val:
+            batch_size: 1
+            num_workers: 1
+            persistent_workers: true
+            seed: 3456
+dataset:
+    value:
+        re10k:
+            augment: true
+            background_color:
+                - 0
+                - 0
+                - 0
+            baseline_max: 1e+10
+            baseline_min: 0.001
+            cameras_are_circular: false
+            dynamic_context_views: true
+            input_image_shape:
+                - 256
+                - 256
+            make_baseline_1: true
+            max_context_views_per_gpu: 24
+            max_fov: 100
+            name: re10k
+            original_image_shape:
+                - 360
+                - 640
+            overfit_to_scene: null
+            relative_pose: true
+            roots:
+                - datasets/re10k
+            skip_bad_shape: true
+            view_sampler:
+                initial_max_distance_between_context_views: 25
+                initial_min_distance_between_context_views: 25
+                max_distance_between_context_views: 90
+                min_distance_between_context_views: 45
+                min_distance_to_context_views: 0
+                name: bounded
+                num_context_views: 2
+                num_target_set: 3
+                num_target_views: 4
+                same_target_gap: false
+                warm_up_steps: 1000
+density_control_loss:
+    value:
+        error_score:
+            grad_scale: 10000
+            log_scale: false
+            mode: original
+            weight: 0.01
+direct_loss:
+    value:
+        l1:
+            weight: 0.8
+        ssim:
+            weight: 0.2
+mode:
+    value: train
+model:
+    value:
+        decoder:
+            background_color:
+                - 0
+                - 0
+                - 0
+            make_scale_invariant: false
+            name: splatting_cuda
+        density_control:
+            aggregation_mode: mean
+            aux_refine: false
+            grad_mode: absgrad
+            gs_param_dim: 256
+            latent_dim: 128
+            mean_dim: 32
+            name: density_control_module
+            num_heads: 1
+            num_latents: 64
+            num_level: 3
+            num_self_attn_per_block: 2
+            refine_error: false
+            refinement_hidden_dim: 32
+            refinement_layer_num: 1
+            refinement_type: voxelize
+            score_mode: absgrad
+            use_depth: false
+            use_mean_features: true
+            use_refine_module: false
+            voxel_size: 0.001
+            voxelize_activate: true
+        encoder:
+            align_corners: false
+            gs_param_dim: 256
+            head_mode: pcd
+            input_image_shape:
+                - 518
+                - 518
+            name: dcsplat
+            num_level: 3
+            use_voxelize: true
+optimizer:
+    value:
+        accumulate: 1
+        backbone_lr_multiplier: 0.1
+        backbone_trainable: T+H
+        lr: 0.0002
+        warm_up_steps: 25
+render_loss:
+    value:
+        lpips:
+            apply_after_step: 0
+            weight: 0.05
+        mse:
+            weight: 1
+seed:
+    value: 111123
+test:
+    value:
+        align_pose: false
+        compute_scores: true
+        error_threshold: 0.4
+        error_threshold_list:
+            - 0.2
+            - 0.4
+            - 0.6
+            - 0.8
+            - 1
+        nvs_view_N_list:
+            - 3
+            - 6
+            - 16
+            - 32
+            - 64
+        output_path: test/ablation/re10k
+        pose_align_steps: 100
+        pred_intrinsic: false
+        rot_opt_lr: 0.005
+        save_active_mask_image: false
+        save_compare: false
+        save_error_score_image: false
+        save_image: false
+        save_video: false
+        threshold_mode: ratio
+        trans_opt_lr: 0.005
+train:
+    value:
+        align_corners: false
+        beta_dist_param:
+            - 0.5
+            - 4
+        cam_scale_mode: sum
+        camera_loss: 10
+        context_view_train: false
+        ext_scale_detach: false
+        extended_visualization: false
+        intrinsic_scaling: false
+        one_sample_validation: null
+        print_log_every_n_steps: 10
+        scene_scale_reg_loss: 0.01
+        train_aux: true
+        train_gs_num: 1
+        train_target_set: true
+        use_refine_aux: false
+        verbose: false
+        vggt_cam_loss: true
+        vggt_distil: false
+trainer:
+    value:
+        gradient_clip_val: 0.5
+        max_steps: 3001
+        num_nodes: 1
+        val_check_interval: 250
+wandb:
+    value:
+        entity: scene-representation-group
+        mode: online
+        name: ABLATION_0225_noRefineModule
+        project: DCSplat
+        tags:
+            - re10k
+            - 256x256
diff --git a/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/output.log b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/output.log
new file mode 100644
index 0000000000000000000000000000000000000000..ed98021f76c967a12e1b1a6f4455e7024246409b
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/output.log
@@ -0,0 +1,800 @@
+LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0,1,2,3,4,5,6,7]
+
+  | Name                   | Type                 | Params | Mode
+------------------------------------------------------------------------
+0 | encoder                | OurSplat             | 888 M  | train
+1 | density_control_module | DensityControlModule | 514    | train
+2 | decoder                | DecoderSplattingCUDA | 0      | train
+3 | render_losses          | ModuleList           | 0      | train
+4 | density_control_losses | ModuleList           | 0      | train
+5 | direct_losses          | ModuleList           | 0      | train
+------------------------------------------------------------------------
+888 M     Trainable params
+0         Non-trainable params
+888 M     Total params
+3,553.936 Total estimated model params size (MB)
+1207      Modules in train mode
+522       Modules in eval mode
+Sanity Checking: |                                                                                           | 0/? [00:00<?, ?it/s][2026-02-25 07:32:30,542][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/data_connector.py:425: The 'val_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=31` in the `DataLoader` to improve performance.
+
+[2026-02-25 07:32:30,543][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user.
+  warnings.warn(  # warn only once
+
+Validation epoch start on rank 0
+Sanity Checking DataLoader 0:   0%|                                                                          | 0/1 [00:00<?, ?it/s]validation step 0; scene = ['306e2b7785657539'];
+target intrinsic:  tensor(0.8595, device='cuda:0') tensor(0.8597, device='cuda:0')
+pred intrinsic:  tensor(0.8779, device='cuda:0') tensor(0.8773, device='cuda:0')
+[rank0]:W0225 07:32:33.024000 137621 site-packages/torch/utils/cpp_extension.py:2425] TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation.
+[rank0]:W0225 07:32:33.024000 137621 site-packages/torch/utils/cpp_extension.py:2425] If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'] to specific architectures.
+[2026-02-25 07:32:33,093][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:32:33,103][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/utilities/data.py:79: Trying to infer the `batch_size` from an ambiguous collection. The batch size we found is 1. To avoid any miscalculations, use `self.log(..., batch_size=batch_size)`.
+
+Setting up [LPIPS] perceptual loss: trunk [vgg], v[0.1], spatial [off]
+[2026-02-25 07:32:33,104][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 07:32:33,104][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+Loading model from: /venv/main/lib/python3.12/site-packages/lpips/weights/v0.1/vgg.pth
+[2026-02-25 07:32:34,792][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/functional.py:554: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /pytorch/aten/src/ATen/native/TensorShape.cpp:4322.)
+  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
+
+Sanity Checking DataLoader 0: 100%|██████████████████████████████████████████████████████████████████| 1/1 [00:03<00:00,  0.26it/s][2026-02-25 07:32:35,076][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/psnr', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 07:32:35,077][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/lpips', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 07:32:35,077][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/ssim', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 07:32:35,078][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/gaussian_num_ratio', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 07:32:35,078][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('info/global_step', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+Epoch 0: |                                                                                                   | 0/? [00:00<?, ?it/s]context = [[34, 36, 50, 53, 54, 60, 63, 70, 76, 78, 79, 80, 81, 88, 92, 94, 102, 110, 112, 114, 122, 125, 126, 131]]target = [[126, 96, 109, 55, 99, 116, 43, 60, 113, 85, 103, 90, 130, 62, 76, 123, 35, 102, 125, 128, 98, 67, 129, 79]]
+[2026-02-25 07:32:44,871][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [57, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [57, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 07:32:44,967][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                               | 9/? [00:41<00:00,  0.22it/s, v_num=cys0]train step 10; scene = [['08c26703c4987851']]; loss = 0.906105
+Epoch 0: |                                                                              | 10/? [00:45<00:00,  0.22it/s, v_num=cys0]context = [[98, 107, 112, 119, 122, 123], [21, 22, 27, 36, 38, 46], [63, 66, 67, 77, 84, 88], [56, 62, 65, 68, 78, 81]]target = [[105, 110, 116, 112, 111, 101], [29, 37, 36, 22, 40, 45], [73, 66, 75, 87, 83, 64], [79, 69, 75, 58, 61, 62]]
+Epoch 0: |                                                                              | 19/? [01:17<00:00,  0.24it/s, v_num=cys0]train step 20; scene = [['4012c15c8381568b'], ['af08406c5a9a43a0'], ['9f9f9beffb86fad7'], ['fc8d08df6c875cb0']]; loss = 0.231639
+Epoch 0: |                                                                              | 20/? [01:21<00:00,  0.25it/s, v_num=cys0]context = [[144, 152, 157, 164, 166, 169, 171, 177], [201, 211, 216, 221, 228, 230, 233, 234], [11, 15, 16, 23, 30, 38, 43, 44]]target = [[153, 170, 149, 169, 145, 174, 165, 157], [229, 216, 205, 206, 203, 213, 233, 215], [37, 38, 39, 35, 15, 24, 19, 25]]
+Epoch 0: |                                                                              | 24/? [01:35<00:00,  0.25it/s, v_num=cys0][2026-02-25 07:34:17,416][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+Epoch 0: |                                                                              | 29/? [01:53<00:00,  0.26it/s, v_num=cys0]train step 30; scene = [['00980329a3221f1c'], ['1e7c432d2207b6f2'], ['af2748330e5243d0']]; loss = 0.180660
+Epoch 0: |                                                                              | 30/? [01:56<00:00,  0.26it/s, v_num=cys0]context = [[2, 7, 10, 15, 20, 24, 25, 29, 30, 32, 33, 37, 40, 44, 49, 60, 64, 74, 76, 81, 89, 91, 95, 99]]target = [[79, 84, 35, 43, 89, 44, 63, 58, 48, 13, 65, 7, 96, 27, 51, 20, 60, 15, 85, 59, 22, 66, 17, 62]]
+Epoch 0: |                                                                              | 39/? [02:29<00:00,  0.26it/s, v_num=cys0]train step 40; scene = [['79a9385753d426bc'], ['593538382d2dc847'], ['c9c67636b9d521be']]; loss = 0.146315
+Epoch 0: |                                                                              | 40/? [02:32<00:00,  0.26it/s, v_num=cys0]context = [[200, 208, 223, 225], [6, 8, 12, 31], [104, 107, 111, 129], [69, 80, 86, 95], [11, 15, 18, 37], [52, 60, 70, 78]]target = [[213, 203, 214, 201], [29, 11, 15, 28], [111, 122, 120, 115], [77, 90, 89, 78], [18, 24, 29, 14], [64, 67, 63, 74]]
+Epoch 0: |                                                                              | 49/? [03:04<00:00,  0.27it/s, v_num=cys0]train step 50; scene = [['579a11551b3315d9'], ['c9dd64b7415e788e'], ['6f3fb517d1798d03']]; loss = 0.139808
+Epoch 0: |                                                                              | 50/? [03:08<00:00,  0.27it/s, v_num=cys0]context = [[20, 22, 25, 28, 36, 47], [10, 13, 15, 22, 26, 35], [31, 47, 49, 51, 52, 56], [109, 110, 115, 116, 125, 135]]target = [[40, 24, 23, 30, 37, 33], [19, 32, 25, 24, 26, 28], [35, 52, 50, 55, 45, 36], [131, 123, 130, 110, 115, 134]]
+Epoch 0: |                                                                              | 59/? [03:41<00:00,  0.27it/s, v_num=cys0]train step 60; scene = [['07916b8004a8e336'], ['e51ef9945ae527c4'], ['db84f84b1d775bb8'], ['92ed61f8e16b7e67']]; loss = 0.140808
+Epoch 0: |                                                                              | 60/? [03:44<00:00,  0.27it/s, v_num=cys0]context = [[96, 97, 101, 102, 105, 106, 107, 114, 124, 134, 138, 145], [76, 78, 82, 83, 96, 98, 100, 111, 112, 118, 119, 125]]target = [[144, 124, 137, 102, 107, 130, 119, 129, 118, 123, 133, 126], [98, 108, 96, 100, 113, 123, 77, 78, 112, 109, 80, 102]]
+Epoch 0: |                                                                              | 69/? [04:18<00:00,  0.27it/s, v_num=cys0]train step 70; scene = [['c34efa1505a0cfaa'], ['a3d0cca9fb57fd85'], ['43d0e6dce7bb1e95'], ['d8c2f0a3734cb493']]; loss = 0.118299
+Epoch 0: |                                                                              | 70/? [04:21<00:00,  0.27it/s, v_num=cys0]context = [[204, 208, 211, 218, 232, 234, 237, 239, 240, 241, 247, 253], [6, 7, 10, 12, 35, 36, 37, 38, 40, 45, 48, 55]]target = [[230, 236, 222, 227, 220, 252, 216, 221, 213, 217, 249, 240], [18, 24, 25, 12, 9, 42, 34, 52, 41, 33, 43, 48]]
+Epoch 0: |                                                                              | 79/? [04:53<00:00,  0.27it/s, v_num=cys0]train step 80; scene = [['24d756c820744e31'], ['cd6c21656a85e9b9'], ['f3b24cf238154fc0']]; loss = 0.096660
+Epoch 0: |                                                                              | 80/? [04:57<00:00,  0.27it/s, v_num=cys0]context = [[4, 30], [52, 79], [61, 87], [12, 40], [83, 109], [3, 29], [221, 249], [198, 227], [9, 38], [46, 72], [0, 26], [123, 150]]target = [[27, 26], [58, 57], [70, 77], [26, 20], [87, 98], [26, 14], [223, 224], [221, 200], [23, 33], [67, 64], [4, 25], [140, 148]]
+Epoch 0: |                                                                              | 89/? [05:31<00:00,  0.27it/s, v_num=cys0]train step 90; scene = [['617b4bc98d7e0bb6'], ['666e4a9aba27bb64']]; loss = 0.081214
+Epoch 0: |                                                                              | 90/? [05:34<00:00,  0.27it/s, v_num=cys0]context = [[134, 135, 138, 146, 147, 151, 156, 161, 162, 164, 166, 167, 168, 169, 187, 189, 197, 210, 215, 224, 225, 228, 230, 231]]target = [[143, 149, 191, 182, 151, 226, 165, 140, 208, 171, 179, 223, 168, 136, 194, 207, 227, 144, 187, 185, 145, 218, 139, 170]]
+Epoch 0: |                                                                              | 99/? [06:08<00:00,  0.27it/s, v_num=cys0]train step 100; scene = [['12fee7f1978d52f1'], ['c963bb60939e2d81']]; loss = 0.117273
+Epoch 0: |                                                                             | 100/? [06:12<00:00,  0.27it/s, v_num=cys0]context = [[40, 44, 47, 48, 49, 50, 56, 63, 77, 79, 88, 89], [21, 33, 34, 35, 36, 40, 43, 46, 47, 60, 62, 70]]target = [[58, 81, 76, 64, 68, 72, 51, 87, 77, 65, 88, 45], [31, 47, 25, 35, 55, 22, 48, 65, 29, 40, 63, 67]]
+Epoch 0: |                                                                             | 109/? [06:44<00:00,  0.27it/s, v_num=cys0]train step 110; scene = [['47396d5a5299873e']]; loss = 0.131424
+Epoch 0: |                                                                             | 110/? [06:48<00:00,  0.27it/s, v_num=cys0]context = [[20, 22, 29, 30, 31, 34, 37, 45, 47, 51, 56, 69], [18, 22, 24, 27, 31, 37, 44, 48, 49, 60, 61, 67]]target = [[49, 22, 36, 59, 63, 60, 45, 66, 38, 28, 26, 64], [61, 65, 25, 20, 63, 60, 26, 22, 33, 45, 37, 35]]
+Epoch 0: |                                                                             | 119/? [07:21<00:00,  0.27it/s, v_num=cys0]train step 120; scene = [['9bd7044e7cbf8e60'], ['76e44cf6b5658b26']]; loss = 0.105025
+Epoch 0: |                                                                             | 120/? [07:25<00:00,  0.27it/s, v_num=cys0]context = [[7, 14, 26, 28, 31, 34, 37, 40], [10, 20, 26, 32, 35, 39, 41, 43], [17, 23, 26, 28, 44, 46, 49, 50]]target = [[14, 26, 10, 22, 30, 13, 31, 11], [24, 31, 36, 14, 40, 34, 41, 13], [24, 44, 34, 48, 46, 36, 26, 38]]
+Epoch 0: |                                                                             | 129/? [07:58<00:00,  0.27it/s, v_num=cys0]train step 130; scene = [['a8cef6a851fbea3c'], ['b6699f4d039a5b06'], ['55cf2bbe9e017ea4'], ['6b0dd861e1ab1fec'], ['14db202c335af709'], ['8b6ff6c5153a7794'], ['b75f3820760d835c'], ['f7dbc855fd2a7669'], ['cfb20f8971e6a591'], ['95f2be7bb8303f50'], ['ff422469e034ae11'], ['5a2ad43377e9d18d']]; loss = 0.167970
+Epoch 0: |                                                                             | 130/? [08:02<00:00,  0.27it/s, v_num=cys0]context = [[8, 15, 26, 30, 32, 35, 40, 46, 47, 48, 50, 53, 54, 62, 69, 70, 72, 76, 80, 86, 90, 99, 100, 105]]target = [[26, 12, 49, 100, 89, 24, 10, 81, 37, 63, 52, 17, 39, 70, 16, 56, 40, 55, 43, 34, 72, 28, 48, 45]]
+Epoch 0: |                                                                             | 139/? [08:34<00:00,  0.27it/s, v_num=cys0]train step 140; scene = [['f62a962df5c26a1a'], ['b076420679a04731']]; loss = 0.096117
+Epoch 0: |                                                                             | 140/? [08:37<00:00,  0.27it/s, v_num=cys0]context = [[90, 95, 97, 104, 118, 121], [14, 29, 34, 35, 36, 44], [12, 14, 22, 27, 34, 41], [6, 10, 16, 18, 34, 35]]target = [[99, 104, 93, 113, 110, 96], [29, 37, 22, 21, 39, 41], [19, 25, 39, 27, 29, 13], [26, 28, 9, 7, 13, 34]]
+Epoch 0: |                                                                             | 149/? [09:11<00:00,  0.27it/s, v_num=cys0]train step 150; scene = [['a52d26a78b04aebd']]; loss = 0.074294
+Epoch 0: |                                                                             | 150/? [09:14<00:00,  0.27it/s, v_num=cys0]context = [[115, 132, 134, 145], [16, 32, 39, 44], [57, 63, 78, 88], [9, 11, 23, 36], [15, 26, 28, 44], [1, 8, 22, 33]]target = [[124, 132, 116, 119], [27, 24, 33, 29], [81, 61, 85, 79], [28, 26, 29, 16], [28, 29, 37, 22], [9, 4, 32, 27]]
+Epoch 0: |                                                                             | 159/? [09:47<00:00,  0.27it/s, v_num=cys0]train step 160; scene = [['268fbffc6c479d5b']]; loss = 0.068108
+Epoch 0: |                                                                             | 160/? [09:51<00:00,  0.27it/s, v_num=cys0]context = [[18, 25, 26, 37, 42, 46, 49, 51, 53, 64, 65, 67], [69, 75, 78, 79, 82, 84, 94, 95, 104, 108, 117, 118]]target = [[53, 27, 22, 32, 41, 38, 50, 43, 47, 48, 23, 19], [74, 70, 114, 115, 90, 89, 88, 92, 94, 110, 107, 101]]
+Epoch 0: |                                                                             | 169/? [10:24<00:00,  0.27it/s, v_num=cys0]train step 170; scene = [['719e2e8912e4eed3'], ['a3e51565a737569f']]; loss = 0.106682
+Epoch 0: |                                                                             | 170/? [10:28<00:00,  0.27it/s, v_num=cys0]context = [[14, 18, 20, 21, 24, 27, 33, 40, 45, 47, 48, 51, 52, 60, 64, 70, 75, 77, 80, 85, 90, 98, 102, 111]]target = [[29, 93, 32, 39, 81, 108, 72, 107, 51, 35, 16, 36, 70, 18, 34, 92, 94, 47, 23, 74, 50, 77, 19, 37]]
+Epoch 0: |                                                                             | 179/? [11:01<00:00,  0.27it/s, v_num=cys0]train step 180; scene = [['f44b9aa76a94a0a3']]; loss = 0.068154
+Epoch 0: |                                                                             | 180/? [11:05<00:00,  0.27it/s, v_num=cys0]context = [[0, 6, 17, 22, 26, 28, 33, 41, 50, 55, 56, 57, 71, 76, 79, 81, 84, 85, 86, 87, 89, 95, 96, 97]]target = [[37, 49, 12, 78, 9, 16, 84, 13, 5, 4, 6, 38, 80, 51, 43, 68, 64, 46, 56, 24, 25, 72, 36, 21]]
+Epoch 0: |                                                                             | 189/? [11:38<00:00,  0.27it/s, v_num=cys0]train step 190; scene = [['71bb669d936a5718'], ['a47203cfd5e0a478'], ['4b009f82cf5c7098']]; loss = 0.082014
+Epoch 0: |                                                                             | 190/? [11:42<00:00,  0.27it/s, v_num=cys0]context = [[9, 11, 19, 25, 33, 43, 46, 47, 48, 54, 57, 58, 63, 70, 72, 75, 79, 80, 83, 84, 85, 96, 99, 106]]target = [[30, 53, 82, 31, 55, 12, 20, 72, 104, 70, 24, 52, 21, 32, 102, 71, 35, 11, 46, 10, 15, 74, 33, 26]]
+Epoch 0: |                                                                             | 199/? [12:15<00:00,  0.27it/s, v_num=cys0]train step 200; scene = [['dd5ec950a01c42a0'], ['6d0db0358f7e051e'], ['983fe650a925ec1b']]; loss = 0.101481
+Epoch 0: |                                                                             | 200/? [12:19<00:00,  0.27it/s, v_num=cys0]context = [[8, 10, 14, 15, 22, 23, 27, 30, 36, 38, 39, 47, 64, 65, 67, 78, 79, 80, 83, 86, 93, 96, 98, 105]]target = [[73, 63, 10, 27, 89, 35, 44, 58, 97, 71, 17, 24, 66, 87, 50, 12, 23, 11, 31, 45, 69, 96, 98, 94]]
+[2026-02-25 07:45:01,533][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 209/? [12:59<00:00,  0.27it/s, v_num=cys0]train step 210; scene = [['9be9b273b3c22c61'], ['4b5883872c9b860c']]; loss = 0.073700
+Epoch 0: |                                                                             | 210/? [13:02<00:00,  0.27it/s, v_num=cys0]context = [[34, 35, 37, 54, 56, 59, 61, 68, 69, 76, 90, 96, 98, 101, 103, 107, 116, 119, 120, 121, 122, 125, 126, 131]]target = [[95, 50, 119, 121, 84, 107, 72, 52, 80, 42, 127, 94, 79, 98, 46, 128, 73, 75, 106, 92, 37, 110, 96, 56]]
+Epoch 0: |                                                                             | 219/? [13:36<00:00,  0.27it/s, v_num=cys0]train step 220; scene = [['a3b6faa8d238d993'], ['df9ba36fbe753843']]; loss = 0.070477
+Epoch 0: |                                                                             | 220/? [13:40<00:00,  0.27it/s, v_num=cys0]context = [[39, 71, 74], [21, 41, 51], [28, 55, 59], [2, 31, 37], [15, 25, 48], [57, 64, 93], [76, 87, 105], [31, 53, 64]]target = [[52, 43, 59], [28, 25, 49], [48, 58, 51], [26, 9, 6], [30, 32, 47], [84, 88, 59], [104, 87, 88], [42, 52, 59]]
+Epoch 0: |                                                                             | 229/? [14:13<00:00,  0.27it/s, v_num=cys0]train step 230; scene = [['ca04de3c55cd1ca0'], ['3d90d586b33daa63'], ['d1772c09b4b6d95f'], ['03d05f69a1cab4f8'], ['60d296908f43a97a'], ['37c400e282bc481e']]; loss = 0.090698
+Epoch 0: |                                                                             | 230/? [14:17<00:00,  0.27it/s, v_num=cys0]context = [[203, 204, 208, 209, 210, 233], [41, 42, 43, 52, 60, 76], [2, 8, 9, 23, 29, 31], [70, 74, 92, 97, 98, 100]]target = [[226, 223, 210, 217, 228, 222], [46, 70, 71, 75, 43, 56], [24, 23, 29, 25, 3, 21], [95, 81, 74, 73, 98, 92]]
+Epoch 0: |                                                                             | 239/? [14:50<00:00,  0.27it/s, v_num=cys0]train step 240; scene = [['9794641b7e015578']]; loss = 0.128480
+Epoch 0: |                                                                             | 240/? [14:54<00:00,  0.27it/s, v_num=cys0]context = [[62, 65, 68, 71, 80, 86, 87, 96, 99, 101, 103, 111], [134, 138, 139, 142, 151, 163, 171, 172, 173, 174, 181, 183]]target = [[87, 89, 96, 93, 103, 71, 65, 77, 63, 98, 102, 105], [165, 177, 159, 147, 138, 152, 171, 141, 181, 146, 161, 176]]
+Epoch 0: |                                                                             | 249/? [15:28<00:00,  0.27it/s, v_num=cys0]train step 250; scene = [['93dff1b985f2c7f9']]; loss = 0.094821
+Epoch 0: |                                                                             | 250/? [15:31<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 250; scene = ['49b8f80c849dc341'];
+target intrinsic:  tensor(0.8891, device='cuda:0') tensor(0.8894, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8543, device='cuda:0') tensor(0.8496, device='cuda:0')
+[2026-02-25 07:48:10,917][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 250/? [15:33<00:00,  0.27it/s, v_num=cys0]context = [[113, 115, 120, 131, 135, 137, 144, 152], [16, 20, 23, 28, 32, 35, 39, 52], [14, 18, 25, 26, 28, 29, 51, 53]]target = [[149, 139, 118, 151, 137, 141, 121, 130], [45, 51, 28, 50, 35, 37, 27, 23], [29, 26, 22, 51, 24, 34, 47, 45]]
+Epoch 0: |                                                                             | 259/? [16:04<00:00,  0.27it/s, v_num=cys0]train step 260; scene = [['b2288bf7003d5d4d']]; loss = 0.087019
+Epoch 0: |                                                                             | 260/? [16:08<00:00,  0.27it/s, v_num=cys0]context = [[206, 208, 212, 216, 221, 236], [58, 60, 68, 69, 70, 93], [15, 19, 26, 34, 41, 44], [15, 18, 23, 34, 42, 44]]target = [[225, 216, 234, 214, 207, 232], [88, 76, 91, 67, 64, 70], [20, 29, 22, 36, 23, 40], [42, 28, 31, 30, 16, 37]]
+Epoch 0: |                                                                             | 269/? [16:41<00:00,  0.27it/s, v_num=cys0]train step 270; scene = [['013ec74a4fde6737'], ['78e816776b064fc4'], ['1b778f72bbee1f27'], ['c71549de92ecb2e4'], ['8e16c8644efeec52'], ['35c5fc80e85db7cd'], ['34c8c62d878eca66'], ['203a5fd3a45ac4a7']]; loss = 0.086147
+Epoch 0: |                                                                             | 270/? [16:45<00:00,  0.27it/s, v_num=cys0]context = [[160, 174, 177, 178, 192, 200], [34, 52, 54, 56, 58, 67], [78, 85, 101, 102, 109, 118], [123, 135, 136, 144, 150, 155]]target = [[169, 195, 175, 174, 194, 189], [55, 53, 57, 48, 40, 66], [90, 108, 110, 84, 109, 83], [131, 130, 127, 124, 139, 133]]
+Epoch 0: |                                                                             | 279/? [17:19<00:00,  0.27it/s, v_num=cys0]train step 280; scene = [['75335793f866b96d'], ['e9d9dc952f5bbd83']]; loss = 0.054612
+Epoch 0: |                                                                             | 280/? [17:23<00:00,  0.27it/s, v_num=cys0]context = [[2, 31], [39, 73], [69, 98], [58, 92], [109, 146], [27, 59], [83, 114], [199, 231], [13, 50], [77, 109], [72, 105], [7, 45]]target = [[20, 26], [71, 57], [92, 77], [61, 72], [125, 129], [44, 51], [110, 101], [211, 202], [27, 17], [98, 104], [80, 89], [36, 37]]
+Epoch 0: |                                                                             | 289/? [17:55<00:00,  0.27it/s, v_num=cys0]train step 290; scene = [['51252022ddf74fb9'], ['8dd73309b133b8bf'], ['9e8db62a9b3cbd5e'], ['d41e59ee023e977b'], ['ce1a9465dc08ef4c'], ['e7887dec76685627']]; loss = 0.075656
+Epoch 0: |                                                                             | 290/? [17:58<00:00,  0.27it/s, v_num=cys0]context = [[74, 83, 85, 88, 89, 93, 100, 101, 108, 119, 122, 123], [17, 20, 23, 24, 28, 30, 32, 35, 48, 52, 64, 66]]target = [[110, 79, 95, 92, 111, 109, 99, 89, 77, 98, 81, 102], [28, 29, 52, 36, 54, 64, 49, 23, 40, 58, 31, 27]]
+Epoch 0: |                                                                             | 299/? [18:32<00:00,  0.27it/s, v_num=cys0]train step 300; scene = [['0c5d83212982c0ec'], ['00793a8a3b268d7c'], ['47a9b1e96499a466'], ['a1fb990016d7b3af']]; loss = 0.060977
+Epoch 0: |                                                                             | 300/? [18:36<00:00,  0.27it/s, v_num=cys0]context = [[83, 88, 89, 91, 92, 100, 102, 104, 107, 111, 113, 115, 129, 132, 134, 135, 140, 142, 145, 150, 154, 173, 175, 180]]target = [[115, 140, 130, 169, 114, 163, 175, 177, 104, 152, 151, 89, 154, 123, 132, 85, 122, 119, 155, 91, 117, 171, 158, 157]]
+Epoch 0: |                                                                             | 309/? [19:09<00:00,  0.27it/s, v_num=cys0]train step 310; scene = [['9b73ab94b5c43711'], ['8c845b940aa8244c'], ['b2789c1a5c127a02'], ['3db6c0e172d18826']]; loss = 0.084985
+Epoch 0: |                                                                             | 310/? [19:12<00:00,  0.27it/s, v_num=cys0]context = [[30, 47, 66], [16, 31, 46], [10, 28, 40], [43, 68, 75], [21, 43, 63], [15, 29, 46], [36, 44, 66], [17, 39, 59]]target = [[51, 41, 31], [29, 43, 44], [23, 35, 33], [64, 71, 63], [48, 59, 27], [45, 29, 27], [60, 42, 50], [34, 26, 47]]
+Epoch 0: |                                                                             | 319/? [19:46<00:00,  0.27it/s, v_num=cys0]train step 320; scene = [['591cd9d079cd7842'], ['3dd7802a2c93a865']]; loss = 0.103070
+Epoch 0: |                                                                             | 320/? [19:49<00:00,  0.27it/s, v_num=cys0]context = [[4, 22, 28, 31, 34, 37, 45, 46], [26, 28, 38, 41, 46, 55, 57, 59], [129, 139, 145, 150, 151, 158, 167, 169]]target = [[7, 41, 13, 5, 20, 27, 32, 44], [29, 32, 54, 28, 40, 52, 43, 33], [137, 143, 146, 152, 130, 131, 145, 133]]
+Epoch 0: |                                                                             | 329/? [20:23<00:00,  0.27it/s, v_num=cys0]train step 330; scene = [['30d9f6321281dade'], ['2a08fac923c9e50d']]; loss = 0.078926
+Epoch 0: |                                                                             | 330/? [20:27<00:00,  0.27it/s, v_num=cys0]context = [[4, 6, 12, 14, 24, 41, 42, 44, 45, 48, 49, 53], [102, 106, 112, 113, 118, 122, 125, 130, 134, 140, 149, 151]]target = [[50, 42, 7, 28, 11, 33, 20, 45, 25, 8, 22, 16], [135, 134, 113, 128, 139, 112, 117, 108, 122, 127, 138, 120]]
+Epoch 0: |                                                                             | 339/? [20:59<00:00,  0.27it/s, v_num=cys0]train step 340; scene = [['bd9f2096d355b1b8'], ['07d3325178e7a790'], ['8204d757ce43dda8']]; loss = 0.085941
+Epoch 0: |                                                                             | 340/? [21:03<00:00,  0.27it/s, v_num=cys0]context = [[87, 88, 93, 105, 107, 111, 116, 118, 123, 131, 132, 136], [104, 111, 116, 121, 122, 127, 131, 132, 136, 137, 142, 153]]target = [[114, 108, 107, 130, 95, 129, 118, 104, 116, 96, 109, 99], [127, 106, 145, 129, 114, 109, 143, 150, 111, 133, 146, 122]]
+Epoch 0: |                                                                             | 349/? [21:37<00:00,  0.27it/s, v_num=cys0]train step 350; scene = [['9d0bfbe5b7f98545'], ['06a16655c8e8ad9c']]; loss = 0.114770
+Epoch 0: |                                                                             | 350/? [21:40<00:00,  0.27it/s, v_num=cys0]context = [[33, 52, 53, 56, 58, 64, 73, 75], [100, 102, 115, 117, 130, 134, 139, 142], [223, 225, 240, 242, 255, 256, 263, 267]]target = [[41, 60, 53, 63, 72, 38, 37, 73], [122, 141, 116, 138, 110, 119, 118, 135], [260, 237, 261, 264, 238, 255, 243, 253]]
+Epoch 0: |                                                                             | 359/? [22:14<00:00,  0.27it/s, v_num=cys0]train step 360; scene = [['73b27f4f150327af'], ['169aaaf51ef3849c'], ['068a8406f1a383d8'], ['a9936b77895f33b3']]; loss = 0.079604
+Epoch 0: |                                                                             | 360/? [22:18<00:00,  0.27it/s, v_num=cys0]context = [[0, 14, 18, 23, 25, 32, 42, 45, 46, 49, 53, 55, 59, 60, 61, 74, 78, 80, 84, 85, 89, 90, 92, 97]]target = [[50, 81, 69, 64, 38, 86, 91, 16, 27, 33, 44, 90, 48, 23, 8, 79, 6, 39, 42, 36, 82, 78, 59, 54]]
+Epoch 0: |                                                                             | 369/? [22:51<00:00,  0.27it/s, v_num=cys0]train step 370; scene = [['8673faf0a9d48165'], ['99a0790d72e6c2af'], ['6cbbe9075b0d2138']]; loss = 0.061778
+Epoch 0: |                                                                             | 370/? [22:54<00:00,  0.27it/s, v_num=cys0]context = [[61, 63, 65, 72, 73, 74, 76, 89, 92, 94, 100, 101, 120, 124, 126, 127, 136, 140, 144, 145, 150, 152, 156, 158]]target = [[121, 100, 147, 98, 148, 143, 126, 63, 73, 141, 79, 119, 115, 106, 153, 101, 120, 71, 91, 62, 105, 84, 151, 145]]
+Epoch 0: |                                                                             | 379/? [23:28<00:00,  0.27it/s, v_num=cys0]train step 380; scene = [['656330f47c5df010'], ['6dfb89a98e14ca66']]; loss = 0.067288
+Epoch 0: |                                                                             | 380/? [23:32<00:00,  0.27it/s, v_num=cys0]context = [[210, 215, 223, 225, 231, 234, 235, 249], [28, 32, 35, 53, 54, 56, 57, 68], [149, 150, 154, 155, 174, 177, 181, 183]]target = [[241, 236, 218, 230, 223, 228, 239, 227], [55, 46, 44, 49, 30, 33, 60, 45], [157, 160, 161, 164, 170, 181, 168, 166]]
+Epoch 0: |                                                                             | 389/? [24:05<00:00,  0.27it/s, v_num=cys0]train step 390; scene = [['723f94d150ab09f2'], ['393cdfb7e832d285'], ['14900b71ac66b7bd'], ['452625cd6b071b87'], ['281599bbab3e73dd'], ['0a2b42e240751d33']]; loss = 0.072931
+Epoch 0: |                                                                             | 390/? [24:09<00:00,  0.27it/s, v_num=cys0]context = [[48, 52, 80], [14, 47, 60], [142, 159, 187], [78, 120, 121], [6, 24, 45], [47, 68, 90], [15, 19, 48], [9, 20, 52]]target = [[51, 50, 70], [45, 15, 56], [172, 148, 157], [120, 98, 109], [20, 33, 11], [83, 84, 73], [45, 30, 32], [16, 36, 45]]
+Epoch 0: |                                                                             | 399/? [24:41<00:00,  0.27it/s, v_num=cys0]train step 400; scene = [['4303746d8f23f16b'], ['0fe8246bb7e2fe40'], ['b7d77240852d6a52'], ['6e5505414fd63528'], ['44985936f68c3a36'], ['1550f1b4fff1f2a4'], ['cea3d842c3285c65'], ['b34bb5f53856d34f']]; loss = 0.097521
+Epoch 0: |                                                                             | 400/? [24:45<00:00,  0.27it/s, v_num=cys0]context = [[9, 20, 28, 31, 34, 53], [189, 196, 216, 223, 230, 232], [74, 77, 86, 88, 98, 118], [132, 145, 157, 168, 179, 181]]target = [[51, 32, 14, 50, 45, 34], [212, 226, 201, 205, 215, 222], [93, 113, 107, 94, 83, 79], [161, 139, 177, 151, 156, 172]]
+[2026-02-25 07:57:27,231][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 409/? [25:18<00:00,  0.27it/s, v_num=cys0]train step 410; scene = [['144e1ec915e46d29'], ['b290b6a0afa1dac7'], ['b3d84dba6581c3d9']]; loss = 0.060998
+Epoch 0: |                                                                             | 410/? [25:22<00:00,  0.27it/s, v_num=cys0]context = [[26, 40, 41, 50, 55, 58], [159, 174, 179, 192, 198, 201], [2, 3, 5, 7, 24, 40], [187, 204, 205, 208, 217, 221]]target = [[38, 32, 44, 52, 42, 27], [198, 191, 196, 162, 163, 169], [33, 17, 23, 37, 15, 39], [214, 211, 195, 220, 210, 192]]
+Epoch 0: |                                                                             | 419/? [25:55<00:00,  0.27it/s, v_num=cys0]train step 420; scene = [['a1dff9c50d92dc9c']]; loss = 0.056380
+Epoch 0: |                                                                             | 420/? [25:59<00:00,  0.27it/s, v_num=cys0]context = [[11, 13, 14, 26, 27, 41, 46, 51], [151, 155, 161, 167, 172, 176, 191, 195], [49, 52, 55, 57, 59, 74, 79, 90]]target = [[25, 44, 17, 33, 37, 24, 38, 40], [193, 191, 194, 162, 158, 176, 163, 171], [87, 85, 67, 58, 86, 68, 83, 78]]
+Epoch 0: |                                                                             | 429/? [26:32<00:00,  0.27it/s, v_num=cys0]train step 430; scene = [['36664e22fd10a141'], ['0474328f4cefd619']]; loss = 0.055051
+Epoch 0: |                                                                             | 430/? [26:35<00:00,  0.27it/s, v_num=cys0]context = [[123, 130, 131, 132, 136, 154, 169, 173], [35, 36, 47, 49, 65, 67, 76, 86], [33, 34, 39, 41, 59, 64, 65, 73]]target = [[145, 147, 167, 130, 146, 170, 138, 137], [59, 38, 49, 42, 70, 76, 39, 40], [56, 44, 36, 35, 38, 45, 48, 58]]
+Epoch 0: |                                                                             | 439/? [27:09<00:00,  0.27it/s, v_num=cys0]train step 440; scene = [['342099a48847f4f6'], ['5ad0327426e3718b'], ['c25b314716aa6b10'], ['c91e2b5399b14430'], ['e1d9ade67e615bd8'], ['46df912c9748215b']]; loss = 0.069442
+Epoch 0: |                                                                             | 440/? [27:13<00:00,  0.27it/s, v_num=cys0]context = [[0, 33], [10, 57], [36, 82], [46, 79], [62, 99], [33, 79], [11, 52], [1, 34], [0, 37], [26, 62], [18, 67], [96, 138]]target = [[22, 18], [37, 23], [44, 54], [69, 50], [68, 69], [69, 56], [49, 41], [27, 30], [31, 17], [42, 32], [59, 54], [108, 115]]
+Epoch 0: |                                                                             | 449/? [27:47<00:00,  0.27it/s, v_num=cys0]train step 450; scene = [['e19c6facac1c9624'], ['5244830b7357365b'], ['b80c2522b1070e2f'], ['6ea0ff32c8ea695c'], ['2f311b2bbbeb5940'], ['3f7992e72a096099']]; loss = 0.072574
+Epoch 0: |                                                                             | 450/? [27:50<00:00,  0.27it/s, v_num=cys0]context = [[41, 42, 44, 49, 57, 66, 68, 73, 75, 76, 83, 91], [4, 5, 15, 24, 32, 35, 37, 38, 45, 47, 49, 56]]target = [[79, 77, 90, 46, 81, 83, 56, 68, 89, 70, 84, 57], [7, 8, 40, 37, 32, 10, 25, 35, 23, 43, 48, 52]]
+Epoch 0: |                                                                             | 459/? [28:24<00:00,  0.27it/s, v_num=cys0]train step 460; scene = [['46fb6702ed1b9967'], ['bdc3f978b0d3aa8f']]; loss = 0.057634
+Epoch 0: |                                                                             | 460/? [28:28<00:00,  0.27it/s, v_num=cys0]context = [[65, 66, 69, 78, 84, 85, 90, 91, 99, 103, 112, 116], [39, 40, 52, 63, 64, 68, 72, 73, 83, 87, 90, 92]]target = [[99, 86, 101, 76, 84, 77, 105, 66, 89, 75, 100, 106], [84, 61, 68, 82, 59, 47, 51, 71, 85, 56, 57, 50]]
+Epoch 0: |                                                                             | 469/? [29:00<00:00,  0.27it/s, v_num=cys0]train step 470; scene = [['2c88995e05a17d17'], ['2b1f47da224557a3'], ['62216d162b71b5b4'], ['61d39a97cb69d99f'], ['42000d5a83b48ee4'], ['cc8480640599f9f3']]; loss = 0.074535
+Epoch 0: |                                                                             | 470/? [29:03<00:00,  0.27it/s, v_num=cys0]context = [[1, 6, 15, 26, 28, 41, 46, 52], [0, 1, 3, 5, 14, 26, 32, 37], [52, 56, 58, 79, 87, 88, 93, 97]]target = [[24, 22, 32, 14, 42, 21, 12, 37], [14, 10, 11, 25, 12, 34, 7, 32], [62, 60, 80, 87, 63, 55, 78, 88]]
+Epoch 0: |                                                                             | 479/? [29:36<00:00,  0.27it/s, v_num=cys0]train step 480; scene = [['2e3bb7fb33e1ed30'], ['7460f503eb18fa6a'], ['bde49071d2088850'], ['e80016be3043dfa4']]; loss = 0.079864
+Epoch 0: |                                                                             | 480/? [29:40<00:00,  0.27it/s, v_num=cys0]context = [[9, 15, 37, 47], [0, 5, 7, 35], [101, 113, 134, 153], [64, 88, 93, 107], [18, 44, 56, 62], [31, 46, 58, 75]]target = [[45, 27, 28, 18], [10, 8, 14, 3], [128, 143, 110, 129], [87, 84, 90, 94], [29, 31, 40, 36], [52, 72, 33, 56]]
+Epoch 0: |                                                                             | 489/? [30:14<00:00,  0.27it/s, v_num=cys0]train step 490; scene = [['83085493f4bc18d2']]; loss = 0.084496
+Epoch 0: |                                                                             | 490/? [30:17<00:00,  0.27it/s, v_num=cys0]context = [[0, 3, 7, 12, 38, 40, 45, 48], [25, 28, 57, 62, 69, 70, 76, 78], [25, 34, 37, 39, 41, 53, 65, 67]]target = [[15, 7, 39, 47, 42, 26, 25, 5], [63, 40, 59, 31, 75, 34, 47, 39], [58, 35, 53, 33, 65, 44, 38, 31]]
+Epoch 0: |                                                                             | 499/? [30:51<00:00,  0.27it/s, v_num=cys0]train step 500; scene = [['1241bcb5732a9502'], ['d33a9e90e1416efb']]; loss = 0.042771
+Epoch 0: |                                                                             | 500/? [30:54<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 500; scene = ['73d6f935f31b3fd4'];
+target intrinsic:  tensor(0.8576, device='cuda:0') tensor(0.8579, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8827, device='cuda:0') tensor(0.8867, device='cuda:0')
+[2026-02-25 08:03:33,811][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 500/? [30:56<00:00,  0.27it/s, v_num=cys0]context = [[1, 21, 23, 29, 39, 48, 49, 52], [17, 18, 20, 28, 34, 39, 45, 51], [99, 103, 106, 110, 116, 136, 138, 140]]target = [[4, 46, 41, 43, 40, 37, 39, 25], [30, 48, 22, 42, 44, 31, 38, 37], [112, 108, 133, 116, 111, 125, 123, 127]]
+Epoch 0: |                                                                             | 509/? [31:28<00:00,  0.27it/s, v_num=cys0]train step 510; scene = [['0eed4548041bea8e'], ['277a96ce456580f4']]; loss = 0.054919
+Epoch 0: |                                                                             | 510/? [31:31<00:00,  0.27it/s, v_num=cys0]context = [[3, 13, 21, 24, 27, 28, 30, 37, 42, 44, 46, 55, 62, 65, 69, 71, 73, 80, 83, 84, 89, 95, 97, 100]]target = [[51, 81, 54, 89, 55, 45, 10, 19, 21, 68, 48, 67, 33, 69, 8, 25, 64, 83, 37, 29, 24, 82, 99, 46]]
+Epoch 0: |                                                                             | 519/? [32:05<00:00,  0.27it/s, v_num=cys0]train step 520; scene = [['625e3aa0ff734714'], ['395802511d26f32e'], ['39343936591c28de']]; loss = 0.084008
+Epoch 0: |                                                                             | 520/? [32:09<00:00,  0.27it/s, v_num=cys0]context = [[13, 16, 17, 22, 25, 30, 35, 39, 53, 56, 62, 68, 78, 87, 90, 92, 93, 95, 99, 100, 103, 106, 107, 110]]target = [[39, 96, 58, 28, 16, 59, 44, 17, 26, 83, 103, 31, 57, 35, 107, 51, 27, 77, 46, 30, 100, 91, 93, 97]]
+Epoch 0: |                                                                             | 529/? [32:42<00:00,  0.27it/s, v_num=cys0]train step 530; scene = [['0c199c575b699444'], ['70d878da47f984e4'], ['15f77c76ea744f99'], ['e54b5eec8cc47776'], ['1969ed97e68d83d9'], ['c7cf9b63dc3e5830'], ['bcef3076b93012b1'], ['ab2680bf91942e23']]; loss = 0.084372
+Epoch 0: |                                                                             | 530/? [32:46<00:00,  0.27it/s, v_num=cys0]context = [[78, 84, 85, 92, 124, 127], [14, 15, 24, 52, 64, 68], [1, 6, 16, 20, 22, 40], [33, 51, 61, 65, 66, 70]]target = [[104, 103, 94, 96, 88, 82], [15, 53, 58, 47, 56, 60], [31, 10, 18, 26, 23, 24], [37, 69, 62, 65, 49, 64]]
+Epoch 0: |                                                                             | 539/? [33:19<00:00,  0.27it/s, v_num=cys0]train step 540; scene = [['a071d9276f6a9272']]; loss = 0.068220
+Epoch 0: |                                                                             | 540/? [33:23<00:00,  0.27it/s, v_num=cys0]context = [[35, 46, 47, 49, 52, 54, 55, 56, 57, 65, 66, 85], [51, 54, 55, 59, 68, 77, 82, 86, 87, 93, 105, 106]]target = [[69, 68, 75, 48, 44, 79, 53, 74, 60, 81, 72, 71], [77, 73, 88, 59, 67, 61, 102, 62, 93, 75, 95, 105]]
+Epoch 0: |                                                                             | 549/? [33:56<00:00,  0.27it/s, v_num=cys0]train step 550; scene = [['836250796ea45b6c']]; loss = 0.087670
+Epoch 0: |                                                                             | 550/? [34:00<00:00,  0.27it/s, v_num=cys0]context = [[30, 34, 39, 45, 47, 79, 80, 83], [16, 21, 22, 24, 33, 42, 51, 60], [115, 128, 129, 138, 141, 150, 156, 163]]target = [[79, 51, 62, 77, 46, 54, 33, 49], [27, 59, 20, 41, 21, 48, 58, 30], [133, 125, 124, 159, 122, 129, 155, 119]]
+Epoch 0: |                                                                             | 559/? [34:34<00:00,  0.27it/s, v_num=cys0]train step 560; scene = [['d70ca840b3c5aec9'], ['65c3f29c43dd1e63'], ['d3917d0a1eda2a1f'], ['5c83dfc8f9ab44fa']]; loss = 0.061901
+Epoch 0: |                                                                             | 560/? [34:37<00:00,  0.27it/s, v_num=cys0]context = [[62, 66, 67, 68, 71, 87, 88, 92, 100, 108, 110, 122], [153, 160, 165, 168, 171, 177, 178, 181, 183, 185, 200, 205]]target = [[95, 104, 83, 68, 115, 81, 74, 106, 119, 79, 90, 72], [204, 165, 196, 185, 197, 198, 195, 187, 155, 188, 154, 201]]
+Epoch 0: |                                                                             | 569/? [35:11<00:00,  0.27it/s, v_num=cys0]train step 570; scene = [['9d8ddcdbe1f7ac42'], ['721df0f45094ca34'], ['fdbfe35f5940d3ad']]; loss = 0.047884
+Epoch 0: |                                                                             | 570/? [35:15<00:00,  0.27it/s, v_num=cys0]context = [[23, 34, 41, 42, 48, 54, 56, 59, 60, 63, 66, 71, 74, 79, 88, 90, 92, 93, 95, 97, 110, 111, 119, 120]]target = [[30, 27, 28, 53, 31, 75, 85, 54, 77, 111, 76, 25, 44, 52, 33, 41, 69, 89, 73, 68, 26, 93, 83, 119]]
+Epoch 0: |                                                                             | 579/? [35:49<00:00,  0.27it/s, v_num=cys0]train step 580; scene = [['88a0267e41b851f0'], ['df71fbb70b19cbc3'], ['1c713c10ecf5a0c9']]; loss = 0.058984
+Epoch 0: |                                                                             | 580/? [35:51<00:00,  0.27it/s, v_num=cys0]context = [[9, 22, 36, 39, 43, 68], [70, 85, 95, 98, 116, 121], [10, 28, 45, 47, 50, 57], [132, 144, 154, 158, 168, 188]]target = [[67, 45, 11, 46, 50, 30], [114, 94, 71, 104, 117, 76], [22, 12, 24, 51, 50, 13], [180, 169, 183, 173, 182, 171]]
+Epoch 0: |                                                                             | 589/? [36:25<00:00,  0.27it/s, v_num=cys0]train step 590; scene = [['3f732b63cdd0729e'], ['9be3165beb073d95'], ['42a6c835ff830674'], ['f928d960cbfae15a'], ['140b10a4f6bb5aa5'], ['cc8e19c8ad1846f4']]; loss = 0.093385
+Epoch 0: |                                                                             | 590/? [36:29<00:00,  0.27it/s, v_num=cys0]context = [[198, 200, 201, 204, 209, 210, 224, 234], [6, 14, 18, 23, 25, 27, 58, 59], [13, 29, 40, 51, 55, 66, 70, 71]]target = [[203, 208, 227, 217, 222, 223, 200, 211], [11, 46, 45, 19, 20, 18, 47, 7], [16, 51, 40, 15, 55, 44, 23, 61]]
+Epoch 0: |                                                                             | 599/? [37:02<00:00,  0.27it/s, v_num=cys0]train step 600; scene = [['cb734fdc69e9900e']]; loss = 0.064087
+Epoch 0: |                                                                             | 600/? [37:06<00:00,  0.27it/s, v_num=cys0]context = [[10, 16, 24, 31, 33, 34, 36, 41, 49, 59, 63, 67], [160, 162, 167, 172, 173, 180, 183, 187, 189, 205, 212, 214]]target = [[32, 54, 42, 14, 39, 60, 66, 33, 62, 22, 47, 49], [182, 199, 163, 179, 203, 200, 174, 209, 164, 172, 194, 167]]
+[2026-02-25 08:09:48,816][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 609/? [37:41<00:00,  0.27it/s, v_num=cys0]train step 610; scene = [['ed9409fa128e193b'], ['a5c03b0c5fb7203e']]; loss = 0.040113
+Epoch 0: |                                                                             | 610/? [37:45<00:00,  0.27it/s, v_num=cys0]context = [[36, 37, 39, 50, 51, 52, 57, 59, 72, 73, 86, 94], [45, 46, 54, 57, 66, 75, 77, 81, 83, 89, 97, 101]]target = [[47, 37, 89, 79, 92, 78, 81, 90, 38, 56, 84, 69], [84, 91, 85, 69, 79, 53, 55, 48, 87, 49, 92, 63]]
+Epoch 0: |                                                                             | 619/? [38:18<00:00,  0.27it/s, v_num=cys0]train step 620; scene = [['7898a828b7203ca4'], ['9c269fce78f0dd27'], ['e1e317857deb7afc'], ['30124191dafb3383'], ['c39f1a9a73797efe'], ['a640a55439a43108']]; loss = 0.058145
+Epoch 0: |                                                                             | 620/? [38:22<00:00,  0.27it/s, v_num=cys0]context = [[2, 15, 16, 21, 26, 29, 47, 56], [6, 12, 15, 37, 45, 47, 50, 68], [9, 18, 23, 26, 39, 48, 50, 60]]target = [[38, 7, 32, 36, 16, 21, 53, 22], [50, 67, 28, 43, 15, 41, 44, 23], [12, 54, 16, 24, 58, 32, 30, 47]]
+Epoch 0: |                                                                             | 629/? [38:55<00:00,  0.27it/s, v_num=cys0]train step 630; scene = [['dd3bbf1f7f832e83'], ['0ff7277275e55096'], ['5f45c360d76a3b12']]; loss = 0.051831
+Epoch 0: |                                                                             | 630/? [38:59<00:00,  0.27it/s, v_num=cys0]context = [[29, 42, 44, 45, 76, 81, 87, 93], [77, 84, 94, 96, 100, 102, 121, 125], [67, 68, 70, 89, 90, 91, 96, 107]]target = [[83, 55, 36, 89, 45, 62, 90, 63], [106, 92, 114, 99, 117, 82, 119, 107], [99, 68, 75, 92, 79, 85, 101, 90]]
+Epoch 0: |                                                                             | 639/? [39:33<00:00,  0.27it/s, v_num=cys0]train step 640; scene = [['867edbda9bb8ef59'], ['1d83764e77e159d8'], ['e318dafa4071cef9'], ['169f09c33ee35289']]; loss = 0.093050
+Epoch 0: |                                                                             | 640/? [39:36<00:00,  0.27it/s, v_num=cys0]context = [[52, 57, 67, 75, 77, 80, 85, 86, 99, 105, 112, 117], [6, 11, 22, 23, 32, 39, 42, 47, 48, 53, 63, 64]]target = [[114, 60, 111, 110, 57, 69, 68, 80, 62, 99, 77, 90], [27, 15, 46, 56, 51, 60, 41, 63, 49, 16, 9, 43]]
+Epoch 0: |                                                                             | 649/? [40:09<00:00,  0.27it/s, v_num=cys0]train step 650; scene = [['23668135f32e0126'], ['daca15248046e480'], ['174ebd189316bd92']]; loss = 0.053341
+Epoch 0: |                                                                             | 650/? [40:13<00:00,  0.27it/s, v_num=cys0]context = [[6, 23, 25, 26, 28, 32, 37, 38, 39, 40, 50, 57], [0, 2, 3, 11, 17, 19, 24, 25, 38, 40, 45, 53]]target = [[15, 47, 23, 31, 12, 30, 20, 55, 33, 11, 22, 10], [51, 9, 13, 34, 52, 30, 26, 45, 38, 27, 11, 14]]
+Epoch 0: |                                                                             | 659/? [40:46<00:00,  0.27it/s, v_num=cys0]train step 660; scene = [['60499200285c9abe']]; loss = 0.044716
+Epoch 0: |                                                                             | 660/? [40:49<00:00,  0.27it/s, v_num=cys0]context = [[57, 65, 76, 86, 93, 99, 100, 122], [31, 34, 42, 47, 60, 80, 81, 83], [13, 16, 19, 30, 33, 49, 50, 55]]target = [[85, 75, 88, 114, 107, 70, 116, 79], [82, 35, 67, 56, 68, 61, 74, 58], [54, 44, 23, 36, 31, 39, 45, 28]]
+Epoch 0: |                                                                             | 669/? [41:22<00:00,  0.27it/s, v_num=cys0]train step 670; scene = [['7665ff641f430aa5']]; loss = 0.040170
+Epoch 0: |                                                                             | 670/? [41:26<00:00,  0.27it/s, v_num=cys0]context = [[50, 57, 61, 69, 82, 87, 98, 101], [26, 33, 44, 47, 53, 56, 62, 66], [11, 33, 39, 48, 56, 67, 75, 78]]target = [[67, 96, 71, 75, 73, 57, 55, 89], [65, 39, 63, 59, 60, 47, 55, 41], [18, 56, 21, 64, 32, 59, 33, 25]]
+Epoch 0: |                                                                             | 679/? [42:00<00:00,  0.27it/s, v_num=cys0]train step 680; scene = [['43c939b11c5fed4a']]; loss = 0.081909
+Epoch 0: |                                                                             | 680/? [42:03<00:00,  0.27it/s, v_num=cys0]context = [[47, 54, 60, 64, 79, 80, 83, 93], [12, 13, 21, 30, 34, 45, 53, 55], [57, 75, 78, 82, 96, 103, 113, 117]]target = [[80, 62, 75, 54, 57, 92, 53, 58], [38, 32, 15, 44, 42, 51, 27, 39], [81, 87, 106, 94, 99, 103, 73, 79]]
+Epoch 0: |                                                                             | 689/? [42:36<00:00,  0.27it/s, v_num=cys0]train step 690; scene = [['1848b8b363d0d2b9'], ['afe6b05d0554a880']]; loss = 0.062938
+Epoch 0: |                                                                             | 690/? [42:40<00:00,  0.27it/s, v_num=cys0]context = [[12, 18, 20, 27, 36, 37, 38, 43, 48, 49, 52, 58, 59, 62, 67, 81, 83, 86, 93, 95, 101, 104, 108, 109]]target = [[34, 24, 43, 31, 87, 30, 51, 54, 52, 94, 86, 21, 44, 97, 61, 95, 38, 60, 49, 73, 41, 19, 65, 67]]
+Epoch 0: |                                                                             | 699/? [43:13<00:00,  0.27it/s, v_num=cys0]train step 700; scene = [['674ef9fb9cf20f9f'], ['8624ee0839cb6e4c'], ['caed302f388b799f']]; loss = 0.056062
+Epoch 0: |                                                                             | 700/? [43:15<00:00,  0.27it/s, v_num=cys0]context = [[31, 32, 41, 46, 53, 54, 55, 57, 65, 68, 73, 74, 80, 85, 100, 105, 108, 109, 113, 114, 116, 118, 126, 128]]target = [[117, 52, 85, 57, 37, 45, 78, 100, 125, 35, 113, 66, 105, 103, 61, 83, 88, 40, 116, 60, 79, 32, 102, 107]]
+Epoch 0: |                                                                             | 709/? [43:49<00:00,  0.27it/s, v_num=cys0]train step 710; scene = [['db6cd90de8fee2ff'], ['7a20ba81fb778529'], ['970350268b239272']]; loss = 0.054809
+Epoch 0: |                                                                             | 710/? [43:53<00:00,  0.27it/s, v_num=cys0]context = [[3, 12, 16, 31, 41, 45, 49, 53, 54, 55, 67, 69], [208, 211, 217, 221, 222, 227, 230, 231, 239, 247, 252, 267]]target = [[38, 34, 68, 25, 44, 63, 48, 7, 42, 6, 64, 28], [228, 253, 259, 266, 244, 238, 249, 234, 232, 241, 220, 213]]
+Epoch 0: |                                                                             | 719/? [44:26<00:00,  0.27it/s, v_num=cys0]train step 720; scene = [['f63d2df8871ce70c'], ['0fdeda15097ed4a4']]; loss = 0.048772
+Epoch 0: |                                                                             | 720/? [44:30<00:00,  0.27it/s, v_num=cys0]context = [[55, 57, 71, 72, 75, 85, 90, 92, 98, 99, 102, 108, 112, 116, 125, 126, 130, 131, 135, 141, 144, 145, 146, 152]]target = [[61, 145, 86, 66, 62, 119, 100, 139, 105, 125, 58, 101, 140, 132, 118, 128, 65, 141, 151, 78, 104, 107, 150, 138]]
+Epoch 0: |                                                                             | 729/? [45:04<00:00,  0.27it/s, v_num=cys0]train step 730; scene = [['232abb354c423e81'], ['d34926c73ae1277e']]; loss = 0.038772
+Epoch 0: |                                                                             | 730/? [45:08<00:00,  0.27it/s, v_num=cys0]context = [[45, 50, 54, 58, 66, 70, 72, 78, 85, 88, 94, 105, 109, 110, 111, 117, 120, 122, 126, 127, 128, 133, 135, 142]]target = [[95, 96, 50, 114, 135, 124, 104, 100, 49, 119, 139, 62, 92, 123, 58, 46, 57, 112, 116, 90, 54, 101, 85, 81]]
+Epoch 0: |                                                                             | 739/? [45:41<00:00,  0.27it/s, v_num=cys0]train step 740; scene = [['19f7966006ad778d'], ['dde0212418df7ca9'], ['ad75e36b74f6b033'], ['ea97e5ae55e56208'], ['9d29b0289133ab4e'], ['282938f90821bdef']]; loss = 0.107187
+Epoch 0: |                                                                             | 740/? [45:44<00:00,  0.27it/s, v_num=cys0]context = [[8, 13, 19, 22, 30, 31, 36, 45, 46, 48, 54, 57, 58, 59, 60, 61, 65, 73, 77, 83, 86, 88, 93, 105]]target = [[46, 101, 90, 61, 31, 23, 37, 95, 18, 67, 32, 100, 93, 35, 89, 45, 10, 70, 60, 97, 85, 81, 66, 79]]
+Epoch 0: |                                                                             | 749/? [46:18<00:00,  0.27it/s, v_num=cys0]train step 750; scene = [['f85921f42c5d98d7'], ['a95dacbd3ea3db36']]; loss = 0.045763
+Epoch 0: |                                                                             | 750/? [46:22<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 750; scene = ['91fda69e1cda4602'];
+target intrinsic:  tensor(0.8937, device='cuda:0') tensor(0.8939, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9339, device='cuda:0') tensor(0.9340, device='cuda:0')
+[2026-02-25 08:19:01,130][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 750/? [46:23<00:00,  0.27it/s, v_num=cys0]context = [[1, 38, 46, 49, 52, 59], [213, 218, 227, 233, 252, 257], [99, 105, 109, 124, 125, 153], [8, 17, 20, 37, 62, 65]]target = [[54, 19, 39, 57, 7, 4], [252, 219, 223, 216, 248, 237], [148, 136, 102, 103, 149, 132], [12, 52, 13, 44, 63, 39]]
+Epoch 0: |                                                                             | 759/? [46:56<00:00,  0.27it/s, v_num=cys0]train step 760; scene = [['75617c97bff1e873'], ['ff02f88545dfa566']]; loss = 0.038173
+Epoch 0: |                                                                             | 760/? [46:59<00:00,  0.27it/s, v_num=cys0]context = [[71, 73, 111], [2, 14, 56], [29, 82, 88], [120, 144, 168], [181, 214, 250], [0, 10, 71], [12, 53, 54], [14, 27, 87]]target = [[101, 74, 84], [43, 51, 38], [82, 55, 72], [156, 139, 147], [198, 214, 207], [22, 5, 40], [36, 13, 39], [68, 35, 70]]
+Epoch 0: |                                                                             | 769/? [47:33<00:00,  0.27it/s, v_num=cys0]train step 770; scene = [['62b0d4ee613af70f'], ['f7926eb1096de201'], ['c63b37ec347f0d0e'], ['b43d9f7c70f5caa0']]; loss = 0.073092
+Epoch 0: |                                                                             | 770/? [47:37<00:00,  0.27it/s, v_num=cys0]context = [[13, 29, 30, 31, 32, 34, 41, 42, 46, 52, 63, 66], [146, 149, 157, 166, 177, 180, 181, 193, 197, 204, 211, 213]]target = [[35, 55, 52, 59, 24, 51, 63, 18, 42, 44, 33, 61], [183, 158, 171, 175, 163, 196, 177, 181, 210, 167, 187, 174]]
+Epoch 0: |                                                                             | 779/? [48:10<00:00,  0.27it/s, v_num=cys0]train step 780; scene = [['b41f4db8b8a42a71']]; loss = 0.075497
+Epoch 0: |                                                                             | 780/? [48:14<00:00,  0.27it/s, v_num=cys0]context = [[43, 50, 53, 55, 57, 58, 59, 60, 68, 81, 94, 96, 102, 106, 108, 110, 112, 121, 125, 126, 130, 131, 136, 140]]target = [[74, 122, 93, 98, 70, 84, 49, 136, 77, 117, 135, 138, 123, 89, 119, 45, 129, 105, 50, 58, 63, 103, 82, 121]]
+Epoch 0: |                                                                             | 789/? [48:47<00:00,  0.27it/s, v_num=cys0]train step 790; scene = [['d79666d294813d8e']]; loss = 0.157637
+Epoch 0: |                                                                             | 790/? [48:51<00:00,  0.27it/s, v_num=cys0]context = [[59, 79, 88, 93, 100, 107], [3, 16, 19, 31, 46, 51], [41, 49, 54, 56, 72, 83], [226, 229, 235, 244, 255, 272]]target = [[103, 98, 81, 75, 74, 72], [48, 44, 20, 31, 15, 46], [64, 82, 70, 50, 47, 60], [251, 249, 271, 250, 259, 260]]
+Epoch 0: |                                                                             | 799/? [49:24<00:00,  0.27it/s, v_num=cys0]train step 800; scene = [['cb797cd30542e55c']]; loss = 0.057544
+Epoch 0: |                                                                             | 800/? [49:28<00:00,  0.27it/s, v_num=cys0]context = [[10, 19, 20, 39, 42, 55], [1, 6, 11, 20, 30, 42], [0, 3, 4, 19, 42, 45], [21, 35, 52, 82, 85, 88]]target = [[38, 39, 26, 21, 37, 40], [11, 40, 18, 27, 3, 24], [43, 33, 8, 24, 5, 19], [83, 71, 66, 32, 63, 47]]
+[2026-02-25 08:22:10,768][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 809/? [50:02<00:00,  0.27it/s, v_num=cys0]train step 810; scene = [['9bd08fc9288bef8b']]; loss = 0.061432
+Epoch 0: |                                                                             | 810/? [50:05<00:00,  0.27it/s, v_num=cys0]context = [[5, 6, 14, 17, 19, 23, 30, 37, 40, 49, 53, 55], [3, 6, 15, 17, 27, 34, 38, 40, 42, 46, 56, 65]]target = [[16, 8, 48, 23, 44, 49, 10, 7, 18, 50, 22, 15], [34, 18, 9, 42, 31, 38, 45, 26, 43, 6, 49, 44]]
+Epoch 0: |                                                                             | 819/? [50:39<00:00,  0.27it/s, v_num=cys0]train step 820; scene = [['49952f737be91dd2']]; loss = 0.076896
+Epoch 0: |                                                                             | 820/? [50:43<00:00,  0.27it/s, v_num=cys0]context = [[22, 23, 26, 28, 31, 43, 51, 53, 57, 65, 67, 68, 71, 76, 77, 78, 80, 84, 88, 90, 107, 112, 116, 119]]target = [[35, 45, 31, 72, 24, 71, 89, 54, 25, 61, 44, 97, 56, 47, 82, 87, 50, 110, 108, 75, 103, 67, 86, 117]]
+Epoch 0: |                                                                             | 829/? [51:14<00:00,  0.27it/s, v_num=cys0]train step 830; scene = [['2f9c9d1b56eb7f75'], ['1c392cc98b3a7642']]; loss = 0.066223
+Epoch 0: |                                                                             | 830/? [51:18<00:00,  0.27it/s, v_num=cys0]context = [[42, 67, 90, 119], [18, 54, 72, 90], [43, 67, 73, 102], [2, 22, 40, 43], [23, 40, 52, 67], [19, 37, 75, 78]]target = [[86, 112, 87, 94], [59, 40, 20, 47], [58, 97, 66, 77], [34, 19, 32, 17], [62, 37, 35, 54], [54, 22, 31, 50]]
+Epoch 0: |                                                                             | 839/? [51:51<00:00,  0.27it/s, v_num=cys0]train step 840; scene = [['c51c7bc0c8151abb'], ['a0d16e79ab441c4f']]; loss = 0.143271
+Epoch 0: |                                                                             | 840/? [51:55<00:00,  0.27it/s, v_num=cys0]context = [[1, 18, 22, 27, 29, 35, 38, 49], [3, 12, 30, 36, 43, 47, 54, 55], [17, 36, 55, 65, 72, 77, 86, 88]]target = [[5, 36, 48, 11, 2, 39, 35, 37], [6, 54, 43, 18, 30, 15, 10, 11], [52, 59, 56, 77, 25, 68, 22, 27]]
+Epoch 0: |                                                                             | 849/? [52:28<00:00,  0.27it/s, v_num=cys0]train step 850; scene = [['818df63ddd1cf294'], ['3002f9cbe7f00e6c'], ['8ec42c5dfea6823b'], ['64ae75e57c6aa0a4']]; loss = 0.127301
+Epoch 0: |                                                                             | 850/? [52:32<00:00,  0.27it/s, v_num=cys0]context = [[31, 75, 86], [20, 22, 73], [7, 55, 63], [3, 62, 68], [10, 43, 54], [52, 92, 111], [0, 34, 42], [51, 90, 102]]target = [[40, 57, 36], [39, 67, 34], [12, 9, 52], [29, 12, 39], [49, 13, 17], [106, 93, 85], [39, 20, 9], [79, 60, 53]]
+Epoch 0: |                                                                             | 859/? [53:05<00:00,  0.27it/s, v_num=cys0]train step 860; scene = [['eb0aa1a4fb58c50c']]; loss = 0.045324
+Epoch 0: |                                                                             | 860/? [53:09<00:00,  0.27it/s, v_num=cys0]context = [[147, 155, 165, 189, 194, 195], [9, 15, 44, 47, 49, 82], [5, 22, 23, 51, 61, 64], [97, 98, 140, 149, 160, 170]]target = [[194, 192, 155, 151, 174, 158], [22, 27, 54, 18, 81, 31], [25, 27, 29, 23, 6, 59], [137, 162, 98, 111, 152, 164]]
+Epoch 0: |                                                                             | 869/? [53:42<00:00,  0.27it/s, v_num=cys0]train step 870; scene = [['c357fbd8aca05570']]; loss = 0.051791
+Epoch 0: |                                                                             | 870/? [53:46<00:00,  0.27it/s, v_num=cys0]context = [[44, 45, 82, 100], [70, 93, 96, 147], [144, 176, 184, 205], [0, 51, 65, 67], [105, 106, 130, 152], [5, 6, 55, 60]]target = [[81, 66, 67, 97], [116, 118, 137, 91], [200, 193, 197, 176], [11, 13, 22, 32], [120, 117, 127, 130], [17, 14, 58, 27]]
+Epoch 0: |                                                                             | 879/? [54:20<00:00,  0.27it/s, v_num=cys0]train step 880; scene = [['c672fa3960b73528'], ['a60e4127f167ac93'], ['a9739ec3a34012af']]; loss = 0.059302
+Epoch 0: |                                                                             | 880/? [54:24<00:00,  0.27it/s, v_num=cys0]context = [[15, 85, 91], [0, 68, 74], [73, 85, 121], [4, 53, 80], [15, 36, 63], [42, 81, 92], [0, 45, 49], [65, 97, 132]]target = [[49, 21, 35], [47, 44, 69], [119, 89, 106], [57, 76, 6], [28, 32, 55], [62, 43, 49], [18, 44, 24], [90, 104, 71]]
+Epoch 0: |                                                                             | 889/? [54:56<00:00,  0.27it/s, v_num=cys0]train step 890; scene = [['40a3f4f9389dd20c'], ['b14ec6f019932d8d'], ['4b9ed7532c875dab'], ['10c36bd5ef5f5a6b'], ['bc9a64096787007d'], ['d58a26d24f2776b2'], ['46f2228076e6f3f7'], ['399668567ff33ad7']]; loss = 0.077907
+Epoch 0: |                                                                             | 890/? [54:59<00:00,  0.27it/s, v_num=cys0]context = [[32, 34, 35, 37, 39, 44, 52, 59, 73, 84, 87, 93], [20, 21, 22, 34, 43, 47, 52, 57, 64, 69, 72, 96]]target = [[87, 80, 41, 73, 35, 37, 86, 66, 72, 39, 50, 83], [28, 22, 50, 33, 26, 79, 75, 73, 29, 77, 60, 37]]
+Epoch 0: |                                                                             | 899/? [55:33<00:00,  0.27it/s, v_num=cys0]train step 900; scene = [['711ade236bebd618']]; loss = 0.059360
+Epoch 0: |                                                                             | 900/? [55:36<00:00,  0.27it/s, v_num=cys0]context = [[6, 7, 14, 15, 17, 19, 39, 43, 44, 45, 47, 48, 49, 53, 57, 61, 67, 69, 72, 73, 78, 90, 92, 103]]target = [[18, 19, 38, 66, 14, 96, 50, 55, 41, 57, 45, 83, 12, 100, 24, 53, 85, 40, 54, 58, 20, 44, 86, 95]]
+Epoch 0: |                                                                             | 909/? [56:10<00:00,  0.27it/s, v_num=cys0]train step 910; scene = [['6b495ce9634d2fbb']]; loss = 0.085765
+Epoch 0: |                                                                             | 910/? [56:13<00:00,  0.27it/s, v_num=cys0]context = [[59, 61, 62, 63, 73, 74, 77, 90, 100, 105, 111, 112], [8, 15, 18, 21, 26, 35, 37, 40, 41, 56, 68, 70]]target = [[76, 98, 64, 84, 69, 87, 100, 99, 67, 108, 77, 78], [22, 24, 41, 15, 45, 44, 34, 28, 13, 39, 11, 55]]
+Epoch 0: |                                                                             | 919/? [56:46<00:00,  0.27it/s, v_num=cys0]train step 920; scene = [['013264a550df794f'], ['4203a06d618eeb97']]; loss = 0.105689
+Epoch 0: |                                                                             | 920/? [56:50<00:00,  0.27it/s, v_num=cys0]context = [[14, 16, 18, 28, 30, 37, 45, 49, 50, 61, 63, 65], [97, 98, 114, 115, 121, 134, 137, 140, 143, 144, 169, 179]]target = [[38, 62, 50, 58, 42, 55, 57, 46, 43, 49, 56, 40], [158, 178, 130, 108, 145, 120, 168, 164, 156, 177, 125, 111]]
+Epoch 0: |                                                                             | 929/? [57:22<00:00,  0.27it/s, v_num=cys0]train step 930; scene = [['5747f1d12ad10026'], ['48c9bb29482ddf76'], ['a45bddb856f554a1']]; loss = 0.056401
+Epoch 0: |                                                                             | 930/? [57:26<00:00,  0.27it/s, v_num=cys0]context = [[0, 2, 12, 23, 26, 31, 35, 37, 41, 42, 44, 49], [164, 174, 179, 182, 184, 187, 189, 201, 202, 205, 207, 216]]target = [[33, 7, 10, 26, 44, 36, 39, 30, 43, 11, 38, 27], [214, 173, 176, 191, 182, 190, 175, 210, 215, 167, 195, 198]]
+Epoch 0: |                                                                             | 939/? [58:00<00:00,  0.27it/s, v_num=cys0]train step 940; scene = [['db02cd4ba6a027da'], ['8f5e074629cedd06']]; loss = 0.050028
+Epoch 0: |                                                                             | 940/? [58:03<00:00,  0.27it/s, v_num=cys0]context = [[127, 141, 154, 157, 160, 163, 172, 173, 189, 190, 192, 196], [6, 23, 36, 41, 46, 51, 57, 61, 65, 66, 83, 84]]target = [[157, 179, 153, 148, 161, 137, 194, 177, 128, 162, 166, 174], [38, 41, 32, 28, 8, 37, 36, 9, 77, 18, 80, 26]]
+Epoch 0: |                                                                             | 949/? [58:36<00:00,  0.27it/s, v_num=cys0]train step 950; scene = [['5b98e84e8e7ffef0'], ['d7ca47da5fac7140'], ['eff98653337775a8'], ['ba00608cd351deb0']]; loss = 0.047744
+Epoch 0: |                                                                             | 950/? [58:40<00:00,  0.27it/s, v_num=cys0]context = [[28, 33, 43, 46, 47, 51, 54, 67, 68, 72, 74, 75, 88, 90, 94, 98, 101, 102, 103, 105, 107, 113, 118, 125]]target = [[56, 59, 68, 66, 119, 65, 80, 100, 101, 39, 97, 94, 110, 62, 109, 40, 91, 42, 44, 78, 60, 73, 108, 50]]
+Epoch 0: |                                                                             | 959/? [59:14<00:00,  0.27it/s, v_num=cys0]train step 960; scene = [['f6d65c637ff68de3'], ['9b4d466924c40d8b'], ['b6c9aa729ebc703e']]; loss = 0.112386
+Epoch 0: |                                                                             | 960/? [59:18<00:00,  0.27it/s, v_num=cys0]context = [[2, 10, 11, 14, 19, 21, 23, 33, 39, 45, 49, 52, 53, 55, 58, 61, 71, 74, 80, 83, 84, 90, 94, 99]]target = [[41, 39, 14, 8, 52, 95, 19, 76, 24, 68, 75, 69, 22, 3, 47, 98, 17, 38, 89, 35, 21, 57, 45, 43]]
+Epoch 0: |                                                                             | 969/? [59:51<00:00,  0.27it/s, v_num=cys0]train step 970; scene = [['4b6cbf1f4c87d918']]; loss = 0.046383
+Epoch 0: |                                                                             | 970/? [59:55<00:00,  0.27it/s, v_num=cys0]context = [[20, 23, 30, 51, 53, 56, 60, 70], [145, 151, 184, 191, 192, 195, 215, 219], [36, 61, 63, 93, 106, 109, 111, 114]]target = [[52, 61, 27, 43, 44, 68, 67, 65], [210, 148, 165, 147, 166, 153, 170, 176], [108, 92, 76, 81, 71, 104, 41, 54]]
+Epoch 0: |                                                                           | 979/? [1:00:28<00:00,  0.27it/s, v_num=cys0]train step 980; scene = [['0492d6125268e9ae']]; loss = 0.044542
+Epoch 0: |                                                                           | 980/? [1:00:32<00:00,  0.27it/s, v_num=cys0]context = [[2, 4, 66, 70], [30, 96, 104, 105], [2, 13, 43, 65], [6, 42, 79, 91], [48, 85, 95, 96], [43, 62, 80, 101]]target = [[16, 66, 47, 53], [66, 42, 33, 84], [51, 25, 31, 53], [39, 90, 74, 16], [89, 80, 56, 68], [49, 45, 72, 48]]
+Epoch 0: |                                                                           | 989/? [1:01:06<00:00,  0.27it/s, v_num=cys0]train step 990; scene = [['b0413d361b4e8abb']]; loss = 0.044510
+Epoch 0: |                                                                           | 990/? [1:01:10<00:00,  0.27it/s, v_num=cys0]context = [[14, 19, 24, 29, 30, 38, 39, 40, 53, 57, 58, 65], [0, 1, 21, 33, 35, 38, 39, 43, 49, 58, 59, 64]]target = [[38, 37, 40, 27, 39, 45, 15, 52, 55, 22, 23, 31], [12, 42, 48, 30, 10, 58, 52, 59, 8, 46, 19, 44]]
+Epoch 0: |                                                                           | 999/? [1:01:42<00:00,  0.27it/s, v_num=cys0]train step 1000; scene = [['2bf91de5cd028c93'], ['2abe932fd9d76528'], ['8fabc39ad677dece'], ['188398a54205f797']]; loss = 0.068127
+Epoch 0: |                                                                          | 1000/? [1:01:46<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 1000; scene = ['647f2049bf4cb3f3'];
+target intrinsic:  tensor(0.8998, device='cuda:0') tensor(0.9001, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8811, device='cuda:0') tensor(0.8800, device='cuda:0')
+[2026-02-25 08:34:25,661][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1000/? [1:01:47<00:00,  0.27it/s, v_num=cys0]context = [[176, 185, 186, 188, 208, 224, 228, 246], [0, 6, 9, 11, 37, 43, 46, 54], [0, 16, 23, 40, 41, 44, 55, 68]]target = [[205, 177, 207, 223, 188, 184, 231, 179], [7, 27, 9, 11, 39, 15, 12, 28], [38, 2, 54, 3, 25, 7, 37, 46]]
+[2026-02-25 08:34:29,312][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1009/? [1:02:22<00:00,  0.27it/s, v_num=cys0]train step 1010; scene = [['3b42fa1245f6b00b'], ['c472581eb9a351cb'], ['222255ddcacd20cf'], ['b953af75bafccce8'], ['2c862f06019afc7e'], ['4178cb9ef3e0fd6d'], ['7f9480301fa3e38b'], ['8ccc9f1ead9cad5b']]; loss = 0.066982
+Epoch 0: |                                                                          | 1010/? [1:02:26<00:00,  0.27it/s, v_num=cys0]context = [[0, 73], [12, 74], [41, 117], [34, 98], [3, 58], [117, 195], [117, 198], [33, 101], [0, 71], [45, 102], [36, 110], [82, 153]]target = [[44, 27], [42, 20], [116, 46], [88, 73], [5, 49], [190, 129], [127, 171], [75, 70], [61, 20], [97, 100], [51, 43], [151, 139]]
+Epoch 0: |                                                                          | 1019/? [1:02:59<00:00,  0.27it/s, v_num=cys0]train step 1020; scene = [['3a95efa051965ee0']]; loss = 0.075249
+Epoch 0: |                                                                          | 1020/? [1:03:03<00:00,  0.27it/s, v_num=cys0]context = [[8, 11, 13, 14, 15, 20, 21, 31, 34, 43, 46, 50, 55, 61, 71, 75, 76, 77, 81, 83, 85, 87, 93, 105]]target = [[83, 40, 76, 65, 29, 87, 84, 91, 31, 73, 50, 22, 38, 57, 99, 20, 19, 24, 26, 82, 36, 34, 44, 15]]
+Epoch 0: |                                                                          | 1029/? [1:03:36<00:00,  0.27it/s, v_num=cys0]train step 1030; scene = [['6bd6d8270eee0b81'], ['2ac1cf1adda42447'], ['a017de2545c81e26'], ['4d5593ce15e317ed'], ['437188c182ae6a10'], ['86440d86f60644b7'], ['27508406fddac6dd'], ['7feae1e0e2c1701b'], ['a86dd7257f21bc73'], ['a025554e94654c2f'], ['20ee785f1bc6340d'], ['db6cd90de8fee2ff']]; loss = 0.115055
+Epoch 0: |                                                                          | 1030/? [1:03:40<00:00,  0.27it/s, v_num=cys0]context = [[84, 89, 90, 91, 93, 111, 112, 113, 116, 117, 121, 126, 128, 134, 138, 139, 145, 153, 159, 161, 169, 173, 178, 181]]target = [[147, 138, 152, 92, 157, 97, 115, 154, 98, 113, 142, 137, 177, 165, 158, 146, 120, 139, 136, 122, 104, 85, 128, 144]]
+Epoch 0: |                                                                          | 1039/? [1:04:13<00:00,  0.27it/s, v_num=cys0]train step 1040; scene = [['80fd51689742e978'], ['5eb07ee6a9e4fec2']]; loss = 0.057091
+Epoch 0: |                                                                          | 1040/? [1:04:16<00:00,  0.27it/s, v_num=cys0]context = [[98, 99, 101, 110, 114, 116, 123, 126, 131, 137, 139, 150, 153, 155, 156, 158, 165, 177, 178, 184, 185, 187, 192, 195]]target = [[134, 141, 100, 117, 190, 107, 162, 158, 179, 192, 121, 151, 157, 139, 118, 175, 156, 183, 127, 167, 103, 172, 154, 116]]
+Epoch 0: |                                                                          | 1049/? [1:04:49<00:00,  0.27it/s, v_num=cys0]train step 1050; scene = [['77dd1e7595b8f9a1']]; loss = 0.057489
+Epoch 0: |                                                                          | 1050/? [1:04:53<00:00,  0.27it/s, v_num=cys0]context = [[114, 115, 118, 119, 124, 126, 130, 147, 149, 151, 158, 166, 169, 170, 173, 174, 177, 180, 191, 192, 203, 206, 208, 211]]target = [[209, 128, 160, 132, 164, 207, 201, 145, 117, 163, 120, 202, 151, 119, 124, 162, 189, 133, 184, 198, 153, 116, 188, 193]]
+Epoch 0: |                                                                          | 1059/? [1:05:26<00:00,  0.27it/s, v_num=cys0]train step 1060; scene = [['51b6e63c903275b4'], ['2a9baa599c00613f'], ['1d7dc2d489714b09'], ['9f92d7e79d4a35bb']]; loss = 0.050230
+Epoch 0: |                                                                          | 1060/? [1:05:30<00:00,  0.27it/s, v_num=cys0]context = [[15, 17, 24, 33, 36, 40, 69, 74, 77, 82, 83, 84], [1, 5, 9, 16, 25, 29, 32, 33, 44, 47, 49, 51]]target = [[41, 42, 50, 25, 44, 31, 43, 24, 81, 18, 26, 71], [45, 3, 37, 11, 42, 29, 49, 41, 17, 22, 39, 23]]
+Epoch 0: |                                                                          | 1069/? [1:06:04<00:00,  0.27it/s, v_num=cys0]train step 1070; scene = [['a464ce7c8383dbbb']]; loss = 0.085306
+Epoch 0: |                                                                          | 1070/? [1:06:07<00:00,  0.27it/s, v_num=cys0]context = [[6, 10, 28, 32, 57, 71, 89, 92], [34, 52, 65, 75, 88, 89, 114, 117], [119, 123, 131, 133, 137, 160, 180, 190]]target = [[90, 59, 62, 24, 52, 29, 47, 35], [72, 39, 106, 85, 58, 74, 77, 40], [188, 149, 156, 166, 161, 120, 125, 158]]
+Epoch 0: |                                                                          | 1079/? [1:06:41<00:00,  0.27it/s, v_num=cys0]train step 1080; scene = [['b382af5f342061fa'], ['f6a0556897b15d6b']]; loss = 0.043402
+Epoch 0: |                                                                          | 1080/? [1:06:44<00:00,  0.27it/s, v_num=cys0]context = [[127, 150, 162, 189], [25, 41, 98, 102], [6, 17, 20, 68], [24, 48, 78, 100], [154, 155, 175, 235], [41, 55, 86, 109]]target = [[180, 163, 148, 162], [45, 57, 100, 82], [16, 24, 44, 47], [50, 42, 47, 92], [206, 210, 196, 211], [80, 89, 82, 62]]
+Epoch 0: |                                                                          | 1089/? [1:07:18<00:00,  0.27it/s, v_num=cys0]train step 1090; scene = [['5beb85aaf29d1242']]; loss = 0.066593
+Epoch 0: |                                                                          | 1090/? [1:07:21<00:00,  0.27it/s, v_num=cys0]context = [[27, 37, 53, 65, 72, 76, 79, 83, 86, 97, 104, 109], [10, 17, 18, 32, 38, 45, 46, 52, 67, 68, 72, 79]]target = [[103, 63, 65, 74, 40, 85, 44, 107, 57, 58, 53, 29], [18, 61, 49, 74, 69, 21, 73, 59, 30, 67, 11, 43]]
+Epoch 0: |                                                                          | 1099/? [1:07:55<00:00,  0.27it/s, v_num=cys0]train step 1100; scene = [['db811a2460c4f9b5'], ['7db8c4965bba509a'], ['a7a79393e5bb8108'], ['2d77b1dd90856337']]; loss = 0.084522
+Epoch 0: |                                                                          | 1100/? [1:07:58<00:00,  0.27it/s, v_num=cys0]context = [[1, 10, 15, 20, 27, 35, 38, 44, 51, 52, 53, 70], [40, 44, 45, 46, 52, 54, 56, 63, 88, 95, 96, 98]]target = [[18, 63, 10, 69, 55, 26, 40, 21, 59, 51, 42, 48], [84, 92, 51, 54, 77, 50, 59, 47, 86, 43, 89, 87]]
+Epoch 0: |                                                                          | 1109/? [1:08:31<00:00,  0.27it/s, v_num=cys0]train step 1110; scene = [['a815d5a5f2ec9562'], ['3be8c5ae6c95c9b4']]; loss = 0.045937
+Epoch 0: |                                                                          | 1110/? [1:08:35<00:00,  0.27it/s, v_num=cys0]context = [[0, 2, 11, 15, 25, 28, 29, 32, 33, 35, 40, 44, 47, 49, 55, 56, 66, 72, 74, 84, 86, 88, 96, 97]]target = [[15, 85, 74, 83, 43, 47, 21, 40, 39, 66, 3, 26, 29, 9, 38, 23, 94, 36, 62, 19, 12, 44, 76, 31]]
+Epoch 0: |                                                                          | 1119/? [1:09:08<00:00,  0.27it/s, v_num=cys0]train step 1120; scene = [['3ab1ce6779776017']]; loss = 0.101712
+Epoch 0: |                                                                          | 1120/? [1:09:12<00:00,  0.27it/s, v_num=cys0]context = [[9, 16, 17, 21, 23, 30, 31, 43, 55, 57, 59, 61], [11, 25, 51, 66, 72, 76, 82, 87, 91, 94, 97, 99]]target = [[56, 22, 30, 35, 59, 36, 38, 23, 34, 11, 26, 51], [19, 73, 37, 71, 50, 43, 17, 80, 25, 54, 28, 24]]
+Epoch 0: |                                                                          | 1129/? [1:09:45<00:00,  0.27it/s, v_num=cys0]train step 1130; scene = [['fd0ebd5afbfd1acf'], ['edf6636dfd51ba3c'], ['7b2c118f021e6902'], ['b3356e816a130b87'], ['78cbac3ff58f2e41'], ['94c654bd3e031bcb'], ['b281bf93286a0573'], ['d9dce3382830aea6'], ['070a524bacb9aa38'], ['cae0139a521aa052'], ['5645a008715acf0a'], ['23174a6cd65a0731']]; loss = 0.121455
+Epoch 0: |                                                                          | 1130/? [1:09:48<00:00,  0.27it/s, v_num=cys0]context = [[8, 50, 81, 84], [2, 42, 44, 54], [25, 26, 79, 87], [25, 30, 88, 92], [45, 100, 129, 135], [17, 68, 77, 83]]target = [[62, 80, 76, 55], [11, 29, 39, 13], [42, 27, 75, 58], [31, 63, 29, 48], [72, 58, 93, 106], [66, 77, 60, 59]]
+Epoch 0: |                                                                          | 1139/? [1:10:21<00:00,  0.27it/s, v_num=cys0]train step 1140; scene = [['090a038e9d844b4e'], ['d28a2455cc34badb']]; loss = 0.046314
+Epoch 0: |                                                                          | 1140/? [1:10:24<00:00,  0.27it/s, v_num=cys0]context = [[19, 20, 28, 32, 34, 36, 38, 40, 52, 54, 63, 65, 68, 77, 78, 87, 94, 98, 100, 105, 107, 113, 114, 116]]target = [[32, 23, 97, 28, 51, 21, 114, 76, 46, 84, 74, 31, 98, 108, 77, 82, 72, 49, 103, 47, 80, 39, 59, 70]]
+Epoch 0: |                                                                          | 1149/? [1:10:58<00:00,  0.27it/s, v_num=cys0]train step 1150; scene = [['82347d56a4a55c27'], ['ee03755cce11b682']]; loss = 0.071285
+Epoch 0: |                                                                          | 1150/? [1:11:02<00:00,  0.27it/s, v_num=cys0]context = [[13, 16, 50, 52, 55, 62], [55, 56, 79, 89, 111, 117], [35, 43, 73, 74, 82, 85], [30, 56, 58, 80, 102, 108]]target = [[39, 59, 30, 34, 42, 27], [91, 98, 71, 70, 84, 68], [42, 44, 39, 68, 69, 50], [42, 68, 72, 55, 87, 51]]
+Epoch 0: |                                                                          | 1159/? [1:11:36<00:00,  0.27it/s, v_num=cys0]train step 1160; scene = [['255998558abc7172']]; loss = 0.082432
+Epoch 0: |                                                                          | 1160/? [1:11:40<00:00,  0.27it/s, v_num=cys0]context = [[13, 16, 18, 19, 21, 25, 28, 29, 36, 38, 43, 47, 49, 50, 53, 56, 67, 68, 69, 71, 72, 93, 108, 110]]target = [[47, 34, 77, 80, 28, 94, 97, 19, 29, 26, 68, 40, 53, 20, 60, 30, 92, 70, 83, 48, 106, 15, 63, 41]]
+Epoch 0: |                                                                          | 1169/? [1:12:13<00:00,  0.27it/s, v_num=cys0]train step 1170; scene = [['771aa992eae9a574'], ['4f9716bb3dc7feec'], ['fbed2318ae410b31'], ['8e1b4054949b6a46']]; loss = 0.059541
+Epoch 0: |                                                                          | 1170/? [1:12:17<00:00,  0.27it/s, v_num=cys0]context = [[207, 214, 223, 228, 242, 250, 268, 269], [35, 58, 71, 96, 97, 109, 110, 114], [0, 27, 28, 46, 50, 60, 68, 75]]target = [[216, 210, 265, 253, 237, 223, 241, 234], [105, 106, 100, 96, 73, 51, 109, 93], [2, 54, 60, 70, 28, 22, 51, 62]]
+Epoch 0: |                                                                          | 1179/? [1:12:50<00:00,  0.27it/s, v_num=cys0]train step 1180; scene = [['18c880b4b5ef683e']]; loss = 0.077895
+Epoch 0: |                                                                          | 1180/? [1:12:54<00:00,  0.27it/s, v_num=cys0]context = [[48, 60, 70, 80, 90, 109, 110, 114], [3, 8, 11, 23, 28, 34, 62, 82], [14, 17, 19, 24, 30, 34, 59, 67]]target = [[85, 101, 84, 113, 106, 74, 87, 98], [81, 68, 48, 46, 67, 69, 56, 22], [31, 33, 45, 20, 17, 61, 37, 44]]
+Epoch 0: |                                                                          | 1189/? [1:13:27<00:00,  0.27it/s, v_num=cys0]train step 1190; scene = [['019fdd708d7163bd'], ['7046980b0d3d3c63'], ['f1af5d4039ce3a2c']]; loss = 0.056714
+Epoch 0: |                                                                          | 1190/? [1:13:30<00:00,  0.27it/s, v_num=cys0]context = [[0, 7, 8, 11, 15, 20, 21, 28, 30, 33, 38, 42, 47, 54, 59, 63, 66, 71, 73, 79, 81, 88, 90, 97]]target = [[83, 9, 22, 57, 21, 76, 37, 10, 7, 13, 43, 47, 19, 6, 51, 71, 95, 78, 24, 92, 30, 28, 55, 84]]
+Epoch 0: |                                                                          | 1199/? [1:14:04<00:00,  0.27it/s, v_num=cys0]train step 1200; scene = [['5070ea042c65de0d'], ['19f950b6900d6176'], ['02768fad99be3290'], ['ed6c9a5913622c3d'], ['0a4cbe699be68e5c'], ['0892e76375b283ba'], ['efd001b0d5127d61'], ['9adb745c741c85e0'], ['e4774e728791dc20'], ['23710dc8de8b4a49'], ['347d7c1f3516f732'], ['97283dc038203c65']]; loss = 0.117074
+Epoch 0: |                                                                          | 1200/? [1:14:08<00:00,  0.27it/s, v_num=cys0]context = [[44, 48, 49, 54, 63, 66, 67, 70, 76, 77, 78, 84, 90, 100, 101, 105, 106, 111, 113, 116, 120, 123, 125, 141]]target = [[63, 111, 132, 61, 108, 104, 110, 81, 59, 89, 131, 126, 73, 100, 134, 74, 121, 84, 80, 51, 140, 133, 55, 83]]
+[2026-02-25 08:46:50,776][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1209/? [1:14:45<00:00,  0.27it/s, v_num=cys0]train step 1210; scene = [['7ba6259f378c70f8']]; loss = 0.047970
+Epoch 0: |                                                                          | 1210/? [1:14:48<00:00,  0.27it/s, v_num=cys0]context = [[77, 82, 86, 95, 101, 107, 111, 112, 113, 117, 120, 130], [9, 29, 33, 44, 49, 53, 54, 58, 67, 68, 81, 96]]target = [[99, 107, 90, 119, 117, 113, 92, 120, 121, 123, 86, 94], [32, 87, 79, 70, 51, 71, 33, 67, 53, 82, 41, 56]]
+Epoch 0: |                                                                          | 1219/? [1:15:21<00:00,  0.27it/s, v_num=cys0]train step 1220; scene = [['0049c83bad21bbdf'], ['a73f4bf6ce8b2a00'], ['f7bef32e09ab061e'], ['175e1611df44964f']]; loss = 0.039392
+Epoch 0: |                                                                          | 1220/? [1:15:25<00:00,  0.27it/s, v_num=cys0]context = [[44, 47, 55, 57, 65, 80, 85, 87, 88, 94, 112, 116], [25, 37, 49, 59, 61, 70, 82, 84, 86, 94, 97, 100]]target = [[96, 81, 62, 98, 79, 76, 52, 95, 59, 65, 56, 72], [34, 26, 32, 50, 44, 97, 67, 53, 75, 66, 59, 40]]
+Epoch 0: |                                                                          | 1229/? [1:15:58<00:00,  0.27it/s, v_num=cys0]train step 1230; scene = [['342ff7b7111b53c1'], ['9ebe2434b1d68246'], ['b84cce034b55e1e4']]; loss = 0.043910
+Epoch 0: |                                                                          | 1230/? [1:16:02<00:00,  0.27it/s, v_num=cys0]context = [[5, 8, 11, 16, 19, 23, 25, 34, 35, 36, 43, 49, 51, 57, 60, 68, 77, 78, 86, 88, 97, 98, 101, 102]]target = [[84, 48, 64, 18, 41, 19, 60, 25, 52, 49, 51, 50, 75, 38, 42, 81, 92, 32, 69, 67, 26, 100, 65, 61]]
+Epoch 0: |                                                                          | 1239/? [1:16:36<00:00,  0.27it/s, v_num=cys0]train step 1240; scene = [['d7ac888a45c3c904'], ['60d988ddbc2d04f1'], ['617340307747e227'], ['969c91a2507e2d81'], ['af197296b340b564'], ['d488b5f08aadff0c']]; loss = 0.076657
+Epoch 0: |                                                                          | 1240/? [1:16:39<00:00,  0.27it/s, v_num=cys0]context = [[9, 16, 18, 25, 28, 30, 41, 43, 46, 51, 53, 55, 56, 57, 59, 68, 71, 72, 74, 77, 87, 95, 105, 106]]target = [[72, 53, 26, 17, 11, 65, 42, 97, 60, 75, 93, 10, 45, 15, 102, 38, 101, 54, 23, 98, 13, 68, 16, 46]]
+Epoch 0: |                                                                          | 1249/? [1:17:12<00:00,  0.27it/s, v_num=cys0]train step 1250; scene = [['d51d569c27d0b2b5'], ['e33b2a9076f25c4d'], ['2f330103819454c6'], ['944e92ff3fea78eb'], ['5a6387d05cd51e02'], ['2f269b68e14e256a'], ['55f1b82ae9c5571f'], ['1490c145692a1899'], ['b1b24b049d5a5da4'], ['834e851000651b8f'], ['4bc6b34a301aac73'], ['97caecafbfa7f1f6']]; loss = 0.103517
+Epoch 0: |                                                                          | 1250/? [1:17:16<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 1250; scene = ['70b0a33083333dc9'];
+target intrinsic:  tensor(0.8872, device='cuda:0') tensor(0.8874, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8753, device='cuda:0') tensor(0.8759, device='cuda:0')
+[2026-02-25 08:49:55,355][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1250/? [1:17:17<00:00,  0.27it/s, v_num=cys0]context = [[0, 2, 3, 7, 9, 14, 16, 19, 27, 30, 32, 34, 39, 43, 56, 62, 65, 73, 75, 80, 82, 85, 91, 97]]target = [[7, 68, 16, 60, 94, 67, 34, 41, 53, 11, 1, 12, 30, 6, 61, 76, 22, 32, 25, 81, 88, 33, 9, 45]]
+Epoch 0: |                                                                          | 1259/? [1:17:50<00:00,  0.27it/s, v_num=cys0]train step 1260; scene = [['1a1f0a618bad5d30'], ['3046f14dad8f19f2'], ['8b4aecd7318912e3']]; loss = 0.050226
+Epoch 0: |                                                                          | 1260/? [1:17:53<00:00,  0.27it/s, v_num=cys0]context = [[44, 48, 50, 51, 55, 60, 62, 66, 72, 78, 79, 85, 91, 98, 99, 100, 102, 112, 117, 121, 123, 125, 129, 141]]target = [[73, 109, 111, 54, 63, 69, 65, 124, 82, 119, 125, 110, 117, 56, 123, 86, 99, 126, 134, 53, 84, 112, 74, 100]]
+Epoch 0: |                                                                          | 1269/? [1:18:27<00:00,  0.27it/s, v_num=cys0]train step 1270; scene = [['a0438ace9b7619cf'], ['336b3140d9d8bebd']]; loss = 0.071758
+Epoch 0: |                                                                          | 1270/? [1:18:31<00:00,  0.27it/s, v_num=cys0]context = [[20, 22, 25, 26, 40, 44, 45, 51, 55, 63, 67, 70, 72, 76, 80, 81, 82, 84, 88, 96, 104, 109, 110, 117]]target = [[84, 115, 116, 28, 110, 27, 111, 85, 100, 39, 61, 63, 68, 56, 101, 113, 108, 94, 73, 62, 67, 98, 47, 55]]
+Epoch 0: |                                                                          | 1279/? [1:19:04<00:00,  0.27it/s, v_num=cys0]train step 1280; scene = [['118f75d62c9a1f46'], ['3bcbf64a56113736'], ['6ca39802dcef328e'], ['0d754618c77e44f6']]; loss = 0.064913
+Epoch 0: |                                                                          | 1280/? [1:19:07<00:00,  0.27it/s, v_num=cys0]context = [[5, 19, 21, 22, 26, 37, 47, 50, 53, 75, 85, 87], [4, 19, 23, 26, 29, 36, 37, 63, 69, 72, 75, 80]]target = [[65, 33, 14, 32, 52, 25, 54, 34, 15, 6, 60, 30], [11, 5, 9, 35, 70, 15, 51, 38, 20, 42, 68, 22]]
+Epoch 0: |                                                                          | 1289/? [1:19:40<00:00,  0.27it/s, v_num=cys0]train step 1290; scene = [['4738c57794aa47f8']]; loss = 0.055996
+Epoch 0: |                                                                          | 1290/? [1:19:43<00:00,  0.27it/s, v_num=cys0]context = [[1, 4, 14, 17, 27, 30, 32, 37, 44, 52, 64, 80], [0, 8, 12, 14, 24, 28, 29, 37, 49, 54, 56, 67]]target = [[64, 28, 44, 35, 23, 68, 42, 61, 3, 77, 52, 14], [1, 11, 56, 59, 13, 63, 55, 61, 38, 9, 18, 37]]
+Epoch 0: |                                                                          | 1299/? [1:20:17<00:00,  0.27it/s, v_num=cys0]train step 1300; scene = [['f18e524ab2d288c3'], ['9dff9a317da1d49d'], ['11a87d6d490e024d']]; loss = 0.052190
+Epoch 0: |                                                                          | 1300/? [1:20:21<00:00,  0.27it/s, v_num=cys0]context = [[5, 6, 8, 12, 30, 45, 58, 64, 65, 75, 79, 85], [22, 26, 27, 28, 39, 45, 47, 55, 56, 57, 66, 72]]target = [[34, 57, 32, 7, 11, 58, 73, 68, 83, 14, 61, 55], [42, 34, 36, 47, 33, 67, 31, 30, 50, 27, 24, 58]]
+Epoch 0: |                                                                          | 1309/? [1:20:54<00:00,  0.27it/s, v_num=cys0]train step 1310; scene = [['589e362118b32d25'], ['d13522abd38eddfe'], ['f1657c6128d2b332'], ['1cb3ecb30e3e9d0e'], ['827c5f3b3886553d'], ['4d27fb96530fe02b']]; loss = 0.115709
+Epoch 0: |                                                                          | 1310/? [1:20:58<00:00,  0.27it/s, v_num=cys0]context = [[20, 21, 27, 31, 36, 40, 48, 51, 55, 62, 69, 72, 74, 75, 83, 89, 96, 97, 99, 104, 105, 106, 114, 117]]target = [[57, 113, 99, 80, 42, 30, 24, 74, 34, 52, 63, 59, 100, 101, 76, 36, 105, 66, 55, 27, 84, 22, 79, 86]]
+Epoch 0: |                                                                          | 1319/? [1:21:31<00:00,  0.27it/s, v_num=cys0]train step 1320; scene = [['c21766bd51fed5bc'], ['47a2a8b326b9f40e'], ['d0c0d78936b6e5ce'], ['e0ee3878561a5fed']]; loss = 0.064375
+Epoch 0: |                                                                          | 1320/? [1:21:34<00:00,  0.27it/s, v_num=cys0]context = [[194, 242], [34, 99], [51, 136], [156, 207], [36, 100], [0, 48], [10, 99], [22, 79], [18, 99], [2, 74], [39, 128], [201, 272]]target = [[225, 201], [63, 36], [58, 72], [193, 181], [39, 77], [13, 1], [29, 68], [36, 58], [45, 66], [58, 18], [100, 50], [234, 262]]
+Epoch 0: |                                                                          | 1329/? [1:22:07<00:00,  0.27it/s, v_num=cys0]train step 1330; scene = [['07fdb102ee3677f5']]; loss = 0.046716
+Epoch 0: |                                                                          | 1330/? [1:22:11<00:00,  0.27it/s, v_num=cys0]context = [[20, 23, 28, 29, 37, 43, 49, 53, 56, 57, 61, 65, 69, 70, 80, 92, 98, 102, 103, 104, 106, 108, 116, 117]]target = [[37, 57, 73, 79, 30, 102, 42, 115, 86, 33, 28, 25, 114, 53, 76, 107, 36, 106, 52, 49, 59, 66, 61, 101]]
+Epoch 0: |                                                                          | 1339/? [1:22:45<00:00,  0.27it/s, v_num=cys0]train step 1340; scene = [['2ff4f3b2475e0e8c'], ['db062134f9dec5f1']]; loss = 0.048112
+Epoch 0: |                                                                          | 1340/? [1:22:49<00:00,  0.27it/s, v_num=cys0]context = [[20, 22, 50, 84], [0, 36, 48, 51], [10, 38, 52, 66], [0, 24, 25, 53], [0, 4, 19, 45], [54, 57, 73, 112]]target = [[42, 28, 78, 23], [3, 12, 46, 44], [57, 60, 15, 41], [47, 3, 31, 2], [42, 13, 23, 28], [102, 69, 98, 70]]
+Epoch 0: |                                                                          | 1349/? [1:23:22<00:00,  0.27it/s, v_num=cys0]train step 1350; scene = [['87e1164e050e9686'], ['3803b2c8fd539a88']]; loss = 0.066346
+Epoch 0: |                                                                          | 1350/? [1:23:26<00:00,  0.27it/s, v_num=cys0]context = [[52, 53, 55, 63, 64, 66, 74, 75, 77, 79, 85, 89, 90, 99, 121, 122, 123, 126, 133, 134, 135, 136, 147, 149]]target = [[99, 80, 109, 136, 92, 134, 112, 76, 135, 68, 116, 138, 54, 145, 144, 143, 61, 72, 102, 77, 90, 85, 139, 58]]
+Epoch 0: |                                                                          | 1359/? [1:23:59<00:00,  0.27it/s, v_num=cys0]train step 1360; scene = [['7db4ed902d003a63']]; loss = 0.069284
+Epoch 0: |                                                                          | 1360/? [1:24:03<00:00,  0.27it/s, v_num=cys0]context = [[13, 26, 58, 83], [117, 127, 165, 171], [12, 45, 50, 66], [21, 23, 67, 101], [26, 34, 95, 97], [42, 73, 81, 98]]target = [[65, 36, 43, 27], [124, 136, 141, 132], [33, 37, 62, 29], [82, 90, 80, 97], [45, 44, 32, 86], [55, 82, 58, 89]]
+Epoch 0: |                                                                          | 1369/? [1:24:36<00:00,  0.27it/s, v_num=cys0]train step 1370; scene = [['36df585860d0ad88']]; loss = 0.035851
+Epoch 0: |                                                                          | 1370/? [1:24:40<00:00,  0.27it/s, v_num=cys0]context = [[41, 42, 45, 55, 59, 62, 67, 68, 69, 73, 79, 87, 90, 94, 104, 114, 117, 120, 121, 125, 128, 130, 135, 138]]target = [[79, 103, 135, 133, 104, 59, 92, 71, 62, 66, 61, 109, 68, 124, 74, 54, 44, 46, 57, 96, 108, 58, 85, 83]]
+Epoch 0: |                                                                          | 1379/? [1:25:13<00:00,  0.27it/s, v_num=cys0]train step 1380; scene = [['256ae648672281d1'], ['7043e8afce176c8c']]; loss = 0.075790
+Epoch 0: |                                                                          | 1380/? [1:25:17<00:00,  0.27it/s, v_num=cys0]context = [[31, 33, 48, 63, 106, 115], [8, 28, 29, 44, 67, 97], [65, 72, 100, 110, 115, 124], [8, 9, 63, 70, 75, 79]]target = [[57, 76, 83, 113, 107, 38], [51, 64, 72, 39, 18, 77], [89, 108, 95, 114, 84, 116], [20, 59, 14, 56, 15, 38]]
+Epoch 0: |                                                                          | 1389/? [1:25:49<00:00,  0.27it/s, v_num=cys0]train step 1390; scene = [['ca65a604c0f00319'], ['937ea87bb5a2047f'], ['a2b430d7bec915d8'], ['c8c116c28ca108b0']]; loss = 0.048112
+Epoch 0: |                                                                          | 1390/? [1:25:53<00:00,  0.27it/s, v_num=cys0]context = [[0, 3, 5, 26, 31, 34, 38, 40, 44, 53, 62, 70], [1, 4, 21, 29, 36, 37, 39, 44, 52, 54, 59, 64]]target = [[36, 5, 37, 41, 54, 19, 44, 21, 40, 34, 59, 32], [59, 51, 53, 46, 27, 9, 21, 52, 24, 54, 32, 48]]
+Epoch 0: |                                                                          | 1399/? [1:26:26<00:00,  0.27it/s, v_num=cys0]train step 1400; scene = [['73fef5139753e974'], ['4b3e117d4f50b167'], ['0e916e63743f841b']]; loss = 0.058952
+Epoch 0: |                                                                          | 1400/? [1:26:30<00:00,  0.27it/s, v_num=cys0]context = [[23, 26, 27, 29, 31, 32, 41, 47, 49, 50, 55, 56, 58, 71, 77, 78, 87, 91, 97, 100, 101, 102, 113, 120]]target = [[55, 52, 103, 28, 68, 66, 49, 93, 76, 109, 105, 112, 100, 101, 31, 70, 64, 116, 77, 56, 46, 48, 32, 85]]
+[2026-02-25 08:59:12,245][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1409/? [1:27:08<00:00,  0.27it/s, v_num=cys0]train step 1410; scene = [['d9ce9620cf31e776'], ['36bf8b485c1284dd']]; loss = 0.051202
+Epoch 0: |                                                                          | 1410/? [1:27:12<00:00,  0.27it/s, v_num=cys0]context = [[44, 47, 51, 55, 57, 59, 61, 62, 68, 72, 78, 80, 82, 88, 91, 97, 109, 112, 116, 118, 135, 138, 140, 141]]target = [[54, 91, 112, 61, 115, 80, 93, 45, 105, 51, 131, 102, 55, 60, 57, 136, 129, 99, 125, 113, 84, 89, 58, 65]]
+Epoch 0: |                                                                          | 1419/? [1:27:44<00:00,  0.27it/s, v_num=cys0]train step 1420; scene = [['df4e1717577f82c1'], ['50b38521b86dd7b6'], ['e4fc6fa4bbd6efa1'], ['54d3668b6c7ed4b8']]; loss = 0.041966
+Epoch 0: |                                                                          | 1420/? [1:27:48<00:00,  0.27it/s, v_num=cys0]context = [[38, 43, 53, 54, 68, 87, 97, 101], [52, 53, 74, 84, 86, 96, 132, 133], [157, 173, 184, 190, 200, 203, 210, 211]]target = [[77, 71, 57, 46, 97, 67, 93, 96], [83, 127, 75, 122, 56, 78, 130, 107], [177, 183, 188, 174, 191, 203, 200, 170]]
+Epoch 0: |                                                                          | 1429/? [1:28:21<00:00,  0.27it/s, v_num=cys0]train step 1430; scene = [['e70c21625fceaebd'], ['4cfdc3ed984caecf'], ['dfc00f3016d34131'], ['f5338d00f9c021a3'], ['89fbc4149a5d7348'], ['d2d3474ebd2be7e8'], ['0b42b8cf14bd1e13'], ['358a2d1387de3df8'], ['abe463f94db2c398'], ['d35b7c2c52c45a54'], ['12f629409a280733'], ['6b09a022387bd762']]; loss = 0.100819
+Epoch 0: |                                                                          | 1430/? [1:28:25<00:00,  0.27it/s, v_num=cys0]context = [[146, 159, 179, 187, 191, 200], [3, 11, 37, 39, 40, 48], [3, 16, 25, 27, 47, 49], [31, 36, 39, 84, 87, 94]]target = [[154, 176, 156, 190, 172, 150], [13, 36, 28, 46, 33, 32], [33, 7, 23, 32, 20, 35], [53, 82, 44, 85, 38, 54]]
+Epoch 0: |                                                                          | 1439/? [1:28:58<00:00,  0.27it/s, v_num=cys0]train step 1440; scene = [['83f69d126eb1528d'], ['e92e4346a6224492']]; loss = 0.046798
+Epoch 0: |                                                                          | 1440/? [1:29:02<00:00,  0.27it/s, v_num=cys0]context = [[1, 19, 33, 39, 51, 53, 57, 61], [4, 18, 19, 43, 44, 47, 49, 75], [40, 51, 69, 84, 88, 93, 99, 104]]target = [[3, 30, 34, 24, 46, 8, 13, 4], [19, 24, 28, 56, 55, 10, 66, 21], [54, 88, 79, 100, 81, 84, 83, 72]]
+Epoch 0: |                                                                          | 1449/? [1:29:36<00:00,  0.27it/s, v_num=cys0]train step 1450; scene = [['c05fd148d2da6d26'], ['b49cc4ec7c6b0050'], ['0d158225b3c47682']]; loss = 0.069016
+Epoch 0: |                                                                          | 1450/? [1:29:40<00:00,  0.27it/s, v_num=cys0]context = [[2, 15, 27, 38, 41, 50, 51, 57, 58, 61, 62, 73], [16, 17, 19, 26, 41, 44, 80, 89, 90, 92, 100, 101]]target = [[37, 21, 47, 67, 70, 31, 10, 57, 34, 30, 50, 52], [84, 77, 99, 81, 19, 59, 24, 78, 64, 25, 30, 90]]
+Epoch 0: |                                                                          | 1459/? [1:30:13<00:00,  0.27it/s, v_num=cys0]train step 1460; scene = [['8d583bfb265295b9']]; loss = 0.053408
+Epoch 0: |                                                                          | 1460/? [1:30:17<00:00,  0.27it/s, v_num=cys0]context = [[143, 146, 157, 160, 179, 202], [1, 13, 19, 36, 45, 46], [175, 189, 227, 237, 252, 257], [22, 29, 47, 51, 54, 67]]target = [[156, 146, 191, 160, 196, 187], [28, 40, 8, 23, 33, 44], [234, 248, 199, 178, 220, 226], [38, 63, 50, 35, 66, 33]]
+Epoch 0: |                                                                          | 1469/? [1:30:51<00:00,  0.27it/s, v_num=cys0]train step 1470; scene = [['6ebab888069161eb']]; loss = 0.041926
+Epoch 0: |                                                                          | 1470/? [1:30:55<00:00,  0.27it/s, v_num=cys0]context = [[28, 29, 32, 37, 43, 46, 50, 65, 67, 71, 76, 77], [132, 133, 136, 137, 140, 141, 145, 147, 169, 171, 200, 204]]target = [[31, 53, 62, 64, 65, 67, 34, 72, 39, 58, 57, 68], [174, 172, 164, 145, 162, 199, 202, 142, 163, 159, 176, 200]]
+Epoch 0: |                                                                          | 1479/? [1:31:28<00:00,  0.27it/s, v_num=cys0]train step 1480; scene = [['964d888d8d08f2aa'], ['58901334e2d813d9'], ['ea4146e3386ff1ac']]; loss = 0.106694
+Epoch 0: |                                                                          | 1480/? [1:31:32<00:00,  0.27it/s, v_num=cys0]context = [[50, 57, 67, 77, 79, 101], [66, 67, 75, 83, 96, 117], [53, 65, 70, 81, 130, 138], [19, 45, 53, 57, 65, 66]]target = [[76, 85, 70, 53, 90, 65], [87, 116, 103, 85, 111, 97], [80, 116, 87, 117, 54, 112], [35, 47, 59, 51, 53, 28]]
+Epoch 0: |                                                                          | 1489/? [1:32:05<00:00,  0.27it/s, v_num=cys0]train step 1490; scene = [['f73db02fdfe72073'], ['a4eeae8de8e2e98e']]; loss = 0.038537
+Epoch 0: |                                                                          | 1490/? [1:32:07<00:00,  0.27it/s, v_num=cys0]context = [[10, 19, 23, 29, 35, 40, 41, 42, 43, 53, 54, 57, 63, 65, 66, 69, 72, 73, 77, 84, 91, 94, 95, 107]]target = [[97, 79, 11, 35, 65, 87, 80, 51, 32, 13, 83, 72, 85, 34, 105, 92, 59, 25, 81, 40, 36, 82, 88, 15]]
+Epoch 0: |                                                                          | 1499/? [1:32:41<00:00,  0.27it/s, v_num=cys0]train step 1500; scene = [['fa1ddac84aafd9b7']]; loss = 0.139373
+Epoch 0: |                                                                          | 1500/? [1:32:45<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 1500; scene = ['45592a7f307bccd0'];
+target intrinsic:  tensor(0.8508, device='cuda:0') tensor(0.8510, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8681, device='cuda:0') tensor(0.8685, device='cuda:0')
+[2026-02-25 09:05:35,984][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1500/? [1:32:58<00:00,  0.27it/s, v_num=cys0]context = [[5, 10, 19, 23, 27, 33, 40, 57, 60, 63, 64, 68], [97, 102, 107, 111, 113, 120, 137, 148, 157, 163, 168, 169]]target = [[54, 31, 25, 44, 17, 36, 32, 65, 16, 11, 23, 38], [99, 138, 103, 159, 164, 104, 123, 135, 120, 124, 130, 148]]
+Epoch 0: |                                                                          | 1509/? [1:33:30<00:00,  0.27it/s, v_num=cys0]train step 1510; scene = [['289a0ef9678c7c70'], ['f4fd1fee2a8f69ee'], ['a88b44d03528e2f2']]; loss = 0.045655
+Epoch 0: |                                                                          | 1510/? [1:33:34<00:00,  0.27it/s, v_num=cys0]context = [[1, 2, 8, 18, 19, 27, 33, 35, 48, 49, 63, 77], [0, 9, 12, 15, 20, 22, 23, 30, 34, 58, 66, 68]]target = [[12, 51, 13, 69, 65, 38, 6, 30, 37, 18, 53, 2], [8, 14, 60, 49, 63, 46, 47, 3, 35, 13, 21, 44]]
+Epoch 0: |                                                                          | 1519/? [1:34:06<00:00,  0.27it/s, v_num=cys0]train step 1520; scene = [['85cff5568e6f52e7'], ['9de970f9a14770a9'], ['a08eb87db37b694a'], ['0a8fda80930b52ae']]; loss = 0.058275
+Epoch 0: |                                                                          | 1520/? [1:34:09<00:00,  0.27it/s, v_num=cys0]context = [[6, 32, 33, 39, 62, 69, 77, 96], [32, 33, 42, 61, 83, 93, 99, 107], [13, 27, 41, 70, 78, 84, 86, 93]]target = [[28, 30, 62, 7, 57, 86, 77, 36], [101, 43, 41, 68, 46, 36, 47, 105], [27, 16, 29, 71, 40, 45, 74, 65]]
+Epoch 0: |                                                                          | 1529/? [1:34:42<00:00,  0.27it/s, v_num=cys0]train step 1530; scene = [['43cc195276cf0c56']]; loss = 0.088894
+Epoch 0: |                                                                          | 1530/? [1:34:46<00:00,  0.27it/s, v_num=cys0]context = [[86, 89, 91, 102, 109, 136, 173, 175], [1, 12, 32, 33, 36, 37, 43, 47], [15, 24, 40, 41, 44, 70, 73, 102]]target = [[138, 158, 117, 118, 170, 120, 165, 133], [37, 44, 31, 32, 24, 9, 40, 38], [16, 27, 88, 46, 69, 52, 73, 44]]
+Epoch 0: |                                                                          | 1539/? [1:35:20<00:00,  0.27it/s, v_num=cys0]train step 1540; scene = [['4b7f3f58b0838d38']]; loss = 0.048801
+Epoch 0: |                                                                          | 1540/? [1:35:23<00:00,  0.27it/s, v_num=cys0]context = [[17, 21, 24, 40, 42, 44, 49, 50, 60, 62, 63, 64, 65, 72, 73, 76, 77, 85, 89, 93, 96, 104, 108, 114]]target = [[104, 84, 41, 96, 107, 103, 94, 44, 111, 97, 83, 54, 100, 57, 27, 38, 113, 29, 18, 99, 76, 101, 55, 36]]
+Epoch 0: |                                                                          | 1549/? [1:35:57<00:00,  0.27it/s, v_num=cys0]train step 1550; scene = [['d6cc1a3af543a7f8'], ['f3dd12e8d9dd4c20'], ['b74b90e3a87f285c'], ['c51f5219c3d09e33'], ['3f6666062c86b73a'], ['3b7bd7e723f069b2'], ['b3992ad0aff60272'], ['30d52bc66d89221d']]; loss = 0.062747
+Epoch 0: |                                                                          | 1550/? [1:36:01<00:00,  0.27it/s, v_num=cys0]context = [[9, 79], [29, 75], [29, 95], [124, 170], [42, 102], [92, 159], [94, 157], [6, 75], [188, 255], [5, 76], [133, 205], [2, 90]]target = [[55, 20], [47, 52], [85, 42], [136, 147], [97, 56], [158, 99], [145, 143], [17, 9], [199, 198], [34, 26], [184, 165], [39, 28]]
+Epoch 0: |                                                                          | 1559/? [1:36:35<00:00,  0.27it/s, v_num=cys0]train step 1560; scene = [['f6a87eade96cceb1']]; loss = 0.048574
+Epoch 0: |                                                                          | 1560/? [1:36:39<00:00,  0.27it/s, v_num=cys0]context = [[9, 17, 23, 26, 33, 40, 46, 62, 63, 68, 69, 71], [117, 126, 137, 140, 146, 163, 169, 173, 178, 186, 201, 205]]target = [[25, 36, 53, 31, 59, 20, 68, 24, 66, 13, 33, 67], [128, 179, 169, 126, 188, 187, 122, 127, 144, 165, 150, 119]]
+Epoch 0: |                                                                          | 1569/? [1:37:11<00:00,  0.27it/s, v_num=cys0]train step 1570; scene = [['ab1c5358d3bb05db']]; loss = 0.043213
+Epoch 0: |                                                                          | 1570/? [1:37:15<00:00,  0.27it/s, v_num=cys0]context = [[6, 8, 14, 38, 44, 69, 83, 89], [31, 47, 49, 50, 68, 77, 89, 91], [32, 46, 50, 61, 66, 84, 89, 100]]target = [[69, 79, 13, 31, 70, 24, 8, 27], [53, 42, 44, 68, 90, 65, 87, 49], [85, 91, 86, 84, 67, 83, 72, 92]]
+Epoch 0: |                                                                          | 1579/? [1:37:48<00:00,  0.27it/s, v_num=cys0]train step 1580; scene = [['572357b6b69cb9ec'], ['326dd7b41ce515ac']]; loss = 0.043743
+Epoch 0: |                                                                          | 1580/? [1:37:51<00:00,  0.27it/s, v_num=cys0]context = [[26, 32, 46, 57, 64, 105], [32, 38, 66, 84, 85, 95], [75, 87, 98, 106, 108, 130], [7, 9, 12, 20, 48, 60]]target = [[83, 29, 52, 46, 45, 39], [41, 73, 88, 76, 54, 82], [78, 99, 126, 113, 86, 105], [29, 45, 35, 34, 32, 16]]
+Epoch 0: |                                                                          | 1589/? [1:38:25<00:00,  0.27it/s, v_num=cys0]train step 1590; scene = [['4558408811e71246'], ['4a3ff4b7939ec268']]; loss = 0.067875
+Epoch 0: |                                                                          | 1590/? [1:38:29<00:00,  0.27it/s, v_num=cys0]context = [[187, 198, 222, 224, 227, 240], [196, 213, 228, 234, 257, 263], [15, 24, 48, 51, 62, 85], [64, 73, 82, 83, 109, 117]]target = [[231, 221, 188, 222, 220, 232], [257, 228, 261, 254, 244, 210], [45, 49, 23, 62, 65, 84], [74, 94, 93, 85, 66, 104]]
+Epoch 0: |                                                                          | 1599/? [1:39:02<00:00,  0.27it/s, v_num=cys0]train step 1600; scene = [['1849a41079039a3a']]; loss = 0.055898
+Epoch 0: |                                                                          | 1600/? [1:39:05<00:00,  0.27it/s, v_num=cys0]context = [[7, 10, 13, 17, 18, 30, 32, 34, 35, 38, 42, 49, 64, 66, 71, 74, 77, 84, 89, 91, 93, 95, 103, 104]]target = [[52, 78, 50, 26, 46, 43, 79, 25, 89, 70, 58, 18, 103, 98, 16, 74, 20, 13, 85, 96, 71, 76, 91, 34]]
+[2026-02-25 09:11:48,010][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1609/? [1:39:45<00:00,  0.27it/s, v_num=cys0]train step 1610; scene = [['6405acc04d687493'], ['a802441c288b598f'], ['e2a52c2d94f21d4d']]; loss = 0.069017
+Epoch 0: |                                                                          | 1610/? [1:39:48<00:00,  0.27it/s, v_num=cys0]context = [[110, 111, 116, 122, 133, 134, 135, 143, 145, 148, 155, 167, 173, 177, 178, 180, 181, 183, 192, 193, 197, 199, 203, 207]]target = [[137, 115, 169, 197, 176, 116, 146, 159, 144, 119, 129, 161, 181, 160, 151, 196, 131, 204, 201, 157, 138, 123, 180, 117]]
+Epoch 0: |                                                                          | 1619/? [1:40:22<00:00,  0.27it/s, v_num=cys0]train step 1620; scene = [['a270004198eddd9b'], ['cc306ca65c5dbc24'], ['1a5022549d2718ae'], ['cdde30f3352829cb']]; loss = 0.067003
+Epoch 0: |                                                                          | 1620/? [1:40:26<00:00,  0.27it/s, v_num=cys0]context = [[72, 79, 103, 136], [6, 15, 31, 60], [1, 48, 52, 80], [95, 135, 145, 162], [136, 166, 176, 206], [0, 7, 45, 61]]target = [[90, 120, 96, 123], [15, 25, 11, 32], [5, 49, 65, 22], [116, 131, 128, 139], [144, 141, 183, 190], [34, 48, 35, 12]]
+Epoch 0: |                                                                          | 1629/? [1:40:59<00:00,  0.27it/s, v_num=cys0]train step 1630; scene = [['22c6139e04c88c0c']]; loss = 0.043042
+Epoch 0: |                                                                          | 1630/? [1:41:03<00:00,  0.27it/s, v_num=cys0]context = [[5, 7, 10, 22, 36, 52], [7, 11, 21, 27, 46, 72], [0, 50, 64, 79, 88, 90], [4, 24, 28, 39, 44, 49]]target = [[41, 40, 13, 50, 28, 25], [39, 20, 47, 50, 25, 55], [75, 69, 56, 70, 5, 13], [27, 35, 24, 13, 14, 40]]
+Epoch 0: |                                                                          | 1639/? [1:41:37<00:00,  0.27it/s, v_num=cys0]train step 1640; scene = [['62ce0e4e9490317d']]; loss = 0.037701
+Epoch 0: |                                                                          | 1640/? [1:41:41<00:00,  0.27it/s, v_num=cys0]context = [[4, 6, 8, 15, 23, 24, 26, 32, 35, 39, 50, 55], [0, 10, 22, 26, 27, 29, 35, 42, 60, 66, 73, 77]]target = [[41, 43, 30, 44, 33, 16, 39, 51, 12, 54, 53, 29], [4, 16, 21, 43, 47, 46, 44, 53, 45, 24, 51, 50]]
+Epoch 0: |                                                                          | 1649/? [1:42:14<00:00,  0.27it/s, v_num=cys0]train step 1650; scene = [['bad5f833677cd32f'], ['2cf78f93fd569aa1']]; loss = 0.068326
+Epoch 0: |                                                                          | 1650/? [1:42:18<00:00,  0.27it/s, v_num=cys0]context = [[9, 71], [34, 102], [46, 110], [10, 61], [137, 186], [1, 52], [128, 195], [10, 100], [198, 258], [16, 67], [12, 66], [5, 66]]target = [[26, 42], [83, 46], [67, 48], [14, 57], [154, 155], [49, 29], [155, 140], [41, 56], [236, 235], [32, 61], [50, 32], [35, 61]]
+Epoch 0: |                                                                          | 1659/? [1:42:51<00:00,  0.27it/s, v_num=cys0]train step 1660; scene = [['1f3f484a027f93d9'], ['12e85fb6e140ee85'], ['197edde42c1eaac3'], ['59058fb21817aa6b']]; loss = 0.057813
+Epoch 0: |                                                                          | 1660/? [1:42:55<00:00,  0.27it/s, v_num=cys0]context = [[24, 30, 39, 55, 81, 105], [182, 183, 195, 207, 215, 233], [25, 65, 77, 78, 93, 105], [20, 40, 46, 77, 82, 103]]target = [[48, 70, 100, 51, 59, 97], [197, 190, 194, 208, 198, 218], [44, 91, 86, 70, 101, 30], [65, 21, 102, 25, 30, 54]]
+Epoch 0: |                                                                          | 1669/? [1:43:28<00:00,  0.27it/s, v_num=cys0]train step 1670; scene = [['25aff15f6c54558b']]; loss = 0.042839
+Epoch 0: |                                                                          | 1670/? [1:43:32<00:00,  0.27it/s, v_num=cys0]context = [[154, 158, 162, 163, 165, 191, 194, 196, 201, 203, 207, 211], [137, 138, 151, 153, 161, 165, 166, 167, 185, 186, 197, 203]]target = [[203, 173, 193, 192, 207, 165, 187, 157, 174, 190, 182, 176], [165, 166, 189, 176, 201, 140, 139, 163, 184, 169, 186, 157]]
+Epoch 0: |                                                                          | 1679/? [1:44:06<00:00,  0.27it/s, v_num=cys0]train step 1680; scene = [['c513a3c2f59aa548'], ['493538f3442cb9fd']]; loss = 0.049257
+Epoch 0: |                                                                          | 1680/? [1:44:09<00:00,  0.27it/s, v_num=cys0]context = [[8, 9, 12, 14, 16, 17, 18, 20, 23, 25, 29, 36, 38, 46, 52, 63, 64, 78, 82, 90, 95, 97, 102, 105]]target = [[19, 24, 49, 13, 77, 40, 51, 54, 85, 10, 74, 88, 50, 94, 78, 71, 68, 87, 44, 95, 63, 81, 82, 96]]
+Epoch 0: |                                                                          | 1689/? [1:44:42<00:00,  0.27it/s, v_num=cys0]train step 1690; scene = [['da5de01cf7e41541']]; loss = 0.052209
+Epoch 0: |                                                                          | 1690/? [1:44:46<00:00,  0.27it/s, v_num=cys0]context = [[166, 168, 170, 180, 181, 183, 184, 191, 219, 234, 239, 240], [52, 57, 62, 79, 87, 99, 104, 107, 111, 115, 118, 124]]target = [[226, 177, 191, 171, 214, 212, 213, 234, 229, 186, 238, 219], [60, 111, 110, 109, 72, 56, 96, 101, 61, 66, 118, 69]]
+Epoch 0: |                                                                          | 1699/? [1:45:19<00:00,  0.27it/s, v_num=cys0]train step 1700; scene = [['3e763e3c28e87eed']]; loss = 0.025079
+Epoch 0: |                                                                          | 1700/? [1:45:22<00:00,  0.27it/s, v_num=cys0]context = [[45, 50, 54, 60, 92, 105, 108, 110], [0, 8, 30, 31, 41, 42, 43, 45], [22, 28, 31, 33, 41, 59, 72, 99]]target = [[71, 52, 86, 64, 72, 69, 73, 59], [14, 36, 2, 1, 35, 4, 16, 20], [74, 61, 93, 94, 31, 42, 26, 28]]
+Epoch 0: |                                                                          | 1709/? [1:45:54<00:00,  0.27it/s, v_num=cys0]train step 1710; scene = [['3727bb4b44708f89']]; loss = 0.092673
+Epoch 0: |                                                                          | 1710/? [1:45:58<00:00,  0.27it/s, v_num=cys0]context = [[59, 62, 71, 76, 105, 110, 112, 114], [1, 2, 12, 23, 30, 34, 44, 48], [75, 78, 86, 87, 90, 114, 117, 122]]target = [[104, 100, 89, 101, 94, 102, 96, 84], [4, 24, 32, 21, 39, 26, 25, 20], [89, 94, 76, 90, 85, 96, 83, 77]]
+Epoch 0: |                                                                          | 1719/? [1:46:31<00:00,  0.27it/s, v_num=cys0]train step 1720; scene = [['d938c73738634cbf'], ['d7f3b4e12f3f3af6'], ['1a17332c1d690519']]; loss = 0.042639
+Epoch 0: |                                                                          | 1720/? [1:46:35<00:00,  0.27it/s, v_num=cys0]context = [[26, 28, 33, 44, 45, 49, 55, 57, 58, 59, 67, 72, 75, 79, 82, 94, 98, 100, 113, 115, 117, 118, 122, 123]]target = [[107, 48, 44, 86, 57, 93, 73, 51, 58, 88, 78, 59, 33, 99, 100, 87, 62, 38, 36, 84, 119, 122, 91, 74]]
+Epoch 0: |                                                                          | 1729/? [1:47:09<00:00,  0.27it/s, v_num=cys0]train step 1730; scene = [['2d9ea631e6423141'], ['b0c6597c77c51a8c']]; loss = 0.053091
+Epoch 0: |                                                                          | 1730/? [1:47:12<00:00,  0.27it/s, v_num=cys0]context = [[87, 105, 125, 127, 131, 138, 143, 150], [1, 14, 30, 33, 45, 48, 57, 66], [8, 18, 21, 26, 59, 92, 93, 94]]target = [[89, 146, 124, 104, 103, 96, 110, 143], [29, 33, 7, 12, 20, 27, 53, 51], [57, 81, 93, 74, 29, 30, 33, 60]]
+Epoch 0: |                                                                          | 1739/? [1:47:46<00:00,  0.27it/s, v_num=cys0]train step 1740; scene = [['56863fb499e2ff9a'], ['f5a3eca31fcdabf8'], ['946597adeb926d39'], ['23eecb3bfb301179'], ['4d999bce04f7516a'], ['b1f32a7b0a8e25ce'], ['72ac5af57c88795c'], ['cc45417d1c5dab8d']]; loss = 0.079800
+Epoch 0: |                                                                          | 1740/? [1:47:50<00:00,  0.27it/s, v_num=cys0]context = [[1, 4, 8, 12, 19, 26, 28, 29, 39, 42, 47, 50, 55, 66, 67, 68, 70, 76, 83, 85, 91, 92, 97, 98]]target = [[72, 40, 83, 3, 21, 13, 14, 61, 46, 53, 36, 57, 6, 10, 12, 37, 7, 22, 19, 78, 56, 25, 54, 80]]
+Epoch 0: |                                                                          | 1749/? [1:48:24<00:00,  0.27it/s, v_num=cys0]train step 1750; scene = [['7351b1a8a7405871'], ['edacb8db81943446']]; loss = 0.064277
+Epoch 0: |                                                                          | 1750/? [1:48:27<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 1750; scene = ['3b273cb40c55db95'];
+target intrinsic:  tensor(1.0504, device='cuda:0') tensor(1.0506, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(1.0343, device='cuda:0') tensor(1.0299, device='cuda:0')
+[2026-02-25 09:21:06,680][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1750/? [1:48:28<00:00,  0.27it/s, v_num=cys0]context = [[6, 12, 18, 21, 26, 27, 30, 33, 34, 35, 50, 55], [2, 5, 11, 19, 32, 36, 39, 44, 45, 46, 47, 56]]target = [[43, 26, 45, 30, 52, 34, 39, 12, 21, 11, 37, 19], [8, 19, 54, 25, 33, 4, 10, 55, 46, 34, 52, 3]]
+Epoch 0: |                                                                          | 1759/? [1:49:01<00:00,  0.27it/s, v_num=cys0]train step 1760; scene = [['616d045c5963ecb2'], ['4bbb9fadc993969a'], ['c8b833961ffd8ac5'], ['321addf5693a66aa']]; loss = 0.047790
+Epoch 0: |                                                                          | 1760/? [1:49:05<00:00,  0.27it/s, v_num=cys0]context = [[111, 116, 130, 132, 142, 146, 147, 158, 160, 164, 176, 181], [19, 20, 21, 24, 28, 31, 47, 58, 61, 70, 71, 77]]target = [[125, 122, 167, 145, 144, 152, 130, 121, 153, 165, 113, 139], [70, 73, 42, 59, 66, 23, 39, 72, 22, 60, 26, 25]]
+Epoch 0: |                                                                          | 1769/? [1:49:38<00:00,  0.27it/s, v_num=cys0]train step 1770; scene = [['8eb7ab71f603c9d4'], ['cc966a9b2af2232f']]; loss = 0.058732
+Epoch 0: |                                                                          | 1770/? [1:49:41<00:00,  0.27it/s, v_num=cys0]context = [[102, 104, 107, 108, 112, 122, 123, 126, 129, 133, 136, 142, 146, 147, 150, 161, 162, 163, 164, 184, 189, 191, 193, 199]]target = [[120, 136, 118, 149, 133, 122, 193, 137, 187, 174, 159, 156, 194, 170, 121, 189, 103, 178, 197, 196, 129, 195, 115, 181]]
+Epoch 0: |                                                                          | 1779/? [1:50:15<00:00,  0.27it/s, v_num=cys0]train step 1780; scene = [['bb28a1c09df8a484']]; loss = 0.040358
+Epoch 0: |                                                                          | 1780/? [1:50:19<00:00,  0.27it/s, v_num=cys0]context = [[11, 70], [25, 75], [78, 152], [51, 119], [24, 81], [136, 193], [20, 96], [30, 109], [7, 87], [58, 129], [66, 145], [27, 74]]target = [[17, 37], [64, 73], [135, 104], [69, 118], [42, 58], [145, 168], [30, 83], [99, 43], [42, 16], [61, 111], [134, 98], [55, 71]]
+Epoch 0: |                                                                          | 1789/? [1:50:52<00:00,  0.27it/s, v_num=cys0]train step 1790; scene = [['48c2b63b452555af'], ['1aa5fdc9ff855ee2']]; loss = 0.067603
+Epoch 0: |                                                                          | 1790/? [1:50:55<00:00,  0.27it/s, v_num=cys0]context = [[32, 36, 37, 42, 43, 52, 63, 67, 68, 73, 74, 90], [27, 40, 53, 55, 57, 63, 68, 79, 80, 81, 100, 102]]target = [[70, 67, 89, 83, 52, 35, 74, 73, 84, 44, 54, 33], [44, 60, 41, 98, 100, 51, 46, 99, 47, 92, 37, 31]]
+Epoch 0: |                                                                          | 1799/? [1:51:29<00:00,  0.27it/s, v_num=cys0]train step 1800; scene = [['b41ece74dd5aa87d'], ['e620345f5468d2e3']]; loss = 0.030837
+Epoch 0: |                                                                          | 1800/? [1:51:33<00:00,  0.27it/s, v_num=cys0]context = [[2, 47], [107, 163], [22, 107], [67, 146], [82, 130], [87, 151], [59, 133], [191, 242], [121, 203], [113, 167], [106, 159], [27, 94]]target = [[37, 42], [161, 109], [92, 85], [142, 139], [116, 104], [106, 134], [116, 87], [201, 206], [142, 129], [159, 155], [116, 134], [57, 36]]
+[2026-02-25 09:24:15,287][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1809/? [1:52:04<00:00,  0.27it/s, v_num=cys0]train step 1810; scene = [['9e1b58ec292d6690'], ['63f6e847b2e2a2e7'], ['aa570297256d99b6'], ['02195bc2cdded89e'], ['8839e65c6f6bb394'], ['c6bd8d3b3367444f']]; loss = 0.074992
+Epoch 0: |                                                                          | 1810/? [1:52:08<00:00,  0.27it/s, v_num=cys0]context = [[3, 14, 18, 21, 23, 31, 32, 34, 36, 63, 70, 78], [24, 32, 33, 47, 62, 69, 70, 73, 80, 85, 86, 88]]target = [[5, 7, 41, 28, 48, 67, 12, 68, 47, 26, 10, 31], [25, 68, 40, 79, 82, 58, 26, 28, 36, 48, 66, 29]]
+Epoch 0: |                                                                          | 1819/? [1:52:42<00:00,  0.27it/s, v_num=cys0]train step 1820; scene = [['8ad3918394be8d98']]; loss = 0.064304
+Epoch 0: |                                                                          | 1820/? [1:52:45<00:00,  0.27it/s, v_num=cys0]context = [[0, 4, 7, 9, 13, 15, 17, 35, 37, 61, 69, 72], [5, 23, 24, 25, 26, 31, 50, 56, 60, 68, 74, 76]]target = [[45, 49, 51, 46, 6, 14, 52, 22, 65, 69, 10, 37], [9, 37, 72, 6, 28, 14, 64, 47, 15, 16, 55, 33]]
+Epoch 0: |                                                                          | 1829/? [1:53:18<00:00,  0.27it/s, v_num=cys0]train step 1830; scene = [['af4565fb713ed79f']]; loss = 0.030642
+Epoch 0: |                                                                          | 1830/? [1:53:22<00:00,  0.27it/s, v_num=cys0]context = [[133, 138, 168, 169, 171, 180, 190, 192, 196, 203, 204, 213], [81, 82, 84, 87, 92, 93, 95, 97, 105, 110, 118, 131]]target = [[147, 196, 136, 153, 201, 185, 205, 161, 169, 149, 210, 189], [114, 103, 99, 119, 107, 110, 88, 124, 86, 122, 126, 120]]
+Epoch 0: |                                                                          | 1839/? [1:53:56<00:00,  0.27it/s, v_num=cys0]train step 1840; scene = [['c690f7b93b2b75c5'], ['f0dbc908f65693ae'], ['ee81b7d704303a2c'], ['a2a3a75cafb630f5']]; loss = 0.078360
+Epoch 0: |                                                                          | 1840/? [1:54:00<00:00,  0.27it/s, v_num=cys0]context = [[95, 106, 107, 115, 116, 118, 120, 126, 128, 130, 133, 139, 148, 149, 150, 153, 159, 163, 173, 178, 182, 183, 184, 192]]target = [[167, 153, 132, 96, 110, 126, 111, 129, 162, 144, 185, 176, 106, 97, 156, 108, 177, 172, 148, 157, 133, 137, 149, 141]]
+Epoch 0: |                                                                          | 1849/? [1:54:33<00:00,  0.27it/s, v_num=cys0]train step 1850; scene = [['a1fa4bf737381a53'], ['2d7243929156069e'], ['201287bd653fa906'], ['2d7243929156069e'], ['d25d6a9b1356b134'], ['4a0f95a3db913b56']]; loss = 0.072569
+Epoch 0: |                                                                          | 1850/? [1:54:36<00:00,  0.27it/s, v_num=cys0]context = [[30, 32, 34, 52, 72, 85, 90, 108], [9, 25, 30, 33, 41, 46, 48, 84], [51, 72, 76, 82, 83, 99, 109, 115]]target = [[46, 41, 107, 87, 40, 64, 57, 53], [35, 56, 83, 38, 62, 21, 53, 37], [109, 111, 114, 53, 86, 67, 91, 104]]
+Epoch 0: |                                                                          | 1859/? [1:55:08<00:00,  0.27it/s, v_num=cys0]train step 1860; scene = [['082779aebfbf4a46']]; loss = 0.036747
+Epoch 0: |                                                                          | 1860/? [1:55:12<00:00,  0.27it/s, v_num=cys0]context = [[3, 17, 39, 49], [112, 121, 139, 157], [128, 136, 147, 178], [2, 10, 33, 87], [25, 50, 81, 98], [28, 48, 69, 113]]target = [[8, 40, 43, 26], [113, 124, 154, 133], [156, 164, 137, 168], [62, 24, 86, 11], [68, 70, 78, 73], [73, 100, 109, 29]]
+Epoch 0: |                                                                          | 1869/? [1:55:44<00:00,  0.27it/s, v_num=cys0]train step 1870; scene = [['01c143c81a4d2145'], ['9b0b82db99ff360a'], ['80f5da17b4119bbf'], ['8c5dce5d79b3d2aa'], ['d3a0a89d951a6101'], ['da1f9f2859b59142'], ['20b791440b7ec0b5'], ['00703cbf7531ef11'], ['4aef5d4b9287e08a'], ['27ed8e7077af6540'], ['04e7f97215df7078'], ['08735c801ab7efb5']]; loss = 0.098273
+Epoch 0: |                                                                          | 1870/? [1:55:48<00:00,  0.27it/s, v_num=cys0]context = [[41, 44, 47, 48, 50, 51, 56, 67, 72, 81, 82, 83, 84, 85, 89, 90, 97, 105, 107, 110, 115, 125, 135, 138]]target = [[79, 108, 90, 99, 85, 65, 59, 69, 73, 126, 131, 100, 114, 52, 87, 47, 123, 54, 43, 137, 94, 88, 70, 75]]
+Epoch 0: |                                                                          | 1879/? [1:56:22<00:00,  0.27it/s, v_num=cys0]train step 1880; scene = [['42b2dabdc5ea4d93']]; loss = 0.032639
+Epoch 0: |                                                                          | 1880/? [1:56:25<00:00,  0.27it/s, v_num=cys0]context = [[17, 29, 31, 34, 36, 39, 48, 59, 64, 66, 67, 79], [42, 47, 51, 56, 57, 61, 64, 70, 74, 83, 92, 93]]target = [[53, 20, 25, 74, 39, 30, 21, 59, 50, 66, 40, 44], [43, 84, 45, 72, 82, 85, 56, 73, 92, 76, 75, 46]]
+Epoch 0: |                                                                          | 1889/? [1:56:59<00:00,  0.27it/s, v_num=cys0]train step 1890; scene = [['ed4855892fd5fa4a'], ['9683be82b6c1e851']]; loss = 0.056587
+Epoch 0: |                                                                          | 1890/? [1:57:03<00:00,  0.27it/s, v_num=cys0]context = [[22, 37, 39, 49, 50, 52, 54, 56, 59, 61, 72, 74, 84, 88, 93, 94, 97, 102, 104, 105, 108, 109, 112, 119]]target = [[59, 53, 102, 34, 27, 44, 100, 45, 32, 65, 36, 42, 87, 82, 85, 26, 80, 60, 90, 113, 86, 78, 40, 68]]
+Epoch 0: |                                                                          | 1899/? [1:57:37<00:00,  0.27it/s, v_num=cys0]train step 1900; scene = [['1844e80dafb62927'], ['b43f32da9a7800ee'], ['11094b4b71a9ab00'], ['0f5daf32c2e8fefd'], ['f8e7bd9e403fc04a'], ['b11afc47f2feb21c'], ['5994966dd0897ead'], ['6d2aae2a7dd35e14'], ['e545aecc18cfa501'], ['82d896f5142ee6dd'], ['140b10a4f6bb5aa5'], ['2ec8edfc07c8841f']]; loss = 0.097580
+Epoch 0: |                                                                          | 1900/? [1:57:40<00:00,  0.27it/s, v_num=cys0]context = [[1, 20, 26, 32, 34, 53], [83, 87, 95, 106, 125, 139], [0, 13, 29, 39, 46, 77], [90, 122, 126, 130, 159, 167]]target = [[36, 12, 39, 17, 4, 7], [92, 86, 107, 112, 135, 137], [32, 69, 36, 28, 50, 38], [132, 161, 105, 165, 142, 102]]
+Epoch 0: |                                                                          | 1909/? [1:58:14<00:00,  0.27it/s, v_num=cys0]train step 1910; scene = [['50f7971dda42084f'], ['159a472b24f1c395'], ['04a191933c5b05ec']]; loss = 0.086071
+Epoch 0: |                                                                          | 1910/? [1:58:18<00:00,  0.27it/s, v_num=cys0]context = [[74, 78, 86, 94, 96, 101, 104, 107, 108, 109, 115, 117, 119, 131, 133, 135, 137, 144, 151, 152, 155, 158, 161, 171]]target = [[108, 142, 163, 121, 87, 111, 149, 96, 120, 145, 97, 122, 147, 101, 123, 95, 94, 132, 86, 89, 99, 140, 104, 136]]
+Epoch 0: |                                                                          | 1919/? [1:58:51<00:00,  0.27it/s, v_num=cys0]train step 1920; scene = [['bf4bbf858718317d'], ['39ed8d2efe760b94'], ['d9644f4985e51a1f']]; loss = 0.046443
+Epoch 0: |                                                                          | 1920/? [1:58:54<00:00,  0.27it/s, v_num=cys0]context = [[9, 79], [15, 63], [28, 73], [101, 167], [5, 70], [53, 102], [2, 75], [86, 135], [1, 58], [5, 84], [1, 76], [135, 208]]target = [[43, 20], [62, 45], [70, 64], [114, 144], [60, 52], [93, 101], [13, 68], [120, 89], [30, 23], [43, 16], [57, 14], [194, 205]]
+Epoch 0: |                                                                          | 1929/? [1:59:26<00:00,  0.27it/s, v_num=cys0]train step 1930; scene = [['6efa62598fececd0'], ['b25a0f4ffca51d79'], ['90dbfac63e6b89be'], ['08a366317a388734']]; loss = 0.075953
+Epoch 0: |                                                                          | 1930/? [1:59:29<00:00,  0.27it/s, v_num=cys0]context = [[3, 19, 23, 28, 32, 40, 48, 50, 53, 60, 63, 64, 66, 70, 74, 78, 80, 82, 87, 93, 96, 97, 98, 100]]target = [[63, 24, 82, 55, 68, 93, 95, 7, 53, 65, 76, 35, 83, 11, 22, 34, 52, 16, 61, 92, 26, 47, 32, 46]]
+Epoch 0: |                                                                          | 1939/? [2:00:03<00:00,  0.27it/s, v_num=cys0]train step 1940; scene = [['1236365ec263ad76']]; loss = 0.045517
+Epoch 0: |                                                                          | 1940/? [2:00:07<00:00,  0.27it/s, v_num=cys0]context = [[10, 11, 16, 23, 26, 33, 34, 43, 44, 48, 54, 58, 70, 73, 75, 78, 82, 86, 89, 91, 102, 105, 106, 107]]target = [[88, 68, 24, 39, 22, 100, 19, 44, 57, 94, 41, 37, 53, 69, 96, 63, 81, 104, 50, 64, 101, 15, 40, 26]]
+Epoch 0: |                                                                          | 1949/? [2:00:41<00:00,  0.27it/s, v_num=cys0]train step 1950; scene = [['2bc8b64aafc5870c'], ['1202c32d91ad3ee3'], ['300571576edc008c']]; loss = 0.104005
+Epoch 0: |                                                                          | 1950/? [2:00:44<00:00,  0.27it/s, v_num=cys0]context = [[11, 14, 25, 42, 47, 49, 58, 66, 68, 69, 79, 81], [39, 55, 56, 60, 63, 70, 74, 86, 93, 96, 99, 104]]target = [[25, 55, 33, 13, 26, 62, 22, 52, 32, 67, 30, 40], [95, 42, 43, 41, 59, 96, 102, 88, 48, 66, 100, 40]]
+Epoch 0: |                                                                          | 1959/? [2:01:17<00:00,  0.27it/s, v_num=cys0]train step 1960; scene = [['3ded3e3c1fe76ee3']]; loss = 0.042777
+Epoch 0: |                                                                          | 1960/? [2:01:21<00:00,  0.27it/s, v_num=cys0]context = [[109, 112, 117, 120, 130, 137, 142, 144, 145, 150, 151, 152, 155, 159, 160, 162, 166, 171, 175, 179, 194, 195, 196, 206]]target = [[168, 171, 125, 169, 131, 150, 188, 139, 145, 152, 126, 184, 205, 142, 119, 196, 148, 155, 185, 197, 154, 203, 143, 147]]
+Epoch 0: |                                                                          | 1969/? [2:01:55<00:00,  0.27it/s, v_num=cys0]train step 1970; scene = [['3b21f48e23e4917f']]; loss = 0.042043
+Epoch 0: |                                                                          | 1970/? [2:01:58<00:00,  0.27it/s, v_num=cys0]context = [[13, 15, 19, 21, 30, 33, 39, 49, 54, 57, 72, 85], [18, 26, 35, 36, 39, 46, 50, 60, 70, 71, 74, 80]]target = [[39, 30, 38, 35, 26, 78, 75, 56, 18, 44, 77, 71], [22, 24, 70, 65, 49, 19, 62, 59, 71, 73, 51, 39]]
+Epoch 0: |                                                                          | 1979/? [2:02:32<00:00,  0.27it/s, v_num=cys0]train step 1980; scene = [['723db63d24c84d1d']]; loss = 0.040032
+Epoch 0: |                                                                          | 1980/? [2:02:36<00:00,  0.27it/s, v_num=cys0]context = [[114, 115, 120, 122, 123, 125, 126, 140, 141, 145, 147, 161, 170, 179, 180, 181, 193, 195, 196, 197, 199, 203, 207, 211]]target = [[201, 148, 171, 190, 135, 143, 189, 141, 205, 151, 187, 206, 183, 144, 200, 147, 181, 184, 120, 207, 152, 128, 170, 204]]
+Epoch 0: |                                                                          | 1989/? [2:03:10<00:00,  0.27it/s, v_num=cys0]train step 1990; scene = [['12eb36bba5c89eeb'], ['40c5311e3b3accef'], ['06d2876e8c40a3b6'], ['2da17464ef895b63'], ['21951a6ae1c4b225'], ['ab78b3eb64029b73'], ['f6ef16edbf87f358'], ['66e4f3268dafe823']]; loss = 0.089924
+Epoch 0: |                                                                          | 1990/? [2:03:13<00:00,  0.27it/s, v_num=cys0]context = [[6, 9, 11, 12, 24, 34, 39, 44, 45, 53, 55, 59, 61, 66, 70, 71, 76, 78, 79, 82, 85, 87, 92, 103]]target = [[32, 63, 22, 65, 39, 36, 57, 91, 64, 13, 66, 60, 14, 71, 26, 41, 23, 101, 29, 93, 92, 90, 98, 73]]
+Epoch 0: |                                                                          | 1999/? [2:03:46<00:00,  0.27it/s, v_num=cys0]train step 2000; scene = [['1ac1373478877088']]; loss = 0.058416
+Epoch 0: |                                                                          | 2000/? [2:03:50<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 2000; scene = ['be75142d4652fe3e'];
+target intrinsic:  tensor(0.9402, device='cuda:0') tensor(0.9404, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8886, device='cuda:0') tensor(0.8881, device='cuda:0')
+[2026-02-25 09:36:29,623][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2000/? [2:03:51<00:00,  0.27it/s, v_num=cys0]context = [[16, 17, 18, 27, 35, 40, 41, 47, 48, 51, 52, 53, 54, 56, 79, 80, 87, 89, 92, 96, 100, 101, 106, 113]]target = [[102, 53, 81, 57, 21, 101, 52, 64, 43, 58, 59, 94, 46, 108, 32, 79, 36, 104, 26, 106, 31, 29, 60, 41]]
+[2026-02-25 09:36:33,850][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2009/? [2:04:28<00:00,  0.27it/s, v_num=cys0]train step 2010; scene = [['e2f2a27bfce53270']]; loss = 0.045383
+Epoch 0: |                                                                          | 2010/? [2:04:32<00:00,  0.27it/s, v_num=cys0]context = [[32, 34, 44, 61, 73, 79], [66, 91, 112, 119, 124, 128], [52, 79, 88, 91, 97, 100], [65, 83, 97, 105, 110, 128]]target = [[68, 72, 41, 73, 74, 40], [74, 67, 110, 72, 105, 98], [93, 66, 74, 77, 83, 65], [81, 72, 117, 80, 83, 114]]
+Epoch 0: |                                                                          | 2019/? [2:05:04<00:00,  0.27it/s, v_num=cys0]train step 2020; scene = [['71f80e4a8d79031c'], ['aec60a6dd8090690'], ['475ee0875ccb635a']]; loss = 0.054234
+Epoch 0: |                                                                          | 2020/? [2:05:08<00:00,  0.27it/s, v_num=cys0]context = [[1, 4, 12, 16, 21, 23, 29, 40, 49, 54, 58, 60, 62, 74, 75, 79, 81, 82, 85, 91, 92, 94, 97, 98]]target = [[52, 69, 55, 22, 50, 24, 41, 46, 80, 58, 32, 92, 54, 81, 38, 23, 3, 29, 97, 40, 25, 39, 78, 73]]
+Epoch 0: |                                                                          | 2029/? [2:05:42<00:00,  0.27it/s, v_num=cys0]train step 2030; scene = [['60b6914bafa3c3f2']]; loss = 0.063915
+Epoch 0: |                                                                          | 2030/? [2:05:46<00:00,  0.27it/s, v_num=cys0]context = [[45, 87, 105], [54, 118, 143], [79, 120, 143], [186, 243, 262], [14, 67, 94], [27, 68, 76], [13, 28, 58], [1, 69, 80]]target = [[68, 84, 80], [134, 122, 96], [131, 97, 136], [244, 198, 209], [34, 77, 22], [70, 43, 37], [43, 32, 33], [15, 54, 19]]
+Epoch 0: |                                                                          | 2039/? [2:06:19<00:00,  0.27it/s, v_num=cys0]train step 2040; scene = [['8e94de9c1bc0732c'], ['0bb8a7807f7095fd']]; loss = 0.034499
+Epoch 0: |                                                                          | 2040/? [2:06:22<00:00,  0.27it/s, v_num=cys0]context = [[25, 79, 84, 91], [12, 32, 41, 66], [10, 17, 31, 64], [69, 71, 109, 114], [31, 49, 67, 84], [14, 27, 46, 84]]target = [[48, 80, 51, 88], [16, 41, 21, 33], [63, 61, 13, 38], [73, 107, 79, 90], [41, 35, 60, 56], [29, 80, 26, 25]]
+Epoch 0: |                                                                          | 2049/? [2:06:56<00:00,  0.27it/s, v_num=cys0]train step 2050; scene = [['c4e12df63403eadf'], ['8359c9726f078a38'], ['e0b75e74fdeffde9']]; loss = 0.072549
+Epoch 0: |                                                                          | 2050/? [2:06:59<00:00,  0.27it/s, v_num=cys0]context = [[152, 157, 160, 167, 191, 197], [71, 81, 105, 126, 127, 139], [10, 12, 47, 54, 66, 83], [85, 88, 106, 117, 127, 136]]target = [[167, 195, 187, 184, 157, 176], [120, 107, 92, 124, 115, 110], [14, 13, 80, 40, 24, 47], [93, 96, 97, 132, 102, 131]]
+Epoch 0: |                                                                          | 2059/? [2:07:32<00:00,  0.27it/s, v_num=cys0]train step 2060; scene = [['22062ed897320134'], ['08e4f6a5098b0d3a'], ['d3057752d15cc3ed']]; loss = 0.099569
+Epoch 0: |                                                                          | 2060/? [2:07:35<00:00,  0.27it/s, v_num=cys0]context = [[164, 195, 232], [18, 42, 81], [5, 55, 93], [175, 199, 220], [6, 42, 65], [2, 7, 60], [149, 190, 195], [33, 42, 82]]target = [[209, 178, 225], [53, 24, 75], [32, 35, 24], [201, 183, 204], [50, 59, 38], [5, 37, 42], [187, 161, 189], [44, 78, 62]]
+Epoch 0: |                                                                          | 2069/? [2:08:09<00:00,  0.27it/s, v_num=cys0]train step 2070; scene = [['85ef0eb4a42e3425'], ['703430ad773c95bc'], ['4fd9d45647d536e5']]; loss = 0.126629
+Epoch 0: |                                                                          | 2070/? [2:08:13<00:00,  0.27it/s, v_num=cys0]context = [[78, 84, 92, 97, 108, 109, 115, 124, 138, 150, 158, 159], [42, 43, 53, 57, 69, 71, 77, 83, 84, 86, 88, 92]]target = [[127, 113, 118, 102, 153, 144, 132, 139, 125, 129, 84, 79], [57, 81, 79, 87, 72, 61, 48, 88, 54, 84, 44, 65]]
+Epoch 0: |                                                                          | 2079/? [2:08:46<00:00,  0.27it/s, v_num=cys0]train step 2080; scene = [['968b857dda7e955a'], ['8e713ad26d00feac']]; loss = 0.041911
+Epoch 0: |                                                                          | 2080/? [2:08:50<00:00,  0.27it/s, v_num=cys0]context = [[33, 57, 88, 112], [20, 29, 43, 76], [20, 56, 96, 110], [3, 13, 50, 53], [44, 50, 64, 108], [65, 92, 120, 152]]target = [[101, 99, 69, 88], [37, 25, 36, 30], [99, 72, 91, 75], [14, 32, 43, 39], [105, 93, 56, 106], [69, 98, 70, 122]]
+Epoch 0: |                                                                          | 2089/? [2:09:23<00:00,  0.27it/s, v_num=cys0]train step 2090; scene = [['0565bd311bf73bbb'], ['eceeb7a49f302da9'], ['2073f379b98e47e8'], ['c5536f755d325407'], ['06ed257e33ae67f5'], ['b12f64d6002ec745']]; loss = 0.063990
+Epoch 0: |                                                                          | 2090/? [2:09:27<00:00,  0.27it/s, v_num=cys0]context = [[14, 20, 22, 23, 24, 26, 36, 41, 74, 81, 83, 87], [2, 10, 11, 16, 25, 26, 34, 35, 41, 49, 62, 63]]target = [[16, 23, 39, 54, 53, 59, 48, 70, 20, 63, 50, 36], [60, 32, 15, 7, 5, 51, 40, 52, 10, 56, 22, 41]]
+Epoch 0: |                                                                          | 2099/? [2:10:01<00:00,  0.27it/s, v_num=cys0]train step 2100; scene = [['3101309753e2f063'], ['d55fbb3dcd24f08e']]; loss = 0.040141
+Epoch 0: |                                                                          | 2100/? [2:10:05<00:00,  0.27it/s, v_num=cys0]context = [[145, 173, 174, 182, 187, 197, 204, 211], [65, 79, 86, 88, 93, 112, 118, 122], [98, 107, 122, 145, 156, 159, 162, 163]]target = [[208, 167, 187, 146, 173, 180, 166, 172], [108, 72, 116, 104, 92, 115, 83, 68], [160, 128, 158, 103, 134, 125, 101, 109]]
+Epoch 0: |                                                                          | 2109/? [2:10:37<00:00,  0.27it/s, v_num=cys0]train step 2110; scene = [['4fd151a48542df52'], ['c53e0b350f04f159'], ['ec1dcf652eae675d'], ['f001500643d191d6']]; loss = 0.124596
+Epoch 0: |                                                                          | 2110/? [2:10:41<00:00,  0.27it/s, v_num=cys0]context = [[42, 45, 48, 60, 81, 84, 88, 90, 102, 104, 121, 130], [22, 29, 34, 38, 39, 44, 49, 50, 53, 56, 59, 74]]target = [[108, 62, 127, 115, 43, 114, 52, 51, 100, 126, 61, 87], [64, 41, 45, 69, 25, 27, 65, 43, 68, 71, 38, 35]]
+Epoch 0: |                                                                          | 2119/? [2:11:15<00:00,  0.27it/s, v_num=cys0]train step 2120; scene = [['4f515c197a061a67'], ['0c7171edef36d44d'], ['ec5f9801aa83c8aa']]; loss = 0.065628
+Epoch 0: |                                                                          | 2120/? [2:11:19<00:00,  0.27it/s, v_num=cys0]context = [[124, 126, 135, 171], [32, 69, 83, 84], [30, 85, 88, 105], [27, 33, 42, 117], [2, 33, 53, 56], [24, 47, 86, 102]]target = [[165, 167, 132, 145], [80, 75, 52, 61], [73, 92, 100, 67], [87, 74, 112, 34], [45, 41, 49, 35], [80, 73, 95, 30]]
+Epoch 0: |                                                                          | 2129/? [2:11:51<00:00,  0.27it/s, v_num=cys0]train step 2130; scene = [['2dd8ae2b71457753'], ['3670e4d9d26e7534']]; loss = 0.053935
+Epoch 0: |                                                                          | 2130/? [2:11:54<00:00,  0.27it/s, v_num=cys0]context = [[24, 25, 27, 28, 32, 34, 36, 39, 48, 50, 55, 61, 62, 63, 64, 67, 77, 87, 102, 106, 110, 116, 117, 121]]target = [[45, 42, 59, 107, 40, 101, 31, 75, 70, 54, 81, 53, 83, 73, 64, 100, 51, 37, 47, 117, 60, 43, 48, 90]]
+Epoch 0: |                                                                          | 2139/? [2:12:28<00:00,  0.27it/s, v_num=cys0]train step 2140; scene = [['158ecf8d21e5af57'], ['4dfd4268f80ef274'], ['04c4eef824be2a53']]; loss = 0.046748
+Epoch 0: |                                                                          | 2140/? [2:12:32<00:00,  0.27it/s, v_num=cys0]context = [[131, 152, 157, 167, 188, 191], [139, 146, 149, 153, 171, 223], [126, 136, 156, 187, 188, 202], [32, 37, 41, 56, 82, 87]]target = [[189, 157, 139, 148, 184, 180], [151, 183, 209, 179, 185, 140], [159, 166, 155, 170, 182, 131], [78, 79, 34, 85, 50, 59]]
+Epoch 0: |                                                                          | 2149/? [2:13:05<00:00,  0.27it/s, v_num=cys0]train step 2150; scene = [['db62795c284bf764'], ['916b86f95631b480']]; loss = 0.060342
+Epoch 0: |                                                                          | 2150/? [2:13:09<00:00,  0.27it/s, v_num=cys0]context = [[4, 6, 16, 19, 21, 30, 35, 38, 41, 48, 51, 52, 54, 62, 63, 67, 69, 77, 80, 81, 86, 93, 95, 101]]target = [[56, 16, 36, 50, 25, 34, 65, 87, 40, 5, 42, 54, 18, 32, 68, 76, 98, 52, 37, 85, 41, 79, 13, 91]]
+Epoch 0: |                                                                          | 2159/? [2:13:43<00:00,  0.27it/s, v_num=cys0]train step 2160; scene = [['72077f7ff39fc73e'], ['9ae93c878c7dbe8a'], ['4bac79c3a17ed149'], ['208bd0af6bc8937b'], ['84e6b90f0c2567d8'], ['f4ba6d204cb14df7'], ['758770f2884e9a79'], ['36df585860d0ad88'], ['69a6e3951e138ca8'], ['5d986af113fbac56'], ['d46599d6e4a2b451'], ['ee020a8773034321']]; loss = 0.079676
+Epoch 0: |                                                                          | 2160/? [2:13:47<00:00,  0.27it/s, v_num=cys0]context = [[6, 8, 10, 50, 66, 87], [50, 69, 72, 95, 105, 111], [126, 165, 170, 172, 176, 211], [5, 9, 19, 42, 47, 50]]target = [[9, 68, 72, 19, 24, 67], [80, 85, 64, 70, 104, 91], [184, 172, 136, 189, 179, 175], [48, 14, 45, 22, 40, 16]]
+Epoch 0: |                                                                          | 2169/? [2:14:21<00:00,  0.27it/s, v_num=cys0]train step 2170; scene = [['e1d9afaa8899ee32'], ['1060ca933281b55c'], ['fbf253fca4a29e87'], ['5ff87250a0eb913b'], ['d02e6b104723d39a'], ['42b208082fce3bc2'], ['ee6e5709a57be759'], ['a191c34eb75bbaec'], ['0cbafecfcb0f7727'], ['e6461cee8a9474d5'], ['fda8bac8ddac590f'], ['465fa8314b741006']]; loss = 0.121923
+Epoch 0: |                                                                          | 2170/? [2:14:24<00:00,  0.27it/s, v_num=cys0]context = [[5, 20, 23, 37, 49, 52, 59, 76], [11, 28, 31, 35, 41, 43, 44, 57], [5, 15, 30, 36, 47, 59, 61, 62]]target = [[53, 50, 40, 18, 39, 33, 23, 17], [27, 38, 40, 28, 55, 45, 26, 25], [56, 9, 61, 23, 59, 30, 45, 35]]
+Epoch 0: |                                                                          | 2179/? [2:14:57<00:00,  0.27it/s, v_num=cys0]train step 2180; scene = [['91599919681fac69'], ['f9cece9ebde532d0']]; loss = 0.066935
+Epoch 0: |                                                                          | 2180/? [2:15:00<00:00,  0.27it/s, v_num=cys0]context = [[46, 63, 73, 75, 76, 83, 99, 102], [17, 31, 32, 39, 47, 52, 58, 68], [83, 91, 110, 113, 116, 117, 121, 133]]target = [[72, 92, 67, 101, 68, 55, 49, 74], [30, 31, 53, 26, 41, 35, 20, 62], [93, 131, 116, 96, 125, 107, 127, 109]]
+Epoch 0: |                                                                          | 2189/? [2:15:34<00:00,  0.27it/s, v_num=cys0]train step 2190; scene = [['4d48befa72535f0a']]; loss = 0.055863
+Epoch 0: |                                                                          | 2190/? [2:15:38<00:00,  0.27it/s, v_num=cys0]context = [[113, 123, 125, 126, 132, 133, 135, 138, 149, 150, 154, 160, 167, 172, 175, 191, 194, 197, 198, 200, 201, 205, 206, 210]]target = [[120, 130, 147, 187, 142, 115, 202, 171, 117, 163, 191, 170, 139, 169, 181, 155, 150, 141, 192, 146, 188, 174, 205, 179]]
+Epoch 0: |                                                                          | 2199/? [2:16:10<00:00,  0.27it/s, v_num=cys0]train step 2200; scene = [['76ce9bd95ed81200'], ['91ac7d7027dcc46d'], ['3c3e1619744887ca'], ['003d2563b3c1023e']]; loss = 0.056805
+Epoch 0: |                                                                          | 2200/? [2:16:14<00:00,  0.27it/s, v_num=cys0]context = [[11, 13, 17, 19, 42, 43, 50, 51, 58, 60, 61, 63, 65, 71, 74, 77, 82, 85, 89, 95, 99, 101, 103, 108]]target = [[61, 31, 48, 27, 88, 78, 23, 53, 24, 93, 63, 73, 20, 32, 55, 35, 29, 75, 21, 76, 49, 42, 28, 40]]
+[2026-02-25 09:48:56,864][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2209/? [2:16:55<00:00,  0.27it/s, v_num=cys0]train step 2210; scene = [['40873337bb9786ab'], ['4f354a704f70f53b'], ['57cea8a71cab48cf'], ['19c383ad9f21d0d3'], ['a98ec507d448dea2'], ['b663a474fd1e4ca4']]; loss = 0.075115
+Epoch 0: |                                                                          | 2210/? [2:16:57<00:00,  0.27it/s, v_num=cys0]context = [[9, 12, 24, 32, 37, 38, 44, 46, 51, 65, 71, 75], [87, 89, 90, 92, 104, 114, 118, 119, 120, 122, 143, 144]]target = [[61, 31, 40, 37, 52, 13, 14, 68, 57, 25, 70, 54], [95, 130, 128, 134, 91, 114, 133, 102, 129, 98, 138, 116]]
+Epoch 0: |                                                                          | 2219/? [2:17:31<00:00,  0.27it/s, v_num=cys0]train step 2220; scene = [['d85f0322f215aa54'], ['1318e656455d56d2'], ['e5f41c9021b0b7c1']]; loss = 0.059443
+Epoch 0: |                                                                          | 2220/? [2:17:35<00:00,  0.27it/s, v_num=cys0]context = [[12, 17, 27, 29, 36, 37, 44, 50, 52, 54, 55, 65, 75, 80, 81, 83, 84, 85, 86, 89, 91, 94, 95, 109]]target = [[83, 45, 76, 92, 104, 41, 30, 59, 79, 28, 86, 38, 85, 64, 66, 51, 106, 15, 89, 16, 46, 108, 99, 33]]
+Epoch 0: |                                                                          | 2229/? [2:18:08<00:00,  0.27it/s, v_num=cys0]train step 2230; scene = [['9af18a1c4c45a179']]; loss = 0.056686
+Epoch 0: |                                                                          | 2230/? [2:18:12<00:00,  0.27it/s, v_num=cys0]context = [[40, 43, 44, 61, 67, 70, 74, 77, 79, 85, 87, 92, 93, 98, 100, 103, 108, 115, 117, 123, 128, 134, 135, 137]]target = [[119, 88, 102, 62, 79, 131, 53, 122, 46, 128, 130, 59, 54, 76, 71, 57, 107, 126, 93, 65, 72, 115, 69, 66]]
+Epoch 0: |                                                                          | 2239/? [2:18:45<00:00,  0.27it/s, v_num=cys0]train step 2240; scene = [['e454026f5348630e'], ['d6a1f3e13c45df99'], ['4dd9c5fab7e6ec75'], ['352d2bdc1900b5e0']]; loss = 0.045554
+Epoch 0: |                                                                          | 2240/? [2:18:48<00:00,  0.27it/s, v_num=cys0]context = [[5, 53], [14, 60], [21, 91], [60, 113], [2, 85], [31, 99], [27, 81], [3, 54], [15, 101], [11, 69], [0, 45], [193, 272]]target = [[39, 37], [57, 23], [71, 41], [71, 104], [30, 23], [39, 89], [58, 57], [31, 40], [83, 33], [61, 53], [28, 1], [235, 232]]
+Epoch 0: |                                                                          | 2249/? [2:19:22<00:00,  0.27it/s, v_num=cys0]train step 2250; scene = [['5afa3097bb38b159'], ['c0f67af5cd34e8d8']]; loss = 0.051479
+Epoch 0: |                                                                          | 2250/? [2:19:26<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 2250; scene = ['651a7f83ed093001'];
+target intrinsic:  tensor(0.8796, device='cuda:0') tensor(0.8798, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9141, device='cuda:0') tensor(0.9146, device='cuda:0')
+[2026-02-25 09:52:05,306][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2250/? [2:19:27<00:00,  0.27it/s, v_num=cys0]context = [[0, 21, 34, 39, 59, 63, 83, 84], [23, 26, 33, 34, 53, 58, 59, 68], [118, 137, 141, 150, 154, 170, 188, 192]]target = [[14, 79, 44, 19, 83, 18, 29, 23], [48, 24, 60, 52, 35, 61, 29, 64], [136, 187, 135, 171, 137, 131, 147, 125]]
+Epoch 0: |                                                                          | 2259/? [2:19:59<00:00,  0.27it/s, v_num=cys0]train step 2260; scene = [['39cafa06c9c431dc']]; loss = 0.054606
+Epoch 0: |                                                                          | 2260/? [2:20:03<00:00,  0.27it/s, v_num=cys0]context = [[46, 51, 80, 81, 83, 97, 105, 114], [2, 9, 10, 12, 28, 29, 51, 70], [38, 41, 42, 46, 69, 72, 118, 124]]target = [[60, 58, 103, 89, 80, 51, 99, 109], [25, 68, 19, 67, 43, 14, 52, 15], [91, 55, 98, 96, 74, 116, 83, 113]]
+Epoch 0: |                                                                          | 2269/? [2:20:37<00:00,  0.27it/s, v_num=cys0]train step 2270; scene = [['10a2d0db5c8c0962']]; loss = 0.050621
+Epoch 0: |                                                                          | 2270/? [2:20:41<00:00,  0.27it/s, v_num=cys0]context = [[33, 39, 52, 59, 85, 90, 93, 112], [16, 18, 22, 27, 40, 46, 55, 70], [2, 32, 40, 41, 42, 46, 51, 71]]target = [[40, 55, 82, 107, 111, 35, 102, 36], [25, 31, 19, 58, 55, 64, 62, 28], [48, 13, 68, 46, 30, 12, 42, 40]]
+Epoch 0: |                                                                          | 2279/? [2:21:14<00:00,  0.27it/s, v_num=cys0]train step 2280; scene = [['bcbd422c5506815b'], ['dd1a25142bac8d29'], ['ba3e38a071a523f0'], ['2248f7eaf18cca06'], ['d0edef000b9ca743'], ['d9aacea21e110a52']]; loss = 0.069012
+Epoch 0: |                                                                          | 2280/? [2:21:18<00:00,  0.27it/s, v_num=cys0]context = [[103, 107, 133, 140, 142, 146, 151, 159, 168, 169, 172, 177], [2, 4, 9, 11, 12, 17, 28, 38, 40, 46, 56, 59]]target = [[132, 133, 104, 118, 122, 149, 173, 175, 125, 120, 168, 164], [39, 49, 57, 47, 31, 13, 28, 52, 21, 16, 55, 29]]
+Epoch 0: |                                                                          | 2289/? [2:21:51<00:00,  0.27it/s, v_num=cys0]train step 2290; scene = [['fa6cb78e1503d22c'], ['f592013dce557cce']]; loss = 0.053618
+Epoch 0: |                                                                          | 2290/? [2:21:55<00:00,  0.27it/s, v_num=cys0]context = [[30, 35, 41, 49, 50, 59, 65, 78, 89, 91, 113, 116], [126, 127, 128, 134, 152, 160, 162, 166, 171, 172, 176, 177]]target = [[51, 61, 81, 100, 48, 110, 54, 112, 60, 113, 86, 55], [140, 169, 144, 127, 174, 150, 129, 143, 163, 132, 173, 152]]
+Epoch 0: |                                                                          | 2299/? [2:22:28<00:00,  0.27it/s, v_num=cys0]train step 2300; scene = [['e23f945703f61bfc']]; loss = 0.055599
+Epoch 0: |                                                                          | 2300/? [2:22:32<00:00,  0.27it/s, v_num=cys0]context = [[73, 94, 106, 123, 128, 130], [6, 22, 33, 42, 45, 51], [165, 166, 167, 196, 205, 228], [98, 109, 112, 134, 135, 152]]target = [[128, 88, 106, 120, 129, 82], [23, 48, 38, 26, 11, 14], [208, 193, 188, 218, 183, 210], [134, 143, 122, 104, 117, 137]]
+Epoch 0: |                                                                          | 2309/? [2:23:05<00:00,  0.27it/s, v_num=cys0]train step 2310; scene = [['347f4e7f9f627a12'], ['d1d992e581136ac6'], ['59e7800dec3fa9f5'], ['cf6a4349d0ffdfcf']]; loss = 0.045549
+Epoch 0: |                                                                          | 2310/? [2:23:09<00:00,  0.27it/s, v_num=cys0]context = [[0, 5, 8, 13, 15, 43, 47, 51, 61, 68, 72, 74], [8, 13, 17, 27, 29, 38, 49, 53, 56, 60, 63, 65]]target = [[5, 42, 38, 57, 31, 53, 36, 35, 59, 55, 2, 11], [23, 48, 28, 17, 61, 47, 12, 64, 36, 9, 30, 57]]
+Epoch 0: |                                                                          | 2319/? [2:23:42<00:00,  0.27it/s, v_num=cys0]train step 2320; scene = [['60c37b519a01205d'], ['b0ee123a8cfc8e62'], ['c5b4562390525d10'], ['8b950107c02ffaa9']]; loss = 0.072430
+Epoch 0: |                                                                          | 2320/? [2:23:46<00:00,  0.27it/s, v_num=cys0]context = [[8, 11, 13, 15, 17, 25, 29, 39, 40, 45, 48, 58], [4, 18, 21, 36, 47, 50, 60, 63, 71, 76, 79, 82]]target = [[56, 37, 21, 47, 26, 51, 36, 12, 49, 27, 24, 16], [62, 43, 80, 40, 30, 39, 37, 19, 6, 13, 35, 10]]
+Epoch 0: |                                                                          | 2329/? [2:24:18<00:00,  0.27it/s, v_num=cys0]train step 2330; scene = [['5c9b898102b16eae'], ['583ab14553881ee8'], ['e204af0947f704ad'], ['39edfd183c4f5b5b'], ['a85c79f15c396d71'], ['3d7a200dab472990']]; loss = 0.059580
+Epoch 0: |                                                                          | 2330/? [2:24:22<00:00,  0.27it/s, v_num=cys0]context = [[9, 23, 91], [5, 23, 61], [94, 164, 165], [149, 165, 237], [73, 137, 159], [1, 13, 48], [5, 55, 59], [25, 55, 75]]target = [[33, 46, 62], [42, 29, 22], [125, 112, 119], [187, 202, 170], [139, 87, 131], [7, 2, 23], [57, 39, 50], [62, 34, 56]]
+Epoch 0: |                                                                          | 2339/? [2:24:55<00:00,  0.27it/s, v_num=cys0]train step 2340; scene = [['c4005922f59686ae']]; loss = 0.045910
+Epoch 0: |                                                                          | 2340/? [2:24:59<00:00,  0.27it/s, v_num=cys0]context = [[9, 13, 14, 15, 29, 42, 47, 48, 52, 54, 58, 60, 61, 68, 76, 77, 81, 88, 89, 97, 98, 99, 105, 106]]target = [[72, 73, 28, 65, 82, 67, 96, 56, 75, 94, 16, 101, 63, 80, 83, 33, 41, 61, 93, 100, 14, 86, 32, 11]]
+Epoch 0: |                                                                          | 2349/? [2:25:32<00:00,  0.27it/s, v_num=cys0]train step 2350; scene = [['009e573e59c8c393'], ['7c48a30f23ea42c3'], ['739123afdcc19a64'], ['ddc11c891f471dd0'], ['38b7b864ae7b21e9'], ['997656bdb430ad43']]; loss = 0.053339
+Epoch 0: |                                                                          | 2350/? [2:25:36<00:00,  0.27it/s, v_num=cys0]context = [[115, 120, 127, 130, 134, 143, 147, 148, 149, 150, 159, 163, 173, 175, 184, 186, 187, 188, 190, 195, 199, 202, 205, 212]]target = [[148, 147, 199, 184, 190, 152, 134, 189, 127, 120, 201, 163, 186, 165, 171, 193, 117, 210, 183, 211, 195, 156, 204, 145]]
+Epoch 0: |                                                                          | 2359/? [2:26:09<00:00,  0.27it/s, v_num=cys0]train step 2360; scene = [['9a447c89080e9b56'], ['9eb8e7f262b10c23'], ['c2b8b3e74c64553a'], ['b6f9cfe435a0fde7']]; loss = 0.049974
+Epoch 0: |                                                                          | 2360/? [2:26:13<00:00,  0.27it/s, v_num=cys0]context = [[1, 4, 8, 16, 17, 28, 40, 45, 51, 53, 60, 62], [1, 4, 5, 11, 12, 13, 14, 15, 17, 29, 40, 82]]target = [[8, 52, 12, 39, 55, 43, 33, 16, 53, 18, 25, 2], [71, 5, 21, 6, 62, 55, 45, 49, 37, 14, 27, 56]]
+Epoch 0: |                                                                          | 2369/? [2:26:47<00:00,  0.27it/s, v_num=cys0]train step 2370; scene = [['090ced9a667843cb']]; loss = 0.039836
+Epoch 0: |                                                                          | 2370/? [2:26:50<00:00,  0.27it/s, v_num=cys0]context = [[13, 19, 35, 37, 43, 61, 92, 94], [0, 8, 19, 21, 24, 25, 27, 45], [1, 5, 38, 42, 45, 47, 75, 82]]target = [[46, 77, 90, 47, 83, 21, 76, 54], [3, 11, 9, 33, 20, 41, 42, 4], [72, 28, 74, 38, 52, 40, 35, 34]]
+Epoch 0: |                                                                          | 2379/? [2:27:24<00:00,  0.27it/s, v_num=cys0]train step 2380; scene = [['8ef9ff3189c85eee'], ['7e8630a890a85545']]; loss = 0.054193
+Epoch 0: |                                                                          | 2380/? [2:27:28<00:00,  0.27it/s, v_num=cys0]context = [[48, 55, 58, 63, 67, 68, 71, 73, 75, 76, 79, 80, 86, 93, 107, 108, 109, 118, 119, 120, 129, 133, 135, 145]]target = [[133, 59, 121, 87, 96, 110, 67, 101, 49, 142, 60, 134, 109, 132, 98, 122, 103, 100, 80, 57, 65, 86, 128, 143]]
+Epoch 0: |                                                                          | 2389/? [2:28:01<00:00,  0.27it/s, v_num=cys0]train step 2390; scene = [['343a98bd8cfda2de'], ['c09d7898ef37ba32']]; loss = 0.051881
+Epoch 0: |                                                                          | 2390/? [2:28:05<00:00,  0.27it/s, v_num=cys0]context = [[19, 20, 31, 32, 41, 46, 49, 50, 51, 53, 59, 60, 62, 63, 74, 77, 82, 84, 85, 91, 93, 97, 98, 116]]target = [[56, 31, 77, 102, 39, 54, 21, 57, 66, 75, 100, 115, 26, 53, 34, 113, 97, 90, 25, 82, 108, 60, 67, 38]]
+Epoch 0: |                                                                          | 2399/? [2:28:39<00:00,  0.27it/s, v_num=cys0]train step 2400; scene = [['8d758914077e5926']]; loss = 0.049282
+Epoch 0: |                                                                          | 2400/? [2:28:43<00:00,  0.27it/s, v_num=cys0]context = [[114, 117, 119, 120, 131, 136, 141, 148, 149, 151, 163, 168], [44, 47, 52, 76, 77, 78, 85, 88, 89, 90, 97, 99]]target = [[141, 153, 135, 128, 126, 119, 140, 147, 161, 120, 155, 127], [82, 75, 49, 62, 84, 54, 78, 70, 55, 72, 96, 92]]
+[2026-02-25 10:01:25,665][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2409/? [2:29:18<00:00,  0.27it/s, v_num=cys0]train step 2410; scene = [['d35650707515c2cf']]; loss = 0.050781
+Epoch 0: |                                                                          | 2410/? [2:29:21<00:00,  0.27it/s, v_num=cys0]context = [[4, 8, 16, 30, 45, 53, 70, 77], [123, 139, 144, 149, 166, 184, 187, 207], [38, 44, 51, 66, 68, 90, 94, 98]]target = [[7, 17, 15, 70, 47, 48, 35, 72], [190, 168, 157, 135, 200, 160, 161, 128], [84, 75, 51, 67, 64, 46, 95, 49]]
+Epoch 0: |                                                                          | 2419/? [2:29:55<00:00,  0.27it/s, v_num=cys0]train step 2420; scene = [['ea5def9e91076788']]; loss = 0.040168
+Epoch 0: |                                                                          | 2420/? [2:29:59<00:00,  0.27it/s, v_num=cys0]context = [[93, 107, 110, 121, 127, 148], [165, 169, 170, 208, 209, 214], [0, 6, 7, 23, 45, 83], [12, 33, 38, 45, 81, 91]]target = [[131, 99, 135, 127, 142, 120], [167, 177, 176, 213, 171, 191], [22, 45, 64, 29, 67, 51], [47, 31, 65, 87, 60, 85]]
+Epoch 0: |                                                                          | 2429/? [2:30:32<00:00,  0.27it/s, v_num=cys0]train step 2430; scene = [['45fe4bbb2526ca6e'], ['0db2c3f2775880de'], ['4a0f95a3db913b56']]; loss = 0.040114
+Epoch 0: |                                                                          | 2430/? [2:30:35<00:00,  0.27it/s, v_num=cys0]context = [[22, 66, 68, 80, 92, 99], [26, 44, 52, 76, 80, 86], [3, 38, 40, 69, 77, 89], [43, 83, 90, 109, 121, 126]]target = [[93, 85, 61, 37, 46, 63], [48, 57, 64, 35, 27, 84], [43, 41, 48, 23, 25, 67], [111, 55, 61, 105, 114, 116]]
+Epoch 0: |                                                                          | 2439/? [2:31:09<00:00,  0.27it/s, v_num=cys0]train step 2440; scene = [['512f1df8266be984'], ['50c8a233bd82a613']]; loss = 0.068929
+Epoch 0: |                                                                          | 2440/? [2:31:13<00:00,  0.27it/s, v_num=cys0]context = [[96, 106, 114, 124, 131, 158], [48, 54, 56, 57, 87, 102], [2, 8, 16, 43, 50, 59], [22, 27, 37, 53, 69, 77]]target = [[112, 148, 103, 120, 104, 132], [72, 96, 86, 90, 52, 98], [46, 13, 56, 3, 57, 55], [58, 57, 71, 67, 28, 45]]
+Epoch 0: |                                                                          | 2449/? [2:31:45<00:00,  0.27it/s, v_num=cys0]train step 2450; scene = [['ac584e6fc77676a4'], ['19a3235d680e11c4'], ['61b9def6cdf00024'], ['fc86266e2fcb72fd']]; loss = 0.046442
+Epoch 0: |                                                                          | 2450/? [2:31:49<00:00,  0.27it/s, v_num=cys0]context = [[16, 41, 71], [43, 46, 97], [10, 18, 90], [3, 16, 53], [143, 187, 198], [25, 36, 70], [98, 136, 179], [19, 65, 78]]target = [[45, 18, 43], [95, 96, 51], [32, 21, 70], [5, 8, 27], [146, 154, 151], [28, 31, 48], [133, 137, 151], [33, 53, 64]]
+Epoch 0: |                                                                          | 2459/? [2:32:22<00:00,  0.27it/s, v_num=cys0]train step 2460; scene = [['dc5f38f005c3ebd6']]; loss = 0.075203
+Epoch 0: |                                                                          | 2460/? [2:32:26<00:00,  0.27it/s, v_num=cys0]context = [[4, 5, 11, 12, 16, 17, 19, 22, 23, 30, 39, 44, 49, 50, 51, 61, 77, 82, 87, 89, 95, 96, 97, 101]]target = [[53, 36, 13, 57, 86, 80, 97, 72, 84, 27, 61, 94, 35, 51, 54, 67, 43, 66, 62, 89, 7, 92, 10, 47]]
+Epoch 0: |                                                                          | 2469/? [2:33:00<00:00,  0.27it/s, v_num=cys0]train step 2470; scene = [['c7620995ebe9c4e6'], ['06241ebed1658f34'], ['1f214117250f089a'], ['5e31e0691d426537'], ['11c4a7d67bc2629e'], ['715e8695976cdb61'], ['6888f7ca14081419'], ['8fe341dcd0880bd5']]; loss = 0.072283
+Epoch 0: |                                                                          | 2470/? [2:33:04<00:00,  0.27it/s, v_num=cys0]context = [[17, 28, 38, 42, 68, 86], [2, 4, 28, 37, 51, 58], [7, 21, 24, 54, 64, 70], [41, 42, 61, 71, 81, 89]]target = [[55, 30, 52, 28, 44, 76], [22, 36, 3, 57, 34, 9], [16, 35, 34, 61, 56, 43], [56, 87, 85, 73, 68, 50]]
+Epoch 0: |                                                                          | 2479/? [2:33:36<00:00,  0.27it/s, v_num=cys0]train step 2480; scene = [['cf98d3219d144500']]; loss = 0.075524
+Epoch 0: |                                                                          | 2480/? [2:33:40<00:00,  0.27it/s, v_num=cys0]context = [[151, 158, 171, 173, 176, 205, 206, 230], [74, 88, 94, 98, 119, 130, 146, 163], [5, 8, 33, 42, 52, 53, 57, 59]]target = [[177, 223, 212, 220, 205, 207, 225, 222], [102, 132, 115, 112, 145, 156, 119, 100], [12, 23, 49, 20, 55, 32, 10, 39]]
+Epoch 0: |                                                                          | 2489/? [2:34:12<00:00,  0.27it/s, v_num=cys0]train step 2490; scene = [['ebff5d05f1bb086f'], ['880427ff150f7b4d'], ['9dd0efd4b4626604'], ['f52c70025f40e56d']]; loss = 0.060628
+Epoch 0: |                                                                          | 2490/? [2:34:16<00:00,  0.27it/s, v_num=cys0]context = [[19, 27, 31, 35, 36, 42, 57, 60, 63, 64, 65, 70], [21, 26, 31, 32, 50, 56, 62, 67, 74, 81, 90, 93]]target = [[51, 33, 64, 68, 53, 29, 54, 27, 26, 69, 35, 48], [59, 60, 83, 76, 86, 68, 61, 36, 71, 22, 69, 82]]
+Epoch 0: |                                                                          | 2499/? [2:34:48<00:00,  0.27it/s, v_num=cys0]train step 2500; scene = [['7dc5c394263df267']]; loss = 0.062105
+Epoch 0: |                                                                          | 2500/? [2:34:52<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 2500; scene = ['97ef4323919c5e8a'];
+target intrinsic:  tensor(0.8889, device='cuda:0') tensor(0.8892, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9225, device='cuda:0') tensor(0.9195, device='cuda:0')
+[2026-02-25 10:07:31,359][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2500/? [2:34:53<00:00,  0.27it/s, v_num=cys0]context = [[17, 21, 30, 44, 52, 84], [24, 34, 39, 45, 53, 84], [159, 163, 179, 192, 194, 207], [115, 116, 136, 154, 172, 188]]target = [[31, 20, 44, 37, 73, 59], [47, 48, 59, 36, 52, 44], [198, 165, 170, 193, 188, 175], [173, 128, 152, 148, 158, 172]]
+Epoch 0: |                                                                          | 2509/? [2:35:26<00:00,  0.27it/s, v_num=cys0]train step 2510; scene = [['e806eb496df1dfa0'], ['3caa89905a59e7e6'], ['3fc603b4c4531c11'], ['08da23838ee6e23b'], ['89a0f0553f7f2bdb'], ['5080347450bbcb08']]; loss = 0.063907
+Epoch 0: |                                                                          | 2510/? [2:35:30<00:00,  0.27it/s, v_num=cys0]context = [[7, 16, 17, 33, 34, 42, 43, 45, 48, 49, 52, 60], [16, 27, 29, 31, 38, 42, 58, 64, 71, 89, 93, 104]]target = [[59, 11, 58, 34, 22, 51, 55, 46, 9, 12, 31, 56], [90, 95, 100, 19, 82, 44, 96, 91, 97, 28, 88, 77]]
+Epoch 0: |                                                                          | 2519/? [2:36:04<00:00,  0.27it/s, v_num=cys0]train step 2520; scene = [['0a94493d6f4cd4be'], ['1c3a8fcf1547dcf7'], ['c34950273c4d538b'], ['397652e91be52496'], ['d9a133a4747493d2'], ['46dc3900c302593a'], ['e281213868014707'], ['83c959de5ae5b86c']]; loss = 0.083002
+Epoch 0: |                                                                          | 2520/? [2:36:08<00:00,  0.27it/s, v_num=cys0]context = [[22, 26, 34, 62, 68, 89, 103, 111], [16, 23, 45, 57, 69, 77, 101, 105], [9, 21, 23, 24, 56, 60, 66, 68]]target = [[29, 88, 85, 106, 103, 79, 65, 23], [61, 100, 46, 104, 23, 66, 18, 40], [25, 38, 15, 16, 64, 45, 34, 30]]
+Epoch 0: |                                                                          | 2529/? [2:36:41<00:00,  0.27it/s, v_num=cys0]train step 2530; scene = [['871ec8d951a81c62']]; loss = 0.052959
+Epoch 0: |                                                                          | 2530/? [2:36:45<00:00,  0.27it/s, v_num=cys0]context = [[39, 41, 46, 48, 52, 53, 55, 66, 68, 70, 83, 92, 93, 97, 103, 106, 110, 111, 118, 124, 125, 131, 133, 136]]target = [[125, 130, 53, 128, 124, 75, 61, 112, 64, 70, 99, 120, 57, 47, 69, 60, 43, 85, 79, 63, 45, 135, 104, 129]]
+Epoch 0: |                                                                          | 2539/? [2:37:16<00:00,  0.27it/s, v_num=cys0]train step 2540; scene = [['8e78ce55e3180547']]; loss = 0.043926
+Epoch 0: |                                                                          | 2540/? [2:37:19<00:00,  0.27it/s, v_num=cys0]context = [[19, 22, 26, 27, 29, 35, 40, 42, 54, 59, 60, 61, 64, 74, 77, 86, 92, 102, 106, 109, 110, 111, 115, 116]]target = [[89, 69, 46, 77, 65, 57, 110, 49, 70, 97, 38, 58, 31, 40, 101, 74, 20, 88, 92, 90, 80, 28, 94, 37]]
+Epoch 0: |                                                                          | 2549/? [2:37:52<00:00,  0.27it/s, v_num=cys0]train step 2550; scene = [['6dfaec91f745fdd9'], ['3f72021b93b6224a'], ['36be66d194d57ec8']]; loss = 0.072194
+Epoch 0: |                                                                          | 2550/? [2:37:56<00:00,  0.27it/s, v_num=cys0]context = [[47, 51, 57, 66, 73, 85, 93, 96, 97, 107, 113, 121], [26, 41, 43, 44, 54, 64, 69, 75, 78, 82, 85, 86]]target = [[83, 71, 53, 103, 110, 111, 55, 70, 94, 72, 74, 106], [32, 47, 85, 36, 49, 51, 41, 62, 35, 84, 31, 28]]
+Epoch 0: |                                                                          | 2559/? [2:38:29<00:00,  0.27it/s, v_num=cys0]train step 2560; scene = [['7c6d160c26de6887'], ['5ef6fe9ef5309457'], ['66de8729fe760dd8'], ['faba60084d22aa27']]; loss = 0.056428
+Epoch 0: |                                                                          | 2560/? [2:38:32<00:00,  0.27it/s, v_num=cys0]context = [[44, 89, 97], [157, 230, 235], [6, 8, 64], [19, 33, 76], [24, 39, 91], [14, 17, 65], [34, 59, 83], [19, 28, 77]]target = [[93, 77, 73], [195, 166, 202], [62, 61, 20], [61, 65, 30], [82, 90, 72], [32, 60, 21], [67, 74, 43], [39, 27, 37]]
+Epoch 0: |                                                                          | 2569/? [2:39:06<00:00,  0.27it/s, v_num=cys0]train step 2570; scene = [['678d8464781a3de2'], ['9ab12a31f2a3b9fb']]; loss = 0.049756
+Epoch 0: |                                                                          | 2570/? [2:39:10<00:00,  0.27it/s, v_num=cys0]context = [[41, 92], [2, 64], [133, 190], [7, 81], [6, 63], [33, 93], [65, 146], [18, 69], [15, 63], [13, 78], [4, 79], [66, 126]]target = [[90, 71], [12, 37], [175, 173], [56, 25], [7, 56], [61, 69], [132, 93], [28, 45], [17, 31], [21, 61], [69, 76], [120, 115]]
+Epoch 0: |                                                                          | 2579/? [2:39:42<00:00,  0.27it/s, v_num=cys0]train step 2580; scene = [['c40b23830f7437d5'], ['a0c7cabb66c795a4'], ['df63cd7eb6e92486']]; loss = 0.055387
+Epoch 0: |                                                                          | 2580/? [2:39:45<00:00,  0.27it/s, v_num=cys0]context = [[110, 115, 117, 128, 136, 137, 138, 141, 143, 145, 150, 168], [7, 10, 23, 30, 35, 48, 57, 59, 63, 64, 65, 69]]target = [[128, 149, 154, 126, 167, 132, 162, 151, 112, 134, 133, 131], [18, 25, 50, 30, 12, 19, 47, 39, 54, 35, 37, 56]]
+Epoch 0: |                                                                          | 2589/? [2:40:19<00:00,  0.27it/s, v_num=cys0]train step 2590; scene = [['65cdcbcb16f0ebe5'], ['4b12a530a5ab03d0']]; loss = 0.040118
+Epoch 0: |                                                                          | 2590/? [2:40:23<00:00,  0.27it/s, v_num=cys0]context = [[32, 42, 45, 59, 89, 100], [65, 89, 96, 104, 111, 139], [177, 197, 217, 235, 253, 264], [81, 94, 117, 131, 132, 142]]target = [[41, 75, 46, 34, 67, 60], [125, 90, 94, 114, 92, 95], [196, 223, 250, 229, 245, 259], [124, 91, 129, 94, 85, 131]]
+Epoch 0: |                                                                          | 2599/? [2:40:56<00:00,  0.27it/s, v_num=cys0]train step 2600; scene = [['cb3bac70297d52c0']]; loss = 0.038228
+Epoch 0: |                                                                          | 2600/? [2:41:00<00:00,  0.27it/s, v_num=cys0]context = [[3, 4, 7, 15, 37, 44, 46, 49], [71, 83, 85, 107, 113, 118, 123, 135], [4, 6, 26, 29, 32, 45, 51, 52]]target = [[38, 11, 27, 26, 16, 48, 39, 30], [81, 90, 94, 125, 75, 97, 110, 98], [22, 27, 29, 35, 7, 9, 49, 21]]
+[2026-02-25 10:13:42,512][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2609/? [2:41:33<00:00,  0.27it/s, v_num=cys0]train step 2610; scene = [['262404ef111a096b'], ['5ac9dbde9cf64eed'], ['c32ee937dbec7221'], ['201e44ad015728c7'], ['1fb562b09fc361ea'], ['e454026f5348630e'], ['d1772c09b4b6d95f'], ['8e7cc5c956a4cf95']]; loss = 0.056347
+Epoch 0: |                                                                          | 2610/? [2:41:36<00:00,  0.27it/s, v_num=cys0]context = [[70, 73, 74, 76, 77, 79, 82, 88, 109, 110, 115, 124, 125, 136, 141, 148, 150, 151, 155, 158, 159, 160, 161, 167]]target = [[156, 121, 159, 150, 128, 138, 162, 81, 102, 71, 103, 74, 137, 144, 108, 166, 72, 161, 129, 110, 165, 136, 106, 99]]
+Epoch 0: |                                                                          | 2619/? [2:42:09<00:00,  0.27it/s, v_num=cys0]train step 2620; scene = [['4dd74d4d53abb812'], ['9029d8d8f6f0e98a'], ['360253c55aef692f'], ['b07a54eda9cdaccb'], ['ca553e70eff3aa6f'], ['70b45bbd1147fbf0']]; loss = 0.045625
+Epoch 0: |                                                                          | 2620/? [2:42:13<00:00,  0.27it/s, v_num=cys0]context = [[27, 55, 94], [5, 48, 83], [56, 94, 118], [121, 137, 181], [0, 61, 78], [36, 80, 108], [2, 18, 55], [8, 60, 69]]target = [[62, 80, 57], [21, 23, 34], [84, 71, 87], [139, 178, 159], [59, 19, 76], [104, 40, 51], [43, 20, 8], [47, 68, 60]]
+Epoch 0: |                                                                          | 2629/? [2:42:47<00:00,  0.27it/s, v_num=cys0]train step 2630; scene = [['c13076a68aeeb481'], ['0bafd80c9e2ae41b']]; loss = 0.053952
+Epoch 0: |                                                                          | 2630/? [2:42:50<00:00,  0.27it/s, v_num=cys0]context = [[135, 142, 145, 147, 149, 150, 158, 162, 163, 169, 172, 203], [22, 29, 55, 60, 62, 63, 64, 68, 69, 70, 71, 72]]target = [[190, 143, 159, 194, 198, 175, 197, 146, 166, 181, 174, 187], [47, 36, 59, 40, 28, 46, 65, 24, 29, 26, 69, 51]]
+Epoch 0: |                                                                          | 2639/? [2:43:24<00:00,  0.27it/s, v_num=cys0]train step 2640; scene = [['eae08bd3d5c892e7'], ['60acba38edb03894']]; loss = 0.042337
+Epoch 0: |                                                                          | 2640/? [2:43:28<00:00,  0.27it/s, v_num=cys0]context = [[166, 176, 251], [156, 173, 220], [69, 137, 148], [197, 204, 277], [28, 29, 100], [44, 111, 118], [0, 56, 61], [4, 68, 81]]target = [[170, 244, 200], [193, 173, 188], [75, 86, 116], [253, 268, 251], [49, 30, 48], [106, 54, 50], [24, 3, 29], [24, 58, 37]]
+Epoch 0: |                                                                          | 2649/? [2:44:01<00:00,  0.27it/s, v_num=cys0]train step 2650; scene = [['c520ae3405948a0a'], ['94340815d8708eea']]; loss = 0.032260
+Epoch 0: |                                                                          | 2650/? [2:44:05<00:00,  0.27it/s, v_num=cys0]context = [[6, 10, 15, 16, 20, 26, 31, 32, 37, 40, 42, 47, 55, 57, 66, 69, 70, 80, 82, 85, 88, 93, 101, 103]]target = [[72, 23, 99, 11, 66, 101, 81, 102, 30, 39, 54, 38, 17, 84, 69, 32, 22, 18, 26, 41, 79, 40, 57, 46]]
+Epoch 0: |                                                                          | 2659/? [2:44:39<00:00,  0.27it/s, v_num=cys0]train step 2660; scene = [['6666ae9375b1656e'], ['64619aa7bd88899d']]; loss = 0.080083
+Epoch 0: |                                                                          | 2660/? [2:44:43<00:00,  0.27it/s, v_num=cys0]context = [[13, 14, 27, 29, 31, 32, 39, 48, 50, 52, 54, 57, 58, 61, 67, 73, 79, 90, 93, 95, 105, 108, 109, 110]]target = [[83, 48, 82, 104, 25, 53, 89, 72, 93, 51, 41, 86, 24, 49, 32, 98, 78, 73, 88, 45, 44, 99, 58, 79]]
+Epoch 0: |                                                                          | 2669/? [2:45:16<00:00,  0.27it/s, v_num=cys0]train step 2670; scene = [['f0712581f6277ffc'], ['5c4442779124ec3d']]; loss = 0.045611
+Epoch 0: |                                                                          | 2670/? [2:45:20<00:00,  0.27it/s, v_num=cys0]context = [[0, 1, 4, 5, 6, 8, 9, 11, 13, 29, 32, 34, 35, 38, 43, 47, 50, 58, 61, 67, 86, 94, 96, 97]]target = [[38, 81, 77, 72, 58, 43, 36, 5, 7, 23, 27, 62, 13, 30, 52, 55, 19, 54, 84, 76, 49, 65, 73, 91]]
+Epoch 0: |                                                                          | 2679/? [2:45:54<00:00,  0.27it/s, v_num=cys0]train step 2680; scene = [['f276b95e48af6e36'], ['e5db691627ea5357']]; loss = 0.062417
+Epoch 0: |                                                                          | 2680/? [2:45:57<00:00,  0.27it/s, v_num=cys0]context = [[108, 120, 121, 130, 145, 149, 158, 169], [9, 17, 36, 43, 44, 69, 77, 78], [107, 120, 131, 140, 141, 142, 147, 152]]target = [[139, 144, 124, 138, 161, 143, 120, 110], [22, 74, 33, 18, 69, 50, 38, 23], [145, 132, 140, 147, 126, 119, 148, 128]]
+Epoch 0: |                                                                          | 2689/? [2:46:31<00:00,  0.27it/s, v_num=cys0]train step 2690; scene = [['2059d1d6c79bc51b'], ['6b1950140a598578'], ['a20b4125ede06429']]; loss = 0.086323
+Epoch 0: |                                                                          | 2690/? [2:46:34<00:00,  0.27it/s, v_num=cys0]context = [[3, 15, 21, 22, 26, 27, 30, 34, 45, 64, 67, 73], [6, 7, 10, 12, 24, 39, 43, 45, 47, 55, 58, 64]]target = [[49, 18, 47, 23, 50, 70, 62, 17, 42, 46, 27, 68], [59, 12, 7, 26, 50, 58, 57, 39, 23, 49, 55, 24]]
+Epoch 0: |                                                                          | 2699/? [2:47:08<00:00,  0.27it/s, v_num=cys0]train step 2700; scene = [['319bd9ea90f25ea3'], ['12879245713d8124'], ['9bb2a6670058b7b2']]; loss = 0.064606
+Epoch 0: |                                                                          | 2700/? [2:47:12<00:00,  0.27it/s, v_num=cys0]context = [[39, 40, 41, 44, 48, 49, 52, 58, 63, 67, 82, 83, 85, 87, 88, 93, 102, 104, 108, 109, 113, 118, 131, 136]]target = [[43, 106, 81, 41, 105, 135, 98, 78, 103, 44, 47, 61, 45, 52, 124, 48, 91, 59, 58, 46, 85, 109, 74, 56]]
+Epoch 0: |                                                                          | 2709/? [2:47:43<00:00,  0.27it/s, v_num=cys0]train step 2710; scene = [['753238098c2307cc']]; loss = 0.084877
+Epoch 0: |                                                                          | 2710/? [2:47:47<00:00,  0.27it/s, v_num=cys0]context = [[6, 7, 8, 9, 17, 19, 29, 32, 39, 48, 54, 55], [110, 116, 118, 121, 122, 128, 134, 138, 144, 149, 170, 178]]target = [[9, 51, 23, 19, 49, 54, 38, 7, 37, 53, 17, 16], [147, 156, 124, 177, 118, 162, 160, 143, 140, 119, 154, 133]]
+Epoch 0: |                                                                          | 2719/? [2:48:20<00:00,  0.27it/s, v_num=cys0]train step 2720; scene = [['7194b8d204f4a0b6'], ['5c0ddb9de8c16f05'], ['1a87a846ba692048']]; loss = 0.044408
+Epoch 0: |                                                                          | 2720/? [2:48:24<00:00,  0.27it/s, v_num=cys0]context = [[73, 88, 89, 91, 93, 96, 104, 111, 114, 115, 118, 120, 121, 125, 133, 136, 137, 143, 144, 145, 151, 161, 165, 170]]target = [[94, 126, 164, 115, 101, 103, 78, 120, 75, 153, 88, 86, 74, 106, 116, 141, 119, 144, 140, 165, 146, 167, 158, 139]]
+Epoch 0: |                                                                          | 2729/? [2:48:58<00:00,  0.27it/s, v_num=cys0]train step 2730; scene = [['b7a4f7a6d35961d4'], ['b7003ac834dc298b'], ['05596054e7569f2b'], ['b8aed6b43cd738c9']]; loss = 0.064321
+Epoch 0: |                                                                          | 2730/? [2:49:02<00:00,  0.27it/s, v_num=cys0]context = [[117, 186, 194, 196], [34, 38, 40, 79], [10, 13, 37, 77], [102, 103, 113, 167], [9, 42, 52, 80], [136, 174, 192, 218]]target = [[187, 180, 158, 134], [57, 76, 69, 70], [49, 37, 70, 57], [158, 143, 120, 136], [15, 34, 75, 65], [138, 214, 217, 143]]
+Epoch 0: |                                                                          | 2739/? [2:49:35<00:00,  0.27it/s, v_num=cys0]train step 2740; scene = [['d9641b3f6ca0b13d'], ['61ef38380e8172c7'], ['3ec462170f378b3f'], ['bc0b4df5aef0a622'], ['6bef29b74b93e80a'], ['8ef643c4e1cb9baf'], ['aa8259399a115c5f'], ['fa35453daaa8e408']]; loss = 0.067017
+Epoch 0: |                                                                          | 2740/? [2:49:38<00:00,  0.27it/s, v_num=cys0]context = [[0, 1, 7, 8, 9, 10, 11, 18, 22, 31, 32, 39, 49, 53, 59, 66, 69, 72, 74, 78, 81, 82, 93, 97]]target = [[74, 55, 50, 21, 47, 85, 2, 42, 41, 40, 39, 93, 49, 45, 23, 66, 73, 38, 89, 36, 87, 30, 48, 9]]
+Epoch 0: |                                                                          | 2749/? [2:50:12<00:00,  0.27it/s, v_num=cys0]train step 2750; scene = [['2225123ef31a93e4']]; loss = 0.088707
+Epoch 0: |                                                                          | 2750/? [2:50:16<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 2750; scene = ['3e07add8413f8157'];
+target intrinsic:  tensor(0.8521, device='cuda:0') tensor(0.8523, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8763, device='cuda:0') tensor(0.8754, device='cuda:0')
+[2026-02-25 10:22:55,254][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2750/? [2:50:17<00:00,  0.27it/s, v_num=cys0]context = [[12, 14, 17, 18, 20, 21, 32, 34, 38, 40, 54, 57, 65, 67, 72, 77, 82, 85, 92, 96, 98, 102, 105, 109]]target = [[101, 39, 104, 51, 15, 33, 16, 24, 94, 58, 72, 48, 20, 17, 22, 37, 107, 63, 40, 79, 69, 27, 28, 60]]
+Epoch 0: |                                                                          | 2759/? [2:50:50<00:00,  0.27it/s, v_num=cys0]train step 2760; scene = [['03a4631365a95cb2'], ['9f622f50132c1efe'], ['4ec29bd9dfd30c2e'], ['314d09d3997725a7'], ['c437043ac9ced6c1'], ['accfed2fa6ff849d'], ['57a1c0730778cddd'], ['bea0e295ee56d42c'], ['75e5ba7e7fb64fc0'], ['71968459b3168e4f'], ['a08c8a37e64ce67d'], ['bf27ccdada0c2373']]; loss = 0.096702
+Epoch 0: |                                                                          | 2760/? [2:50:53<00:00,  0.27it/s, v_num=cys0]context = [[41, 48, 51, 52, 59, 61, 68, 69, 72, 73, 75, 103], [9, 10, 23, 35, 37, 38, 41, 43, 48, 69, 76, 84]]target = [[56, 74, 53, 76, 101, 97, 93, 42, 89, 83, 91, 79], [14, 17, 28, 50, 58, 32, 54, 30, 61, 34, 83, 15]]
+Epoch 0: |                                                                          | 2769/? [2:51:27<00:00,  0.27it/s, v_num=cys0]train step 2770; scene = [['539e0d3713447384'], ['c49e7882e04566c0'], ['69c9010e3d4209bc']]; loss = 0.038474
+Epoch 0: |                                                                          | 2770/? [2:51:31<00:00,  0.27it/s, v_num=cys0]context = [[0, 9, 19, 31, 34, 68], [7, 20, 31, 36, 41, 53], [1, 15, 29, 56, 58, 74], [0, 19, 21, 35, 43, 50]]target = [[10, 2, 39, 12, 11, 26], [21, 31, 11, 28, 27, 52], [16, 13, 66, 17, 65, 6], [39, 21, 41, 27, 40, 25]]
+Epoch 0: |                                                                          | 2779/? [2:52:04<00:00,  0.27it/s, v_num=cys0]train step 2780; scene = [['ede896927ea91dd6'], ['7badba95b5be610e']]; loss = 0.048480
+Epoch 0: |                                                                          | 2780/? [2:52:08<00:00,  0.27it/s, v_num=cys0]context = [[0, 7, 26, 27, 51, 57], [5, 15, 18, 23, 55, 81], [27, 32, 44, 76, 87, 98], [8, 16, 18, 21, 24, 58]]target = [[48, 19, 22, 26, 32, 17], [57, 65, 27, 75, 51, 22], [85, 50, 79, 82, 30, 89], [41, 25, 12, 52, 32, 47]]
+Epoch 0: |                                                                          | 2789/? [2:52:41<00:00,  0.27it/s, v_num=cys0]train step 2790; scene = [['1e7d7ef1404597f0'], ['ea02d0f42c603c21']]; loss = 0.036799
+Epoch 0: |                                                                          | 2790/? [2:52:45<00:00,  0.27it/s, v_num=cys0]context = [[99, 100, 101, 107, 108, 110, 111, 113, 116, 128, 131, 139, 145, 149, 155, 162, 167, 170, 172, 179, 183, 191, 194, 196]]target = [[122, 186, 110, 138, 149, 133, 161, 112, 141, 184, 189, 178, 100, 131, 188, 155, 174, 195, 185, 171, 137, 160, 176, 170]]
+Epoch 0: |                                                                          | 2799/? [2:53:18<00:00,  0.27it/s, v_num=cys0]train step 2800; scene = [['c8e92789f25baec1'], ['b477406d6064f1a3']]; loss = 0.031863
+Epoch 0: |                                                                          | 2800/? [2:53:22<00:00,  0.27it/s, v_num=cys0]context = [[2, 4, 5, 18, 22, 24, 31, 37, 39, 45, 54, 57, 60, 64, 67, 73, 82, 83, 84, 86, 87, 97, 98, 99]]target = [[23, 55, 39, 36, 80, 52, 28, 29, 89, 63, 56, 61, 98, 78, 40, 9, 77, 82, 69, 47, 43, 68, 81, 44]]
+[2026-02-25 10:26:05,189][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2809/? [2:54:00<00:00,  0.27it/s, v_num=cys0]train step 2810; scene = [['31907a6ebec6ef2e']]; loss = 0.033095
+Epoch 0: |                                                                          | 2810/? [2:54:04<00:00,  0.27it/s, v_num=cys0]context = [[96, 97, 145, 156, 165, 173], [147, 150, 159, 169, 173, 196], [21, 33, 34, 46, 56, 75], [7, 17, 47, 50, 51, 58]]target = [[144, 115, 146, 171, 129, 101], [194, 177, 170, 186, 155, 193], [34, 60, 52, 63, 32, 22], [37, 45, 50, 22, 27, 35]]
+Epoch 0: |                                                                          | 2819/? [2:54:38<00:00,  0.27it/s, v_num=cys0]train step 2820; scene = [['b60f8abb905bb1e1']]; loss = 0.062230
+Epoch 0: |                                                                          | 2820/? [2:54:42<00:00,  0.27it/s, v_num=cys0]context = [[11, 22, 72], [43, 82, 129], [0, 56, 58], [40, 93, 105], [0, 28, 57], [3, 75, 76], [51, 109, 110], [26, 85, 112]]target = [[21, 31, 45], [57, 82, 123], [45, 30, 31], [94, 95, 98], [27, 53, 38], [43, 50, 49], [102, 80, 69], [106, 107, 98]]
+Epoch 0: |                                                                          | 2829/? [2:55:15<00:00,  0.27it/s, v_num=cys0]train step 2830; scene = [['ad1ad6648a30592c'], ['448e12a47b369490'], ['d3dddf816450d1d0'], ['d7fd0247a853f44b'], ['27cf7ba27eed1eae'], ['fd73259e7c2801e8'], ['9869d33078eac049'], ['06779b91ed31eaf1']]; loss = 0.065979
+Epoch 0: |                                                                          | 2830/? [2:55:19<00:00,  0.27it/s, v_num=cys0]context = [[15, 17, 19, 20, 21, 26, 36, 40, 42, 44, 47, 49, 50, 57, 67, 77, 78, 87, 89, 90, 102, 104, 105, 112]]target = [[103, 47, 82, 105, 98, 89, 52, 35, 33, 55, 65, 60, 107, 87, 38, 34, 20, 40, 81, 102, 88, 53, 93, 16]]
+Epoch 0: |                                                                          | 2839/? [2:55:53<00:00,  0.27it/s, v_num=cys0]train step 2840; scene = [['a911c16d8056a34d']]; loss = 0.031964
+Epoch 0: |                                                                          | 2840/? [2:55:57<00:00,  0.27it/s, v_num=cys0]context = [[12, 15, 17, 19, 47, 48, 51, 58], [32, 38, 40, 56, 57, 72, 75, 78], [75, 90, 96, 97, 109, 115, 129, 139]]target = [[53, 44, 45, 42, 50, 17, 36, 54], [71, 72, 37, 48, 54, 39, 74, 52], [104, 134, 124, 132, 129, 127, 137, 121]]
+Epoch 0: |                                                                          | 2849/? [2:56:31<00:00,  0.27it/s, v_num=cys0]train step 2850; scene = [['9753a8291766b5da'], ['3a8f14b855f7f4ee'], ['ba3f2372b7959e95']]; loss = 0.055231
+Epoch 0: |                                                                          | 2850/? [2:56:34<00:00,  0.27it/s, v_num=cys0]context = [[134, 142, 143, 146, 150, 151, 154, 155, 177, 179, 184, 208], [70, 78, 83, 88, 97, 106, 108, 112, 115, 126, 133, 139]]target = [[160, 140, 154, 139, 135, 166, 141, 167, 155, 136, 164, 199], [72, 104, 100, 117, 114, 103, 112, 123, 133, 115, 110, 81]]
+Epoch 0: |                                                                          | 2859/? [2:57:07<00:00,  0.27it/s, v_num=cys0]train step 2860; scene = [['8f4f366720645ea0'], ['fef48b769b17f0ed']]; loss = 0.033012
+Epoch 0: |                                                                          | 2860/? [2:57:11<00:00,  0.27it/s, v_num=cys0]context = [[115, 116, 123, 135, 137, 144, 146, 152, 155, 161, 162, 168, 176, 177, 179, 184, 187, 190, 198, 199, 200, 202, 209, 212]]target = [[195, 155, 181, 142, 154, 202, 144, 161, 136, 132, 137, 192, 148, 189, 188, 174, 178, 133, 156, 134, 124, 197, 199, 166]]
+Epoch 0: |                                                                          | 2869/? [2:57:44<00:00,  0.27it/s, v_num=cys0]train step 2870; scene = [['dff9f50bbc0d0d5d'], ['3baf335bfb54b9c0'], ['c74adb4381e65e99']]; loss = 0.054674
+Epoch 0: |                                                                          | 2870/? [2:57:48<00:00,  0.27it/s, v_num=cys0]context = [[30, 41, 45, 47, 51, 57, 66, 68, 77, 83, 85, 90, 95, 96, 103, 105, 110, 111, 116, 120, 122, 123, 126, 127]]target = [[118, 32, 75, 92, 109, 89, 43, 36, 72, 115, 96, 111, 76, 68, 90, 70, 69, 66, 38, 77, 81, 45, 34, 42]]
+Epoch 0: |                                                                          | 2879/? [2:58:22<00:00,  0.27it/s, v_num=cys0]train step 2880; scene = [['d6bd4b0784843fdd'], ['bd1bc8c20a660a96'], ['f8f0491a2268ee5e'], ['90bcc3b4b3d551a8'], ['57ccd86b6a8979b3'], ['9d66dba4551f79f8'], ['a30f898a6e455745'], ['73973f88f24769be']]; loss = 0.057784
+Epoch 0: |                                                                          | 2880/? [2:58:25<00:00,  0.27it/s, v_num=cys0]context = [[2, 4, 5, 6, 10, 14, 18, 21, 29, 36, 41, 54], [21, 24, 33, 45, 50, 54, 58, 64, 75, 78, 87, 97]]target = [[11, 35, 12, 39, 53, 18, 48, 26, 24, 29, 9, 50], [42, 27, 84, 60, 87, 49, 94, 52, 68, 54, 44, 76]]
+Epoch 0: |                                                                          | 2889/? [2:58:59<00:00,  0.27it/s, v_num=cys0]train step 2890; scene = [['294bfa0dd8a9eada'], ['a069f40a4a017b66']]; loss = 0.058155
+Epoch 0: |                                                                          | 2890/? [2:59:02<00:00,  0.27it/s, v_num=cys0]context = [[3, 8, 10, 16, 25, 28, 29, 35, 36, 37, 43, 55, 60, 69, 72, 75, 82, 84, 87, 90, 91, 94, 95, 100]]target = [[34, 96, 98, 26, 86, 19, 46, 11, 4, 54, 8, 29, 87, 56, 9, 25, 82, 49, 6, 37, 32, 43, 21, 38]]
+Epoch 0: |                                                                          | 2899/? [2:59:35<00:00,  0.27it/s, v_num=cys0]train step 2900; scene = [['cf6618aadac4ddd9'], ['f7e27052900e847e'], ['0b173d0c5951a7f5'], ['3551ff5b8a497fb7']]; loss = 0.039044
+Epoch 0: |                                                                          | 2900/? [2:59:39<00:00,  0.27it/s, v_num=cys0]context = [[38, 95, 96], [42, 65, 102], [13, 22, 77], [198, 232, 261], [13, 56, 82], [7, 21, 79], [0, 49, 80], [83, 129, 157]]target = [[85, 41, 90], [74, 61, 87], [54, 41, 58], [242, 233, 231], [20, 63, 48], [10, 63, 69], [44, 7, 71], [89, 154, 86]]
+Epoch 0: |                                                                          | 2909/? [3:00:12<00:00,  0.27it/s, v_num=cys0]train step 2910; scene = [['b5924605972475e2'], ['64554b0854be0a81']]; loss = 0.034888
+Epoch 0: |                                                                          | 2910/? [3:00:16<00:00,  0.27it/s, v_num=cys0]context = [[0, 15, 38, 47], [103, 127, 147, 175], [174, 186, 243, 250], [14, 30, 57, 62], [88, 115, 122, 134], [64, 65, 75, 127]]target = [[11, 38, 21, 27], [165, 125, 140, 149], [233, 239, 215, 225], [41, 57, 27, 47], [94, 121, 104, 109], [65, 69, 83, 107]]
+Epoch 0: |                                                                          | 2919/? [3:00:49<00:00,  0.27it/s, v_num=cys0]train step 2920; scene = [['d53f3d87d749e474']]; loss = 0.040880
+Epoch 0: |                                                                          | 2920/? [3:00:53<00:00,  0.27it/s, v_num=cys0]context = [[94, 95, 106, 107, 112, 114, 118, 121, 124, 126, 128, 132, 137, 138, 147, 148, 153, 156, 173, 175, 176, 184, 188, 191]]target = [[164, 111, 98, 129, 114, 99, 141, 146, 144, 153, 96, 149, 109, 122, 151, 187, 136, 167, 132, 176, 95, 148, 126, 181]]
+Epoch 0: |                                                                          | 2929/? [3:01:25<00:00,  0.27it/s, v_num=cys0]train step 2930; scene = [['e61532beb3fa8b63'], ['c0803f4d1cb0eeea']]; loss = 0.030268
+Epoch 0: |                                                                          | 2930/? [3:01:29<00:00,  0.27it/s, v_num=cys0]context = [[16, 17, 22, 25, 31, 34, 37, 38, 41, 46, 54, 55, 57, 66, 75, 82, 84, 85, 92, 101, 104, 107, 111, 113]]target = [[109, 112, 73, 65, 111, 102, 34, 28, 20, 49, 46, 85, 17, 81, 50, 58, 22, 89, 52, 19, 63, 87, 59, 25]]
+Epoch 0: |                                                                          | 2939/? [3:02:02<00:00,  0.27it/s, v_num=cys0]train step 2940; scene = [['73aee8654106974f']]; loss = 0.044918
+Epoch 0: |                                                                          | 2940/? [3:02:05<00:00,  0.27it/s, v_num=cys0]context = [[80, 83, 85, 91, 93, 101, 104, 109, 110, 118, 125, 131, 133, 138, 140, 142, 147, 160, 163, 164, 167, 168, 173, 177]]target = [[172, 125, 159, 124, 146, 83, 138, 151, 164, 144, 84, 121, 153, 91, 105, 107, 109, 163, 131, 114, 89, 94, 98, 108]]
+Epoch 0: |                                                                          | 2949/? [3:02:39<00:00,  0.27it/s, v_num=cys0]train step 2950; scene = [['79d5ea6cd3f0fdb2'], ['85fb2f64303a1388'], ['9e3c241e9f50165d'], ['6a94bfa75e7988c8'], ['270022f0b06e71d5'], ['02f1d3b1d43877df'], ['6f4cc17690dcdd2e'], ['70c5a81e8b7868cc'], ['63982f095d5089b0'], ['f7498ea452fed198'], ['d07aa4d1691ccf58'], ['5474e6cd7ccd6d1a']]; loss = 0.064936
+Epoch 0: |                                                                          | 2950/? [3:02:42<00:00,  0.27it/s, v_num=cys0]context = [[1, 20, 21, 24, 25, 26, 27, 30, 36, 39, 50, 53], [60, 74, 75, 77, 82, 90, 99, 106, 111, 113, 114, 133]]target = [[12, 37, 4, 3, 45, 18, 13, 11, 51, 27, 44, 24], [74, 66, 127, 95, 119, 112, 130, 78, 75, 77, 114, 97]]
+Epoch 0: |                                                                          | 2959/? [3:03:16<00:00,  0.27it/s, v_num=cys0]train step 2960; scene = [['0e5f43adc84b0435']]; loss = 0.070053
+Epoch 0: |                                                                          | 2960/? [3:03:20<00:00,  0.27it/s, v_num=cys0]context = [[73, 79, 98, 99, 100, 103, 107, 126], [40, 42, 46, 63, 67, 92, 93, 118], [3, 7, 12, 36, 38, 39, 41, 59]]target = [[113, 108, 119, 82, 81, 84, 86, 124], [90, 106, 43, 103, 69, 87, 113, 111], [18, 41, 24, 30, 7, 20, 53, 54]]
+Epoch 0: |                                                                          | 2969/? [3:03:53<00:00,  0.27it/s, v_num=cys0]train step 2970; scene = [['d3784c9108c25d42']]; loss = 0.038727
+Epoch 0: |                                                                          | 2970/? [3:03:57<00:00,  0.27it/s, v_num=cys0]context = [[10, 11, 17, 29, 30, 31, 33, 39, 48, 60, 63, 64, 69, 70, 72, 87, 91, 92, 93, 95, 99, 100, 104, 107]]target = [[63, 97, 21, 50, 94, 32, 27, 70, 60, 78, 65, 45, 53, 85, 56, 33, 38, 80, 98, 12, 44, 95, 43, 34]]
+Epoch 0: |                                                                          | 2979/? [3:04:31<00:00,  0.27it/s, v_num=cys0]train step 2980; scene = [['058bac6c226fc1a9'], ['7d900c809e896e32']]; loss = 0.038469
+Epoch 0: |                                                                          | 2980/? [3:04:34<00:00,  0.27it/s, v_num=cys0]context = [[64, 69, 70, 71, 83, 85, 86, 93, 94, 96, 111, 112, 113, 117, 123, 138, 139, 140, 141, 148, 149, 153, 159, 161]]target = [[97, 127, 104, 150, 132, 129, 78, 128, 116, 125, 134, 130, 149, 99, 137, 79, 88, 140, 94, 98, 142, 153, 155, 73]]
+Epoch 0: |                                                                          | 2989/? [3:05:07<00:00,  0.27it/s, v_num=cys0]train step 2990; scene = [['51f581faf9000425'], ['89bc971b5dfbd294'], ['511207c07d553599']]; loss = 0.048343
+Epoch 0: |                                                                          | 2990/? [3:05:10<00:00,  0.27it/s, v_num=cys0]context = [[27, 42, 44, 84, 95, 99], [168, 178, 219, 221, 222, 224], [50, 56, 89, 100, 116, 131], [117, 131, 156, 157, 160, 171]]target = [[41, 88, 49, 83, 97, 46], [194, 197, 187, 203, 199, 177], [130, 88, 129, 93, 52, 103], [130, 162, 157, 127, 156, 133]]
+Epoch 0: |                                                                          | 2999/? [3:05:43<00:00,  0.27it/s, v_num=cys0]train step 3000; scene = [['d77284dc9b9d1031']]; loss = 0.086485
+Epoch 0: |                                                                          | 3000/? [3:05:46<00:00,  0.27it/s, v_num=cys0]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 3000; scene = ['a76028640ffa1ef9'];
+target intrinsic:  tensor(0.8569, device='cuda:0') tensor(0.8571, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8954, device='cuda:0') tensor(0.8959, device='cuda:0')
+[2026-02-25 10:38:39,652][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 3000/? [3:06:01<00:00,  0.27it/s, v_num=cys0]context = [[176, 195, 196, 203, 214, 228, 231, 237], [75, 81, 89, 91, 113, 117, 152, 155], [54, 61, 76, 79, 90, 95, 104, 105]]target = [[207, 201, 216, 186, 203, 231, 198, 179], [144, 150, 79, 87, 147, 123, 115, 138], [86, 82, 99, 72, 100, 74, 66, 62]]
+[2026-02-25 10:38:43,134][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 3001/? [3:06:06<00:00,  0.27it/s, v_num=cys0]
+`Trainer.fit` stopped: `max_steps=3001` reached.
+Peak VRAM: 96.070 GB (allocated), 136.279 GB (reserved)
+Total elapsed: 3.12 hours
+Saved memory info to: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_noRefineModule/peak_vram_memory.json
diff --git a/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/requirements.txt b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..7fbf9096f92b53f8bb2a7e5467c79ecbe64faca5
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/requirements.txt
@@ -0,0 +1,172 @@
+wheel==0.45.1
+pytz==2025.2
+easydict==1.13
+antlr4-python3-runtime==4.9.3
+wadler_lindig==0.1.7
+urllib3==2.5.0
+tzdata==2025.2
+typing-inspection==0.4.1
+tabulate==0.9.0
+smmap==5.0.2
+kornia_rs==0.1.9
+setuptools==78.1.1
+safetensors==0.5.3
+PyYAML==6.0.2
+PySocks==1.7.1
+pyparsing==3.2.5
+pydantic_core==2.33.2
+pycparser==2.23
+protobuf==6.32.1
+propcache==0.3.2
+proglog==0.1.12
+fsspec==2024.6.1
+platformdirs==4.4.0
+pip==25.2
+pillow==10.4.0
+frozenlist==1.7.0
+packaging==24.2
+opt_einsum==3.4.0
+numpy==1.26.4
+ninja==1.13.0
+fonttools==4.60.0
+networkx==3.4.2
+multidict==6.6.4
+mdurl==0.1.2
+MarkupSafe==3.0.2
+kiwisolver==1.4.9
+imageio-ffmpeg==0.6.0
+idna==3.7
+hf-xet==1.1.10
+gmpy2==2.2.1
+einops==0.8.1
+filelock==3.17.0
+decorator==4.4.2
+dacite==1.9.2
+cycler==0.12.1
+colorama==0.4.6
+click==8.3.0
+nvidia-nvtx-cu12==12.8.90
+charset-normalizer==3.3.2
+certifi==2025.8.3
+beartype==0.19.0
+attrs==25.3.0
+async-timeout==5.0.1
+annotated-types==0.7.0
+aiohappyeyeballs==2.6.1
+yarl==1.20.1
+tifffile==2025.5.10
+sentry-sdk==2.39.0
+scipy==1.15.3
+pydantic==2.11.9
+pandas==2.3.2
+opencv-python==4.11.0.86
+omegaconf==2.3.0
+markdown-it-py==4.0.0
+lightning-utilities==0.14.3
+lazy_loader==0.4
+jaxtyping==0.2.37
+imageio==2.37.0
+gitdb==4.0.12
+contourpy==1.3.2
+colorspacious==1.1.2
+cffi==1.17.1
+aiosignal==1.4.0
+scikit-video==1.1.11
+scikit-image==0.25.2
+rich==14.1.0
+moviepy==1.0.3
+matplotlib==3.10.6
+hydra-core==1.3.2
+nvidia-nccl-cu12==2.27.3
+huggingface-hub==0.35.1
+GitPython==3.1.45
+brotlicffi==1.0.9.2
+aiohttp==3.12.15
+torchmetrics==1.8.2
+opt-einsum-fx==0.1.4
+kornia==0.8.1
+pytorch-lightning==2.5.1
+lpips==0.1.4
+e3nn==0.6.0
+lightning==2.5.1
+nvidia-cusparselt-cu12==0.7.1
+triton==3.4.0
+nvidia-nvjitlink-cu12==12.8.93
+nvidia-curand-cu12==10.3.9.90
+nvidia-cufile-cu12==1.13.1.3
+nvidia-cuda-runtime-cu12==12.8.90
+nvidia-cuda-nvrtc-cu12==12.8.93
+nvidia-cuda-cupti-cu12==12.8.90
+nvidia-cublas-cu12==12.8.4.1
+nvidia-cusparse-cu12==12.5.8.93
+nvidia-cufft-cu12==11.3.3.83
+nvidia-cudnn-cu12==9.10.2.21
+nvidia-cusolver-cu12==11.7.3.90
+torch==2.8.0+cu128
+torchvision==0.23.0+cu128
+torchaudio==2.8.0+cu128
+torch_scatter==2.1.2+pt28cu128
+gsplat==1.5.3
+wandb==0.25.0
+cuda-bindings==13.0.3
+cuda-pathfinder==1.3.3
+Jinja2==3.1.6
+mpmath==1.3.0
+nvidia-cublas==13.1.0.3
+nvidia-cuda-cupti==13.0.85
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-runtime==13.0.96
+nvidia-cudnn-cu13==9.15.1.9
+nvidia-cufft==12.0.0.61
+nvidia-cufile==1.15.1.6
+nvidia-curand==10.4.0.35
+nvidia-cusolver==12.0.4.66
+nvidia-cusparse==12.6.3.3
+nvidia-cusparselt-cu13==0.8.0
+nvidia-nccl-cu13==2.28.9
+nvidia-nvjitlink==13.0.88
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvtx==13.0.85
+requests==2.32.5
+sentencepiece==0.2.1
+sympy==1.14.0
+torchcodec==0.10.0
+torchdata==0.10.0
+torchtext==0.6.0
+anyio==4.12.0
+asttokens==3.0.1
+comm==0.2.3
+debugpy==1.8.19
+executing==2.2.1
+h11==0.16.0
+httpcore==1.0.9
+httpx==0.28.1
+ipykernel==7.1.0
+ipython==9.8.0
+ipython_pygments_lexers==1.1.1
+ipywidgets==8.1.8
+jedi==0.19.2
+jupyter_client==8.7.0
+jupyter_core==5.9.1
+jupyterlab_widgets==3.0.16
+matplotlib-inline==0.2.1
+nest-asyncio==1.6.0
+parso==0.8.5
+pexpect==4.9.0
+prompt_toolkit==3.0.52
+psutil==7.2.1
+ptyprocess==0.7.0
+pure_eval==0.2.3
+Pygments==2.19.2
+python-dateutil==2.9.0.post0
+pyzmq==27.1.0
+shellingham==1.5.4
+six==1.17.0
+stack-data==0.6.3
+tornado==6.5.4
+tqdm==4.67.1
+traitlets==5.14.3
+typer-slim==0.21.0
+typing_extensions==4.15.0
+wcwidth==0.2.14
+widgetsnbextension==4.0.15
diff --git a/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/wandb-metadata.json b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/wandb-metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..3ad332567fe61601e92ddfb98375e122fc8d6df7
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/wandb-metadata.json
@@ -0,0 +1,93 @@
+{
+  "os":  "Linux-6.8.0-90-generic-x86_64-with-glibc2.39",
+  "python":  "CPython 3.12.12",
+  "startedAt":  "2026-02-25T07:32:27.352870Z",
+  "args":  [
+    "+experiment=re10k_ablation_24v",
+    "wandb.mode=online",
+    "wandb.name=ABLATION_0225_noRefineModule",
+    "model.density_control.use_refine_module=false"
+  ],
+  "program":  "-m src.main",
+  "git":  {
+    "remote":  "git@github.com:K-nowing/CVPR2026.git",
+    "commit":  "2512754c6c27ca5150bf17fbcbdde3f192fd53cc"
+  },
+  "email":  "dna9041@korea.ac.kr",
+  "root":  "/workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_noRefineModule",
+  "host":  "27d18dedec6d",
+  "executable":  "/venv/main/bin/python",
+  "cpu_count":  128,
+  "cpu_count_logical":  256,
+  "gpu":  "NVIDIA H200",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "1170378588160",
+      "used":  "708558733312"
+    }
+  },
+  "memory":  {
+    "total":  "1622948257792"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-2649ab80-a3a6-5a1c-0fa5-12bc11bd75e9"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-e92921d9-c681-246f-af93-637e0dc938ca"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-ffe12ffc-9bb7-82de-5692-1ec0ee2e68d8"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-499e5acd-b6ab-2010-c51b-ee9b5aa65825"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-3b2522d9-1c72-e49b-2c30-96165680b74a"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-a9a280c5-b2f9-dc1e-a8a9-7326a74001ff"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-07d0167b-a6a1-1900-2d27-7c6c11598409"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-8362a999-20d1-c27b-5d18-032d23f859ab"
+    }
+  ],
+  "cudaVersion":  "13.1",
+  "writerId":  "z1winms0ab80rmcbaynf075otkwpygrq"
+}
\ No newline at end of file
diff --git a/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/wandb-summary.json b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/wandb-summary.json
new file mode 100644
index 0000000000000000000000000000000000000000..d7542eb901622c8f6a257d7b32fee515d175036c
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/files/wandb-summary.json
@@ -0,0 +1 @@
+{"trainer/global_step":3001,"_wandb":{"runtime":11183},"loss/aux_1/lpips":0.009582722559571266,"loss/aux_1/error_score":0.23707404732704163,"loss/aux_0/error_score":0.37395116686820984,"_timestamp":1.7720159252726524e+09,"train/psnr_probabilistic":20.766376495361328,"loss/aux_1/mse":0.011007795110344887,"val/psnr":21.622608184814453,"train/error_scores":{"count":1,"filenames":["media/images/train/error_scores_201_99cdf460841ea0543ea7.png"],"captions":[["0621c7675fab1418"]],"_type":"images/separated","width":1328,"height":2120,"format":"png"},"active_mask_imgs":{"filenames":["media/images/active_mask_imgs_198_d9a5bace8f25f1101b30.png"],"captions":["a76028640ffa1ef9"],"_type":"images/separated","width":536,"height":800,"format":"png","count":1},"loss/aux_0/lpips":0.010782335884869099,"lr-AdamW/pg1-momentum":0.9,"epoch":0,"loss/total":0.08648455888032913,"lr-AdamW/pg2":2e-05,"loss/final_3dgs/mse":0.009266000241041183,"error_scores":{"_type":"images/separated","width":800,"height":536,"format":"png","count":1,"filenames":["media/images/error_scores_199_e79b447934cce3e14bdb.png"],"captions":["a76028640ffa1ef9"]},"train/comparison":{"_type":"images/separated","width":1328,"height":2154,"format":"png","count":1,"filenames":["media/images/train/comparison_202_b6bf8b4d2d9219d977fa.png"],"captions":[["0621c7675fab1418"]]},"lr-AdamW/pg2-momentum":0.9,"train/scene_scale":1.0070030689239502,"comparison":{"count":1,"filenames":["media/images/comparison_197_d1042a2aa788751a412f.png"],"captions":["a76028640ffa1ef9"],"_type":"images/separated","width":1064,"height":1098,"format":"png"},"val/gaussian_num_ratio":0.3997650146484375,"loss/aux_0/mse":0.009491334669291973,"_runtime":11183,"info/global_step":3000,"_step":202,"loss/aux_2/mse":0.010797698982059956,"loss/scene_scale_reg":0.00019978880300186574,"lr-AdamW/pg1":2.003594834351718e-05,"val/ssim":0.8318922519683838,"loss/final_3dgs/lpips":0.00893520936369896,"loss/aux_2/lpips":0.009327557869255543,"loss/camera":9.838641562964767e-05,"val/lpips":0.1536986380815506}
\ No newline at end of file
diff --git a/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug-core.log b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug-core.log
new file mode 100644
index 0000000000000000000000000000000000000000..b39b405a3f0902a6bc141d33c9d57bc6d90dc4f2
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug-core.log
@@ -0,0 +1,15 @@
+{"time":"2026-02-25T07:32:27.422522053Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpwu965jc4/port-137621.txt","pid":137621,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-02-25T07:32:27.423426767Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":137621}
+{"time":"2026-02-25T07:32:27.423393077Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-137621-140053-2081743564/socket","Net":"unix"}}
+{"time":"2026-02-25T07:32:27.602024695Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-02-25T07:32:27.611595513Z","level":"INFO","msg":"handleInformInit: received","streamId":"2f0bcys0","id":"1(@)"}
+{"time":"2026-02-25T07:32:28.037979247Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"2f0bcys0","id":"1(@)"}
+{"time":"2026-02-25T07:32:33.742044945Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"v0xp4cjc1l9g"}
+{"time":"2026-02-25T10:38:52.520680299Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
+{"time":"2026-02-25T10:38:52.520838241Z","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-02-25T10:38:52.520822771Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
+{"time":"2026-02-25T10:38:52.520922373Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
+{"time":"2026-02-25T10:38:52.520970143Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-137621-140053-2081743564/socket","Net":"unix"}}
+{"time":"2026-02-25T10:38:53.701442926Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
+{"time":"2026-02-25T10:38:53.701488686Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
+{"time":"2026-02-25T10:38:53.701513197Z","level":"INFO","msg":"server is closed"}
diff --git a/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug-internal.log b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..77a6a5015e15fecf0a7233dd1c15eced819110aa
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug-internal.log
@@ -0,0 +1,11 @@
+{"time":"2026-02-25T07:32:27.611867617Z","level":"INFO","msg":"stream: starting","core version":"0.25.0"}
+{"time":"2026-02-25T07:32:28.03755666Z","level":"INFO","msg":"stream: created new stream","id":"2f0bcys0"}
+{"time":"2026-02-25T07:32:28.037863635Z","level":"INFO","msg":"handler: started","stream_id":"2f0bcys0"}
+{"time":"2026-02-25T07:32:28.037970207Z","level":"INFO","msg":"stream: started","id":"2f0bcys0"}
+{"time":"2026-02-25T07:32:28.038020847Z","level":"INFO","msg":"writer: started","stream_id":"2f0bcys0"}
+{"time":"2026-02-25T07:32:28.038027757Z","level":"INFO","msg":"sender: started","stream_id":"2f0bcys0"}
+{"time":"2026-02-25T10:38:52.520830581Z","level":"INFO","msg":"stream: closing","id":"2f0bcys0"}
+{"time":"2026-02-25T10:38:53.390340772Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-02-25T10:38:53.699950002Z","level":"INFO","msg":"handler: closed","stream_id":"2f0bcys0"}
+{"time":"2026-02-25T10:38:53.700227926Z","level":"INFO","msg":"sender: closed","stream_id":"2f0bcys0"}
+{"time":"2026-02-25T10:38:53.700251656Z","level":"INFO","msg":"stream: closed","id":"2f0bcys0"}
diff --git a/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug.log b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..2dab590a5f31eeb0e6d17e20ae047b827a5d4d4d
--- /dev/null
+++ b/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug.log
@@ -0,0 +1,21 @@
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_setup.py:_flush():81] Current SDK version is 0.25.0
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_setup.py:_flush():81] Configure stats pid to 137621
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug.log
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_noRefineModule/wandb/run-20260225_073227-2f0bcys0/logs/debug-internal.log
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_init.py:init():844] calling init triggers
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
+config: {'model': {'encoder': {'name': 'dcsplat', 'input_image_shape': [518, 518], 'head_mode': 'pcd', 'num_level': 3, 'gs_param_dim': 256, 'align_corners': False, 'use_voxelize': True}, 'decoder': {'name': 'splatting_cuda', 'background_color': [0.0, 0.0, 0.0], 'make_scale_invariant': False}, 'density_control': {'name': 'density_control_module', 'mean_dim': 32, 'gs_param_dim': 256, 'refinement_layer_num': 1, 'num_level': 3, 'grad_mode': 'absgrad', 'use_mean_features': True, 'refinement_type': 'voxelize', 'refinement_hidden_dim': 32, 'aggregation_mode': 'mean', 'num_heads': 1, 'score_mode': 'absgrad', 'latent_dim': 128, 'num_latents': 64, 'num_self_attn_per_block': 2, 'voxel_size': 0.001, 'aux_refine': False, 'refine_error': False, 'use_refine_module': False, 'voxelize_activate': True, 'use_depth': False}}, 'render_loss': {'mse': {'weight': 1.0}, 'lpips': {'weight': 0.05, 'apply_after_step': 0}}, 'density_control_loss': {'error_score': {'weight': 0.01, 'log_scale': False, 'grad_scale': 10000.0, 'mode': 'original'}}, 'direct_loss': {'l1': {'weight': 0.8}, 'ssim': {'weight': 0.2}}, 'wandb': {'project': 'DCSplat', 'entity': 'scene-representation-group', 'name': 'ABLATION_0225_noRefineModule', 'mode': 'online', 'tags': ['re10k', '256x256']}, 'mode': 'train', 'data_loader': {'train': {'num_workers': 16, 'persistent_workers': True, 'batch_size': 16, 'seed': 1234}, 'test': {'num_workers': 4, 'persistent_workers': False, 'batch_size': 1, 'seed': 2345}, 'val': {'num_workers': 1, 'persistent_workers': True, 'batch_size': 1, 'seed': 3456}}, 'optimizer': {'lr': 0.0002, 'warm_up_steps': 25, 'backbone_lr_multiplier': 0.1, 'backbone_trainable': 'T+H', 'accumulate': 1}, 'checkpointing': {'load': None, 'every_n_train_steps': 1500, 'save_top_k': 2, 'save_weights_only': False}, 'train': {'extended_visualization': False, 'print_log_every_n_steps': 10, 'camera_loss': 10.0, 'one_sample_validation': None, 'align_corners': False, 'intrinsic_scaling': False, 'verbose': False, 'beta_dist_param': [0.5, 4.0], 'use_refine_aux': False, 'train_target_set': True, 'train_gs_num': 1, 'ext_scale_detach': False, 'cam_scale_mode': 'sum', 'scene_scale_reg_loss': 0.01, 'train_aux': True, 'vggt_cam_loss': True, 'vggt_distil': False, 'context_view_train': False}, 'test': {'output_path': 'test/ablation/re10k', 'align_pose': False, 'pose_align_steps': 100, 'rot_opt_lr': 0.005, 'trans_opt_lr': 0.005, 'compute_scores': True, 'save_image': False, 'save_video': False, 'save_active_mask_image': False, 'save_error_score_image': False, 'save_compare': False, 'pred_intrinsic': False, 'error_threshold': 0.4, 'error_threshold_list': [0.2, 0.4, 0.6, 0.8, 1.0], 'threshold_mode': 'ratio', 'nvs_view_N_list': [3, 6, 16, 32, 64]}, 'seed': 111123, 'trainer': {'max_steps': 3001, 'val_check_interval': 250, 'gradient_clip_val': 0.5, 'num_nodes': 1}, 'dataset': {'re10k': {'make_baseline_1': True, 'relative_pose': True, 'augment': True, 'background_color': [0.0, 0.0, 0.0], 'overfit_to_scene': None, 'skip_bad_shape': True, 'view_sampler': {'name': 'bounded', 'num_target_views': 4, 'num_context_views': 2, 'min_distance_between_context_views': 45, 'max_distance_between_context_views': 90, 'min_distance_to_context_views': 0, 'warm_up_steps': 1000, 'initial_min_distance_between_context_views': 25, 'initial_max_distance_between_context_views': 25, 'same_target_gap': False, 'num_target_set': 3}, 'name': 're10k', 'roots': ['datasets/re10k'], 'input_image_shape': [256, 256], 'original_image_shape': [360, 640], 'cameras_are_circular': False, 'baseline_min': 0.001, 'baseline_max': 10000000000.0, 'max_fov': 100.0, 'dynamic_context_views': True, 'max_context_views_per_gpu': 24}}, '_wandb': {}}
+2026-02-25 07:32:27,354 INFO    MainThread:137621 [wandb_init.py:init():892] starting backend
+2026-02-25 07:32:27,602 INFO    MainThread:137621 [wandb_init.py:init():895] sending inform_init request
+2026-02-25 07:32:27,609 INFO    MainThread:137621 [wandb_init.py:init():903] backend started and connected
+2026-02-25 07:32:27,613 INFO    MainThread:137621 [wandb_init.py:init():973] updated telemetry
+2026-02-25 07:32:27,622 INFO    MainThread:137621 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
+2026-02-25 07:32:28,628 INFO    MainThread:137621 [wandb_init.py:init():1042] starting run threads in backend
+2026-02-25 07:32:28,738 INFO    MainThread:137621 [wandb_run.py:_console_start():2524] atexit reg
+2026-02-25 07:32:28,738 INFO    MainThread:137621 [wandb_run.py:_redirect():2373] redirect: wrap_raw
+2026-02-25 07:32:28,738 INFO    MainThread:137621 [wandb_run.py:_redirect():2442] Wrapping output streams.
+2026-02-25 07:32:28,738 INFO    MainThread:137621 [wandb_run.py:_redirect():2465] Redirects installed.
+2026-02-25 07:32:28,740 INFO    MainThread:137621 [wandb_init.py:init():1082] run started, returning control to user process
+2026-02-25 10:38:52,520 INFO    wandb-AsyncioManager-main:137621 [service_client.py:_forward_responses():134] Reached EOF.
+2026-02-25 10:38:52,520 INFO    wandb-AsyncioManager-main:137621 [mailbox.py:close():155] Closing mailbox, abandoning 1 handles.
diff --git a/ABLATION_0225_randomSelect/main.log b/ABLATION_0225_randomSelect/main.log
new file mode 100644
index 0000000000000000000000000000000000000000..2e2d669bf1bb7da1cbf2cd2c01c438733d4735c3
--- /dev/null
+++ b/ABLATION_0225_randomSelect/main.log
@@ -0,0 +1,116 @@
+[2026-02-25 10:39:03,453][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 10:39:09,556][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 10:39:09,556][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 10:39:59,700][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/data_connector.py:425: The 'val_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=31` in the `DataLoader` to improve performance.
+
+[2026-02-25 10:39:59,701][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 10:40:02,283][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:40:02,292][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/utilities/data.py:79: Trying to infer the `batch_size` from an ambiguous collection. The batch size we found is 1. To avoid any miscalculations, use `self.log(..., batch_size=batch_size)`.
+
+[2026-02-25 10:40:02,292][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 10:40:02,293][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 10:40:03,984][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/functional.py:554: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /pytorch/aten/src/ATen/native/TensorShape.cpp:4322.)
+  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
+
+[2026-02-25 10:40:04,284][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/psnr', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 10:40:04,285][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/lpips', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 10:40:04,286][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/ssim', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 10:40:04,286][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/gaussian_num_ratio', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 10:40:04,286][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('info/global_step', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 10:40:13,358][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [256, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [256, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 10:40:13,429][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:41:49,278][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 10:52:55,864][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:56:11,418][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:05:44,777][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:12:03,856][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:18:29,813][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:27:57,672][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:31:13,251][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:43:50,712][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:43:54,528][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:56:39,754][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:59:50,137][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:09:23,940][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:16:01,606][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:22:24,120][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:32:00,789][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:35:14,407][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:47:51,723][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:47:56,086][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:00:42,868][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:03:57,852][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:13:32,568][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:19:49,814][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:26:12,524][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_randomSelect/train_ddp_process_1.log b/ABLATION_0225_randomSelect/train_ddp_process_1.log
new file mode 100644
index 0000000000000000000000000000000000000000..5a1d5483ced6e1918676dc16f0a0007e4e3c1bd3
--- /dev/null
+++ b/ABLATION_0225_randomSelect/train_ddp_process_1.log
@@ -0,0 +1,60 @@
+[2026-02-25 10:39:19,964][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 10:39:37,919][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 10:39:37,920][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 10:39:59,701][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 10:40:13,353][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [256, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [256, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 10:40:13,462][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:41:49,277][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 10:52:55,864][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:05:44,777][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:18:29,813][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:31:13,252][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:43:54,528][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:56:39,755][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:09:23,940][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:22:24,117][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:35:14,407][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:47:56,084][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:00:42,869][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:13:32,567][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:26:12,524][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_randomSelect/train_ddp_process_2.log b/ABLATION_0225_randomSelect/train_ddp_process_2.log
new file mode 100644
index 0000000000000000000000000000000000000000..111b7892cd182ebe0ee5b10b4e2ce6fb8d5e4cf2
--- /dev/null
+++ b/ABLATION_0225_randomSelect/train_ddp_process_2.log
@@ -0,0 +1,60 @@
+[2026-02-25 10:39:19,879][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 10:39:48,721][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 10:39:48,721][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 10:39:59,701][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 10:40:13,356][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [256, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [256, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 10:40:13,461][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:41:49,278][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 10:52:55,864][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:05:44,777][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:18:29,813][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:31:13,250][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:43:54,529][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:56:39,755][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:09:23,942][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:22:24,117][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:35:14,408][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:47:56,084][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:00:42,869][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:13:32,568][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:26:12,524][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_randomSelect/train_ddp_process_4.log b/ABLATION_0225_randomSelect/train_ddp_process_4.log
new file mode 100644
index 0000000000000000000000000000000000000000..05c88107de5ea50a790e586c8793ebb3410028f7
--- /dev/null
+++ b/ABLATION_0225_randomSelect/train_ddp_process_4.log
@@ -0,0 +1,60 @@
+[2026-02-25 10:39:19,958][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 10:39:38,019][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 10:39:38,020][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 10:39:59,701][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 10:40:13,354][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [256, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [256, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 10:40:13,462][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:41:49,306][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 10:52:55,864][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:05:44,777][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:18:29,814][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:31:13,251][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:43:54,528][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:56:39,755][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:09:23,940][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:22:24,117][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:35:14,408][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:47:56,084][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:00:42,869][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:13:32,567][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:26:12,524][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_randomSelect/train_ddp_process_5.log b/ABLATION_0225_randomSelect/train_ddp_process_5.log
new file mode 100644
index 0000000000000000000000000000000000000000..d684e3f191e68bd5eb9f5eec217d1483538d2567
--- /dev/null
+++ b/ABLATION_0225_randomSelect/train_ddp_process_5.log
@@ -0,0 +1,60 @@
+[2026-02-25 10:39:19,922][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 10:39:48,615][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 10:39:48,615][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 10:39:59,701][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 10:40:12,614][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [256, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [256, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 10:40:13,463][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:41:49,305][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 10:52:55,865][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:05:44,777][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:18:29,814][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:31:13,251][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:43:54,528][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:56:39,756][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:09:23,941][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:22:24,117][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:35:14,407][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:47:56,084][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:00:42,869][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:13:32,570][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:26:12,524][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_randomSelect/train_ddp_process_6.log b/ABLATION_0225_randomSelect/train_ddp_process_6.log
new file mode 100644
index 0000000000000000000000000000000000000000..0c37bdeb2e7fd15e09653c5fdbdcf73862ecc3c4
--- /dev/null
+++ b/ABLATION_0225_randomSelect/train_ddp_process_6.log
@@ -0,0 +1,60 @@
+[2026-02-25 10:39:19,844][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 10:39:46,759][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 10:39:46,759][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 10:39:59,701][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 10:40:12,842][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [256, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [256, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 10:40:13,462][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 10:41:49,277][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 10:52:55,864][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:05:44,777][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:18:29,813][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:31:13,251][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:43:54,528][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 11:56:39,755][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:09:23,940][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:22:24,118][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:35:14,410][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 12:47:56,084][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:00:42,869][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:13:32,567][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 13:26:12,524][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_targetTrain/.hydra/config.yaml b/ABLATION_0225_targetTrain/.hydra/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..c7490607029458c8748e63cd2a4c04724c4daf09
--- /dev/null
+++ b/ABLATION_0225_targetTrain/.hydra/config.yaml
@@ -0,0 +1,185 @@
+model:
+  encoder:
+    name: dcsplat
+    input_image_shape:
+    - 518
+    - 518
+    head_mode: pcd
+    num_level: 3
+    gs_param_dim: 256
+    align_corners: false
+    use_voxelize: true
+  decoder:
+    name: splatting_cuda
+    background_color:
+    - 0.0
+    - 0.0
+    - 0.0
+    make_scale_invariant: false
+  density_control:
+    name: density_control_module
+    mean_dim: 32
+    gs_param_dim: 256
+    refinement_layer_num: 1
+    num_level: 3
+    grad_mode: absgrad
+    use_mean_features: true
+    refinement_type: voxelize
+    refinement_hidden_dim: 32
+    aggregation_mode: mean
+    num_heads: 1
+    score_mode: absgrad
+    latent_dim: 128
+    num_latents: 64
+    num_self_attn_per_block: 2
+    voxel_size: 0.001
+    aux_refine: false
+    refine_error: false
+    use_refine_module: true
+    voxelize_activate: true
+    use_depth: false
+render_loss:
+  mse:
+    weight: 1.0
+  lpips:
+    weight: 0.05
+    apply_after_step: 0
+density_control_loss:
+  error_score:
+    weight: 0.01
+    log_scale: false
+    grad_scale: 10000.0
+    mode: original
+direct_loss:
+  l1:
+    weight: 0.8
+  ssim:
+    weight: 0.2
+wandb:
+  project: DCSplat
+  entity: scene-representation-group
+  name: ABLATION_0225_targetTrain
+  mode: online
+  tags:
+  - re10k
+  - 256x256
+mode: train
+data_loader:
+  train:
+    num_workers: 16
+    persistent_workers: true
+    batch_size: 16
+    seed: 1234
+  test:
+    num_workers: 4
+    persistent_workers: false
+    batch_size: 1
+    seed: 2345
+  val:
+    num_workers: 1
+    persistent_workers: true
+    batch_size: 1
+    seed: 3456
+optimizer:
+  lr: 0.0002
+  warm_up_steps: 25
+  backbone_lr_multiplier: 0.1
+  backbone_trainable: T+H
+  accumulate: 1
+checkpointing:
+  load: null
+  every_n_train_steps: 1500
+  save_top_k: 2
+  save_weights_only: false
+train:
+  extended_visualization: false
+  print_log_every_n_steps: 10
+  camera_loss: 10.0
+  one_sample_validation: null
+  align_corners: false
+  intrinsic_scaling: false
+  verbose: false
+  beta_dist_param:
+  - 0.5
+  - 4.0
+  use_refine_aux: false
+  train_target_set: true
+  train_gs_num: 1
+  ext_scale_detach: false
+  cam_scale_mode: sum
+  scene_scale_reg_loss: 0.0
+  train_aux: false
+  vggt_cam_loss: true
+  vggt_distil: false
+  context_view_train: false
+test:
+  output_path: test/ablation/re10k
+  align_pose: false
+  pose_align_steps: 100
+  rot_opt_lr: 0.005
+  trans_opt_lr: 0.005
+  compute_scores: true
+  save_image: false
+  save_video: false
+  save_active_mask_image: false
+  save_error_score_image: false
+  save_compare: false
+  pred_intrinsic: false
+  error_threshold: 0.4
+  error_threshold_list:
+  - 0.2
+  - 0.4
+  - 0.6
+  - 0.8
+  - 1.0
+  threshold_mode: ratio
+  nvs_view_N_list:
+  - 3
+  - 6
+  - 16
+  - 32
+  - 64
+seed: 111123
+trainer:
+  max_steps: 3001
+  val_check_interval: 250
+  gradient_clip_val: 0.5
+  num_nodes: 1
+dataset:
+  re10k:
+    make_baseline_1: true
+    relative_pose: true
+    augment: true
+    background_color:
+    - 0.0
+    - 0.0
+    - 0.0
+    overfit_to_scene: null
+    skip_bad_shape: true
+    view_sampler:
+      name: bounded
+      num_target_views: 4
+      num_context_views: 2
+      min_distance_between_context_views: 45
+      max_distance_between_context_views: 90
+      min_distance_to_context_views: 0
+      warm_up_steps: 1000
+      initial_min_distance_between_context_views: 25
+      initial_max_distance_between_context_views: 25
+      same_target_gap: false
+      num_target_set: 3
+    name: re10k
+    roots:
+    - datasets/re10k
+    input_image_shape:
+    - 256
+    - 256
+    original_image_shape:
+    - 360
+    - 640
+    cameras_are_circular: false
+    baseline_min: 0.001
+    baseline_max: 10000000000.0
+    max_fov: 100.0
+    dynamic_context_views: true
+    max_context_views_per_gpu: 24
diff --git a/ABLATION_0225_targetTrain/.hydra/hydra.yaml b/ABLATION_0225_targetTrain/.hydra/hydra.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..2f06af194bf0c5d50e2baf7eed7084ce947ebb71
--- /dev/null
+++ b/ABLATION_0225_targetTrain/.hydra/hydra.yaml
@@ -0,0 +1,166 @@
+hydra:
+  run:
+    dir: outputs/ablation/re10k/${wandb.name}
+  sweep:
+    dir: multirun/${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:
+    - +experiment=re10k_ablation_24v
+    - wandb.mode=online
+    - wandb.name=ABLATION_0225_targetTrain
+    - train.scene_scale_reg_loss=0.0
+    - train.train_aux=false
+  job:
+    name: main
+    chdir: null
+    override_dirname: +experiment=re10k_ablation_24v,train.scene_scale_reg_loss=0.0,train.train_aux=false,wandb.mode=online,wandb.name=ABLATION_0225_targetTrain
+    id: ???
+    num: ???
+    config_name: main
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.3.2
+    version_base: '1.3'
+    cwd: /workspace/code/CVPR2026
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/code/CVPR2026/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain
+    choices:
+      experiment: re10k_ablation_24v
+      dataset@dataset.re10k: re10k
+      dataset/view_sampler_dataset_specific_config@dataset.re10k.view_sampler: bounded_re10k
+      dataset/view_sampler@dataset.re10k.view_sampler: bounded
+      model/density_control: density_control_module
+      model/decoder: splatting_cuda
+      model/encoder: dcsplat
+      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
diff --git a/ABLATION_0225_targetTrain/.hydra/overrides.yaml b/ABLATION_0225_targetTrain/.hydra/overrides.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..b0fda36a6874f74408417426ffc52a71c420f97f
--- /dev/null
+++ b/ABLATION_0225_targetTrain/.hydra/overrides.yaml
@@ -0,0 +1,5 @@
+- +experiment=re10k_ablation_24v
+- wandb.mode=online
+- wandb.name=ABLATION_0225_targetTrain
+- train.scene_scale_reg_loss=0.0
+- train.train_aux=false
diff --git a/ABLATION_0225_targetTrain/wandb/debug-internal.log b/ABLATION_0225_targetTrain/wandb/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..c06eded7785be3e23e697052114b1fdf840b3371
--- /dev/null
+++ b/ABLATION_0225_targetTrain/wandb/debug-internal.log
@@ -0,0 +1,11 @@
+{"time":"2026-02-25T01:41:00.103713472Z","level":"INFO","msg":"stream: starting","core version":"0.25.0"}
+{"time":"2026-02-25T01:41:00.515197271Z","level":"INFO","msg":"stream: created new stream","id":"qetzseh9"}
+{"time":"2026-02-25T01:41:00.515336753Z","level":"INFO","msg":"handler: started","stream_id":"qetzseh9"}
+{"time":"2026-02-25T01:41:00.515557835Z","level":"INFO","msg":"stream: started","id":"qetzseh9"}
+{"time":"2026-02-25T01:41:00.515605436Z","level":"INFO","msg":"writer: started","stream_id":"qetzseh9"}
+{"time":"2026-02-25T01:41:00.515647397Z","level":"INFO","msg":"sender: started","stream_id":"qetzseh9"}
+{"time":"2026-02-25T04:35:28.261315178Z","level":"INFO","msg":"stream: closing","id":"qetzseh9"}
+{"time":"2026-02-25T04:35:29.641674674Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-02-25T04:35:29.934574479Z","level":"INFO","msg":"handler: closed","stream_id":"qetzseh9"}
+{"time":"2026-02-25T04:35:29.934901185Z","level":"INFO","msg":"sender: closed","stream_id":"qetzseh9"}
+{"time":"2026-02-25T04:35:29.934924995Z","level":"INFO","msg":"stream: closed","id":"qetzseh9"}
diff --git a/ABLATION_0225_targetTrain/wandb/debug.log b/ABLATION_0225_targetTrain/wandb/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..e5fc69f9b164886e9cc9a2671e28569335bcb3fd
--- /dev/null
+++ b/ABLATION_0225_targetTrain/wandb/debug.log
@@ -0,0 +1,21 @@
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_setup.py:_flush():81] Current SDK version is 0.25.0
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_setup.py:_flush():81] Configure stats pid to 121656
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug.log
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug-internal.log
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_init.py:init():844] calling init triggers
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
+config: {'model': {'encoder': {'name': 'dcsplat', 'input_image_shape': [518, 518], 'head_mode': 'pcd', 'num_level': 3, 'gs_param_dim': 256, 'align_corners': False, 'use_voxelize': True}, 'decoder': {'name': 'splatting_cuda', 'background_color': [0.0, 0.0, 0.0], 'make_scale_invariant': False}, 'density_control': {'name': 'density_control_module', 'mean_dim': 32, 'gs_param_dim': 256, 'refinement_layer_num': 1, 'num_level': 3, 'grad_mode': 'absgrad', 'use_mean_features': True, 'refinement_type': 'voxelize', 'refinement_hidden_dim': 32, 'aggregation_mode': 'mean', 'num_heads': 1, 'score_mode': 'absgrad', 'latent_dim': 128, 'num_latents': 64, 'num_self_attn_per_block': 2, 'voxel_size': 0.001, 'aux_refine': False, 'refine_error': False, 'use_refine_module': True, 'voxelize_activate': True, 'use_depth': False}}, 'render_loss': {'mse': {'weight': 1.0}, 'lpips': {'weight': 0.05, 'apply_after_step': 0}}, 'density_control_loss': {'error_score': {'weight': 0.01, 'log_scale': False, 'grad_scale': 10000.0, 'mode': 'original'}}, 'direct_loss': {'l1': {'weight': 0.8}, 'ssim': {'weight': 0.2}}, 'wandb': {'project': 'DCSplat', 'entity': 'scene-representation-group', 'name': 'ABLATION_0225_targetTrain', 'mode': 'online', 'tags': ['re10k', '256x256']}, 'mode': 'train', 'data_loader': {'train': {'num_workers': 16, 'persistent_workers': True, 'batch_size': 16, 'seed': 1234}, 'test': {'num_workers': 4, 'persistent_workers': False, 'batch_size': 1, 'seed': 2345}, 'val': {'num_workers': 1, 'persistent_workers': True, 'batch_size': 1, 'seed': 3456}}, 'optimizer': {'lr': 0.0002, 'warm_up_steps': 25, 'backbone_lr_multiplier': 0.1, 'backbone_trainable': 'T+H', 'accumulate': 1}, 'checkpointing': {'load': None, 'every_n_train_steps': 1500, 'save_top_k': 2, 'save_weights_only': False}, 'train': {'extended_visualization': False, 'print_log_every_n_steps': 10, 'camera_loss': 10.0, 'one_sample_validation': None, 'align_corners': False, 'intrinsic_scaling': False, 'verbose': False, 'beta_dist_param': [0.5, 4.0], 'use_refine_aux': False, 'train_target_set': True, 'train_gs_num': 1, 'ext_scale_detach': False, 'cam_scale_mode': 'sum', 'scene_scale_reg_loss': 0.0, 'train_aux': False, 'vggt_cam_loss': True, 'vggt_distil': False, 'context_view_train': False}, 'test': {'output_path': 'test/ablation/re10k', 'align_pose': False, 'pose_align_steps': 100, 'rot_opt_lr': 0.005, 'trans_opt_lr': 0.005, 'compute_scores': True, 'save_image': False, 'save_video': False, 'save_active_mask_image': False, 'save_error_score_image': False, 'save_compare': False, 'pred_intrinsic': False, 'error_threshold': 0.4, 'error_threshold_list': [0.2, 0.4, 0.6, 0.8, 1.0], 'threshold_mode': 'ratio', 'nvs_view_N_list': [3, 6, 16, 32, 64]}, 'seed': 111123, 'trainer': {'max_steps': 3001, 'val_check_interval': 250, 'gradient_clip_val': 0.5, 'num_nodes': 1}, 'dataset': {'re10k': {'make_baseline_1': True, 'relative_pose': True, 'augment': True, 'background_color': [0.0, 0.0, 0.0], 'overfit_to_scene': None, 'skip_bad_shape': True, 'view_sampler': {'name': 'bounded', 'num_target_views': 4, 'num_context_views': 2, 'min_distance_between_context_views': 45, 'max_distance_between_context_views': 90, 'min_distance_to_context_views': 0, 'warm_up_steps': 1000, 'initial_min_distance_between_context_views': 25, 'initial_max_distance_between_context_views': 25, 'same_target_gap': False, 'num_target_set': 3}, 'name': 're10k', 'roots': ['datasets/re10k'], 'input_image_shape': [256, 256], 'original_image_shape': [360, 640], 'cameras_are_circular': False, 'baseline_min': 0.001, 'baseline_max': 10000000000.0, 'max_fov': 100.0, 'dynamic_context_views': True, 'max_context_views_per_gpu': 24}}, '_wandb': {}}
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_init.py:init():892] starting backend
+2026-02-25 01:41:00,093 INFO    MainThread:121656 [wandb_init.py:init():895] sending inform_init request
+2026-02-25 01:41:00,100 INFO    MainThread:121656 [wandb_init.py:init():903] backend started and connected
+2026-02-25 01:41:00,102 INFO    MainThread:121656 [wandb_init.py:init():973] updated telemetry
+2026-02-25 01:41:00,106 INFO    MainThread:121656 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
+2026-02-25 01:41:02,034 INFO    MainThread:121656 [wandb_init.py:init():1042] starting run threads in backend
+2026-02-25 01:41:02,143 INFO    MainThread:121656 [wandb_run.py:_console_start():2524] atexit reg
+2026-02-25 01:41:02,143 INFO    MainThread:121656 [wandb_run.py:_redirect():2373] redirect: wrap_raw
+2026-02-25 01:41:02,143 INFO    MainThread:121656 [wandb_run.py:_redirect():2442] Wrapping output streams.
+2026-02-25 01:41:02,143 INFO    MainThread:121656 [wandb_run.py:_redirect():2465] Redirects installed.
+2026-02-25 01:41:02,146 INFO    MainThread:121656 [wandb_init.py:init():1082] run started, returning control to user process
+2026-02-25 04:35:28,261 INFO    wandb-AsyncioManager-main:121656 [service_client.py:_forward_responses():134] Reached EOF.
+2026-02-25 04:35:28,261 INFO    wandb-AsyncioManager-main:121656 [mailbox.py:close():155] Closing mailbox, abandoning 1 handles.
diff --git a/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/config.yaml b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..d3496f6f2cb7ab9f07eb43673f2e917e7fb12c42
--- /dev/null
+++ b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/config.yaml
@@ -0,0 +1,308 @@
+_wandb:
+    value:
+        cli_version: 0.25.0
+        e:
+            kk43v9bdhoc229qwa3b4aitwhvhz6zhc:
+                args:
+                    - +experiment=re10k_ablation_24v
+                    - wandb.mode=online
+                    - wandb.name=ABLATION_0225_targetTrain
+                    - train.scene_scale_reg_loss=0.0
+                    - train.train_aux=false
+                cpu_count: 128
+                cpu_count_logical: 256
+                cudaVersion: "13.1"
+                disk:
+                    /:
+                        total: "1170378588160"
+                        used: "660778004480"
+                email: dna9041@korea.ac.kr
+                executable: /venv/main/bin/python
+                git:
+                    commit: 2512754c6c27ca5150bf17fbcbdde3f192fd53cc
+                    remote: git@github.com:K-nowing/CVPR2026.git
+                gpu: NVIDIA H200
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-2649ab80-a3a6-5a1c-0fa5-12bc11bd75e9
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-e92921d9-c681-246f-af93-637e0dc938ca
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-ffe12ffc-9bb7-82de-5692-1ec0ee2e68d8
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-499e5acd-b6ab-2010-c51b-ee9b5aa65825
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-3b2522d9-1c72-e49b-2c30-96165680b74a
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-a9a280c5-b2f9-dc1e-a8a9-7326a74001ff
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-07d0167b-a6a1-1900-2d27-7c6c11598409
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "150754820096"
+                      name: NVIDIA H200
+                      uuid: GPU-8362a999-20d1-c27b-5d18-032d23f859ab
+                host: 27d18dedec6d
+                memory:
+                    total: "1622948257792"
+                os: Linux-6.8.0-90-generic-x86_64-with-glibc2.39
+                program: -m src.main
+                python: CPython 3.12.12
+                root: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain
+                startedAt: "2026-02-25T01:40:59.849537Z"
+                writerId: kk43v9bdhoc229qwa3b4aitwhvhz6zhc
+        m:
+            - "1": trainer/global_step
+              "6":
+                - 3
+              "7": []
+            - "2": '*'
+              "5": 1
+              "6":
+                - 1
+              "7": []
+        python_version: 3.12.12
+        t:
+            "1":
+                - 1
+                - 41
+                - 49
+                - 50
+                - 106
+            "2":
+                - 1
+                - 41
+                - 49
+                - 50
+                - 106
+            "3":
+                - 7
+                - 13
+                - 15
+                - 16
+                - 66
+            "4": 3.12.12
+            "5": 0.25.0
+            "12": 0.25.0
+            "13": linux-x86_64
+checkpointing:
+    value:
+        every_n_train_steps: 1500
+        load: null
+        save_top_k: 2
+        save_weights_only: false
+data_loader:
+    value:
+        test:
+            batch_size: 1
+            num_workers: 4
+            persistent_workers: false
+            seed: 2345
+        train:
+            batch_size: 16
+            num_workers: 16
+            persistent_workers: true
+            seed: 1234
+        val:
+            batch_size: 1
+            num_workers: 1
+            persistent_workers: true
+            seed: 3456
+dataset:
+    value:
+        re10k:
+            augment: true
+            background_color:
+                - 0
+                - 0
+                - 0
+            baseline_max: 1e+10
+            baseline_min: 0.001
+            cameras_are_circular: false
+            dynamic_context_views: true
+            input_image_shape:
+                - 256
+                - 256
+            make_baseline_1: true
+            max_context_views_per_gpu: 24
+            max_fov: 100
+            name: re10k
+            original_image_shape:
+                - 360
+                - 640
+            overfit_to_scene: null
+            relative_pose: true
+            roots:
+                - datasets/re10k
+            skip_bad_shape: true
+            view_sampler:
+                initial_max_distance_between_context_views: 25
+                initial_min_distance_between_context_views: 25
+                max_distance_between_context_views: 90
+                min_distance_between_context_views: 45
+                min_distance_to_context_views: 0
+                name: bounded
+                num_context_views: 2
+                num_target_set: 3
+                num_target_views: 4
+                same_target_gap: false
+                warm_up_steps: 1000
+density_control_loss:
+    value:
+        error_score:
+            grad_scale: 10000
+            log_scale: false
+            mode: original
+            weight: 0.01
+direct_loss:
+    value:
+        l1:
+            weight: 0.8
+        ssim:
+            weight: 0.2
+mode:
+    value: train
+model:
+    value:
+        decoder:
+            background_color:
+                - 0
+                - 0
+                - 0
+            make_scale_invariant: false
+            name: splatting_cuda
+        density_control:
+            aggregation_mode: mean
+            aux_refine: false
+            grad_mode: absgrad
+            gs_param_dim: 256
+            latent_dim: 128
+            mean_dim: 32
+            name: density_control_module
+            num_heads: 1
+            num_latents: 64
+            num_level: 3
+            num_self_attn_per_block: 2
+            refine_error: false
+            refinement_hidden_dim: 32
+            refinement_layer_num: 1
+            refinement_type: voxelize
+            score_mode: absgrad
+            use_depth: false
+            use_mean_features: true
+            use_refine_module: true
+            voxel_size: 0.001
+            voxelize_activate: true
+        encoder:
+            align_corners: false
+            gs_param_dim: 256
+            head_mode: pcd
+            input_image_shape:
+                - 518
+                - 518
+            name: dcsplat
+            num_level: 3
+            use_voxelize: true
+optimizer:
+    value:
+        accumulate: 1
+        backbone_lr_multiplier: 0.1
+        backbone_trainable: T+H
+        lr: 0.0002
+        warm_up_steps: 25
+render_loss:
+    value:
+        lpips:
+            apply_after_step: 0
+            weight: 0.05
+        mse:
+            weight: 1
+seed:
+    value: 111123
+test:
+    value:
+        align_pose: false
+        compute_scores: true
+        error_threshold: 0.4
+        error_threshold_list:
+            - 0.2
+            - 0.4
+            - 0.6
+            - 0.8
+            - 1
+        nvs_view_N_list:
+            - 3
+            - 6
+            - 16
+            - 32
+            - 64
+        output_path: test/ablation/re10k
+        pose_align_steps: 100
+        pred_intrinsic: false
+        rot_opt_lr: 0.005
+        save_active_mask_image: false
+        save_compare: false
+        save_error_score_image: false
+        save_image: false
+        save_video: false
+        threshold_mode: ratio
+        trans_opt_lr: 0.005
+train:
+    value:
+        align_corners: false
+        beta_dist_param:
+            - 0.5
+            - 4
+        cam_scale_mode: sum
+        camera_loss: 10
+        context_view_train: false
+        ext_scale_detach: false
+        extended_visualization: false
+        intrinsic_scaling: false
+        one_sample_validation: null
+        print_log_every_n_steps: 10
+        scene_scale_reg_loss: 0
+        train_aux: false
+        train_gs_num: 1
+        train_target_set: true
+        use_refine_aux: false
+        verbose: false
+        vggt_cam_loss: true
+        vggt_distil: false
+trainer:
+    value:
+        gradient_clip_val: 0.5
+        max_steps: 3001
+        num_nodes: 1
+        val_check_interval: 250
+wandb:
+    value:
+        entity: scene-representation-group
+        mode: online
+        name: ABLATION_0225_targetTrain
+        project: DCSplat
+        tags:
+            - re10k
+            - 256x256
diff --git a/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/output.log b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/output.log
new file mode 100644
index 0000000000000000000000000000000000000000..a8155df9e59c2385dfc85cfda6b46b423b74089d
--- /dev/null
+++ b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/output.log
@@ -0,0 +1,800 @@
+LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0,1,2,3,4,5,6,7]
+
+  | Name                   | Type                 | Params | Mode
+------------------------------------------------------------------------
+0 | encoder                | OurSplat             | 888 M  | train
+1 | density_control_module | DensityControlModule | 2.6 M  | train
+2 | decoder                | DecoderSplattingCUDA | 0      | train
+3 | render_losses          | ModuleList           | 0      | train
+4 | density_control_losses | ModuleList           | 0      | train
+5 | direct_losses          | ModuleList           | 0      | train
+------------------------------------------------------------------------
+891 M     Trainable params
+0         Non-trainable params
+891 M     Total params
+3,564.328 Total estimated model params size (MB)
+1231      Modules in train mode
+522       Modules in eval mode
+Sanity Checking: |                                                                                                                                                                      | 0/? [00:00<?, ?it/s][2026-02-25 01:41:04,283][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/data_connector.py:425: The 'val_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=31` in the `DataLoader` to improve performance.
+
+[2026-02-25 01:41:04,285][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user.
+  warnings.warn(  # warn only once
+
+Validation epoch start on rank 0
+Sanity Checking DataLoader 0:   0%|                                                                                                                                                     | 0/1 [00:00<?, ?it/s]validation step 0; scene = ['306e2b7785657539'];
+target intrinsic:  tensor(0.8595, device='cuda:0') tensor(0.8597, device='cuda:0')
+pred intrinsic:  tensor(0.8779, device='cuda:0') tensor(0.8773, device='cuda:0')
+[rank0]:W0225 01:41:06.565000 121656 site-packages/torch/utils/cpp_extension.py:2425] TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation.
+[rank0]:W0225 01:41:06.565000 121656 site-packages/torch/utils/cpp_extension.py:2425] If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'] to specific architectures.
+[2026-02-25 01:41:06,636][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 01:41:06,645][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/utilities/data.py:79: Trying to infer the `batch_size` from an ambiguous collection. The batch size we found is 1. To avoid any miscalculations, use `self.log(..., batch_size=batch_size)`.
+
+Setting up [LPIPS] perceptual loss: trunk [vgg], v[0.1], spatial [off]
+[2026-02-25 01:41:06,646][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 01:41:06,646][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+Loading model from: /venv/main/lib/python3.12/site-packages/lpips/weights/v0.1/vgg.pth
+[2026-02-25 01:41:08,321][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/functional.py:554: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /pytorch/aten/src/ATen/native/TensorShape.cpp:4322.)
+  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
+
+Sanity Checking DataLoader 0: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:03<00:00,  0.26it/s][2026-02-25 01:41:08,603][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/psnr', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 01:41:08,605][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/lpips', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 01:41:08,605][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/ssim', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 01:41:08,605][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/gaussian_num_ratio', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 01:41:08,605][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('info/global_step', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+Epoch 0: |                                                                                                                                                                              | 0/? [00:00<?, ?it/s]context = [[34, 36, 50, 53, 54, 60, 63, 70, 76, 78, 79, 80, 81, 88, 92, 94, 102, 110, 112, 114, 122, 125, 126, 131]]target = [[126, 96, 109, 55, 99, 116, 43, 60, 113, 85, 103, 90, 130, 62, 76, 123, 35, 102, 125, 128, 98, 67, 129, 79]]
+[2026-02-25 01:41:16,604][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [57, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [57, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 01:41:16,706][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                          | 9/? [00:35<00:00,  0.25it/s, v_num=seh9]train step 10; scene = [['36047ec1694f9d49'], ['eac6cedeba1f720b']]; loss = 0.103298
+Epoch 0: |                                                                                                                                                         | 10/? [00:39<00:00,  0.26it/s, v_num=seh9]context = [[241, 243, 246, 266], [36, 42, 60, 61], [91, 100, 108, 116], [19, 27, 28, 44], [9, 11, 19, 34], [19, 24, 42, 44]]target = [[244, 251, 250, 260], [52, 41, 43, 49], [113, 114, 115, 93], [30, 40, 20, 26], [20, 13, 24, 28], [37, 24, 22, 39]]
+Epoch 0: |                                                                                                                                                         | 19/? [01:11<00:00,  0.27it/s, v_num=seh9]train step 20; scene = [['ce3f0dbbab9e2619'], ['35f1002ddc4fbebf'], ['37b468e1f381e86f'], ['bcd63850409a4c42'], ['91ced81093fc5294'], ['7544731be485052b'], ['e4d41618c44cfb3a'], ['f6144e0803cf99db']]; loss = 0.095114
+Epoch 0: |                                                                                                                                                         | 20/? [01:14<00:00,  0.27it/s, v_num=seh9]context = [[33, 35, 49, 53, 55, 56, 62, 66], [55, 60, 70, 71, 80, 83, 87, 88], [1, 4, 8, 9, 10, 15, 24, 34]]target = [[49, 36, 48, 65, 57, 60, 50, 41], [60, 82, 78, 73, 67, 77, 87, 61], [3, 20, 13, 9, 14, 17, 12, 7]]
+Epoch 0: |                                                                                                                                                         | 24/? [01:27<00:00,  0.27it/s, v_num=seh9][2026-02-25 01:42:41,674][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+Epoch 0: |                                                                                                                                                         | 29/? [01:45<00:00,  0.28it/s, v_num=seh9]train step 30; scene = [['1e7c432d2207b6f2'], ['c495f01f294333ee'], ['0636f8d5854771ca'], ['8ef87b7c44d0d34f']]; loss = 0.042072
+Epoch 0: |                                                                                                                                                         | 30/? [01:49<00:00,  0.28it/s, v_num=seh9]context = [[10, 14, 16, 19, 20, 24, 33, 34, 42, 48, 51, 59], [54, 60, 62, 65, 71, 76, 77, 80, 91, 92, 98, 103]]target = [[14, 34, 26, 32, 11, 44, 29, 20, 15, 27, 52, 35], [97, 77, 72, 62, 61, 102, 67, 101, 89, 58, 81, 88]]
+Epoch 0: |                                                                                                                                                         | 39/? [02:20<00:00,  0.28it/s, v_num=seh9]train step 40; scene = [['3931c6a7e69ffff1']]; loss = 0.038852
+Epoch 0: |                                                                                                                                                         | 40/? [02:23<00:00,  0.28it/s, v_num=seh9]context = [[75, 90, 91, 94, 96, 100], [89, 97, 103, 109, 113, 114], [36, 39, 42, 57, 59, 62], [144, 147, 150, 151, 159, 170]]target = [[90, 79, 86, 92, 99, 85], [106, 90, 112, 96, 104, 110], [52, 44, 56, 49, 55, 41], [153, 149, 167, 166, 150, 147]]
+Epoch 0: |                                                                                                                                                         | 49/? [02:55<00:00,  0.28it/s, v_num=seh9]train step 50; scene = [['a559ed659c3d87e2']]; loss = 0.038939
+Epoch 0: |                                                                                                                                                         | 50/? [02:58<00:00,  0.28it/s, v_num=seh9]context = [[171, 177, 196], [2, 12, 29], [53, 67, 80], [160, 161, 185], [234, 237, 260], [5, 27, 32], [139, 140, 164], [4, 30, 31]]target = [[192, 188, 187], [4, 21, 24], [67, 79, 58], [161, 177, 163], [239, 258, 244], [13, 31, 25], [140, 163, 142], [14, 9, 20]]
+Epoch 0: |                                                                                                                                                         | 59/? [03:29<00:00,  0.28it/s, v_num=seh9]train step 60; scene = [['db139805c0d52235']]; loss = 0.024980
+Epoch 0: |                                                                                                                                                         | 60/? [03:33<00:00,  0.28it/s, v_num=seh9]context = [[21, 31, 32, 42, 43, 48, 49, 52, 56, 58, 59, 60, 62, 66, 72, 78, 88, 91, 93, 99, 100, 102, 105, 118]]target = [[106, 62, 43, 74, 92, 26, 38, 65, 85, 100, 68, 70, 81, 98, 115, 32, 107, 59, 110, 96, 77, 116, 114, 35]]
+Epoch 0: |                                                                                                                                                         | 69/? [04:03<00:00,  0.28it/s, v_num=seh9]train step 70; scene = [['cd6c21656a85e9b9'], ['3fe4e74e7531c300'], ['229581725830ce75'], ['47396d5a5299873e']]; loss = 0.033019
+Epoch 0: |                                                                                                                                                         | 70/? [04:07<00:00,  0.28it/s, v_num=seh9]context = [[3, 7, 11, 14, 15, 17, 18, 21, 33, 37, 45, 52], [68, 69, 76, 82, 86, 87, 93, 98, 100, 107, 108, 117]]target = [[27, 45, 17, 44, 9, 5, 34, 30, 43, 22, 11, 31], [84, 76, 90, 97, 87, 110, 105, 111, 95, 102, 113, 116]]
+Epoch 0: |                                                                                                                                                         | 79/? [04:37<00:00,  0.28it/s, v_num=seh9]train step 80; scene = [['1683c797f3feac29'], ['1916ec563ca4d9b6']]; loss = 0.033336
+Epoch 0: |                                                                                                                                                         | 80/? [04:41<00:00,  0.28it/s, v_num=seh9]context = [[9, 17, 27, 28, 35, 37, 40, 42], [36, 37, 43, 47, 52, 61, 63, 69], [67, 71, 74, 90, 95, 96, 99, 100]]target = [[23, 36, 14, 16, 34, 28, 17, 13], [63, 37, 50, 54, 39, 42, 57, 64], [75, 77, 84, 91, 98, 88, 83, 93]]
+Epoch 0: |                                                                                                                                                         | 89/? [05:11<00:00,  0.29it/s, v_num=seh9]train step 90; scene = [['ffae0c358d55ccd6']]; loss = 0.069921
+Epoch 0: |                                                                                                                                                         | 90/? [05:14<00:00,  0.29it/s, v_num=seh9]context = [[37, 39, 40, 43, 44, 45, 49, 51, 52, 54, 55, 59, 62, 74, 80, 83, 89, 92, 95, 96, 102, 118, 133, 134]]target = [[46, 52, 108, 96, 76, 118, 59, 73, 105, 109, 83, 125, 48, 39, 111, 98, 124, 79, 130, 132, 100, 104, 131, 97]]
+Epoch 0: |                                                                                                                                                         | 99/? [05:45<00:00,  0.29it/s, v_num=seh9]train step 100; scene = [['baab02d64ec03a56'], ['d56e9c764f637e1c'], ['a6990abb3591ce5a'], ['8ad7f684b4dddb11']]; loss = 0.036138
+Epoch 0: |                                                                                                                                                        | 100/? [05:48<00:00,  0.29it/s, v_num=seh9]context = [[41, 43, 50, 59, 66, 67, 69, 74], [5, 8, 12, 14, 24, 28, 31, 38], [1, 12, 13, 15, 25, 30, 32, 34]]target = [[66, 55, 53, 56, 44, 54, 59, 57], [30, 12, 29, 19, 18, 20, 17, 32], [11, 29, 21, 8, 16, 31, 22, 23]]
+Epoch 0: |                                                                                                                                                        | 109/? [06:19<00:00,  0.29it/s, v_num=seh9]train step 110; scene = [['124a8e1a8219f20f'], ['3a40230e7d49bd72'], ['47a1772c9348c0be'], ['893b447be20a73fc']]; loss = 0.027057
+Epoch 0: |                                                                                                                                                        | 110/? [06:23<00:00,  0.29it/s, v_num=seh9]context = [[77, 81, 84, 87, 88, 95, 101, 104, 109, 111, 122, 126], [52, 66, 67, 69, 74, 76, 82, 89, 92, 98, 100, 101]]target = [[92, 107, 90, 124, 120, 84, 117, 110, 105, 97, 106, 78], [87, 65, 84, 75, 92, 60, 74, 79, 98, 71, 76, 57]]
+Epoch 0: |                                                                                                                                                        | 119/? [06:54<00:00,  0.29it/s, v_num=seh9]train step 120; scene = [['2054cdda3bb0e2fa']]; loss = 0.022100
+Epoch 0: |                                                                                                                                                        | 120/? [06:58<00:00,  0.29it/s, v_num=seh9]context = [[136, 140, 142, 144, 152, 158, 159, 164, 165, 171, 172, 173, 179, 185, 187, 190, 193, 196, 197, 204, 206, 218, 232, 233]]target = [[190, 137, 170, 210, 220, 168, 155, 172, 146, 149, 202, 183, 166, 232, 203, 177, 225, 208, 139, 180, 199, 154, 216, 161]]
+Epoch 0: |                                                                                                                                                        | 129/? [07:28<00:00,  0.29it/s, v_num=seh9]train step 130; scene = [['40c0d605c4f8c69b'], ['a8b72199cf4cf5e2'], ['b75f3820760d835c']]; loss = 0.031350
+Epoch 0: |                                                                                                                                                        | 130/? [07:31<00:00,  0.29it/s, v_num=seh9]context = [[173, 179, 180, 182, 183, 184, 187, 188, 190, 193, 194, 195, 202, 204, 207, 213, 218, 224, 225, 249, 256, 260, 267, 270]]target = [[215, 236, 199, 243, 185, 249, 266, 259, 204, 189, 228, 188, 245, 203, 182, 198, 251, 195, 261, 238, 237, 257, 240, 212]]
+Epoch 0: |                                                                                                                                                        | 139/? [08:02<00:00,  0.29it/s, v_num=seh9]train step 140; scene = [['35fcc43842bf847e'], ['633e76d7ccbeb3ed'], ['0207b0ec0cc851f6'], ['63341a860ea3a43a'], ['b4e9a9bf77f35701'], ['482ceff1527f21e1']]; loss = 0.034945
+Epoch 0: |                                                                                                                                                        | 140/? [08:06<00:00,  0.29it/s, v_num=seh9]context = [[143, 164, 173], [112, 114, 142], [144, 152, 171], [50, 77, 78], [139, 148, 170], [75, 86, 107], [7, 18, 38], [99, 125, 132]]target = [[154, 148, 172], [129, 115, 128], [151, 168, 166], [65, 61, 57], [163, 141, 160], [82, 97, 106], [33, 11, 22], [112, 111, 103]]
+Epoch 0: |                                                                                                                                                        | 149/? [08:37<00:00,  0.29it/s, v_num=seh9]train step 150; scene = [['bb2afd35d6f8a765'], ['e1da66dcfb584564'], ['fc434865599d2fe6']]; loss = 0.019020
+Epoch 0: |                                                                                                                                                        | 150/? [08:40<00:00,  0.29it/s, v_num=seh9]context = [[76, 77, 96, 98, 112, 114, 116, 117, 121, 123, 124, 125], [3, 8, 17, 22, 23, 24, 26, 41, 42, 44, 51, 52]]target = [[121, 115, 81, 79, 105, 86, 114, 122, 109, 102, 101, 116], [9, 38, 27, 24, 46, 49, 18, 44, 31, 43, 15, 12]]
+Epoch 0: |                                                                                                                                                        | 159/? [09:11<00:00,  0.29it/s, v_num=seh9]train step 160; scene = [['e726107dc0960b84']]; loss = 0.027724
+Epoch 0: |                                                                                                                                                        | 160/? [09:14<00:00,  0.29it/s, v_num=seh9]context = [[91, 98, 100, 109, 113, 118, 119, 124, 128, 131, 138, 140], [59, 61, 71, 73, 75, 81, 86, 96, 97, 100, 106, 108]]target = [[137, 105, 134, 128, 136, 101, 125, 133, 114, 107, 98, 123], [107, 95, 90, 68, 72, 98, 62, 71, 67, 76, 69, 99]]
+Epoch 0: |                                                                                                                                                        | 169/? [09:46<00:00,  0.29it/s, v_num=seh9]train step 170; scene = [['6c383c3e7ece2df7']]; loss = 0.018364
+Epoch 0: |                                                                                                                                                        | 170/? [09:50<00:00,  0.29it/s, v_num=seh9]context = [[4, 5, 7, 12, 16, 19, 36, 39, 41, 47, 50, 51, 54, 58, 60, 63, 69, 71, 74, 75, 76, 86, 91, 101]]target = [[34, 93, 44, 16, 49, 25, 12, 28, 94, 48, 11, 58, 8, 76, 80, 57, 40, 13, 88, 39, 92, 41, 30, 61]]
+Epoch 0: |                                                                                                                                                        | 179/? [10:21<00:00,  0.29it/s, v_num=seh9]train step 180; scene = [['a52931a53e49657f'], ['5a2447dbfc1bdfac']]; loss = 0.018666
+Epoch 0: |                                                                                                                                                        | 180/? [10:25<00:00,  0.29it/s, v_num=seh9]context = [[86, 97, 99, 102, 109, 110, 118, 121], [152, 154, 164, 177, 181, 184, 186, 187], [29, 32, 35, 37, 42, 55, 57, 64]]target = [[100, 118, 98, 109, 94, 120, 116, 103], [162, 153, 181, 171, 159, 154, 183, 160], [33, 58, 50, 40, 35, 56, 34, 52]]
+Epoch 0: |                                                                                                                                                        | 189/? [10:56<00:00,  0.29it/s, v_num=seh9]train step 190; scene = [['fa5df7455c4df198'], ['d3136cedcdc0fc89']]; loss = 0.027796
+Epoch 0: |                                                                                                                                                        | 190/? [10:59<00:00,  0.29it/s, v_num=seh9]context = [[75, 84, 89, 102, 105, 108], [135, 136, 145, 150, 159, 164], [25, 27, 29, 56, 59, 61], [44, 51, 57, 70, 72, 73]]target = [[93, 89, 85, 105, 101, 96], [147, 137, 139, 158, 142, 159], [34, 40, 28, 38, 39, 49], [66, 49, 56, 47, 62, 70]]
+Epoch 0: |                                                                                                                                                        | 199/? [11:29<00:00,  0.29it/s, v_num=seh9]train step 200; scene = [['4d0d26a83b1768dd']]; loss = 0.032775
+Epoch 0: |                                                                                                                                                        | 200/? [11:32<00:00,  0.29it/s, v_num=seh9]context = [[13, 28, 48], [39, 63, 69], [58, 80, 92], [14, 17, 47], [1, 4, 36], [20, 44, 50], [31, 39, 65], [14, 33, 51]]target = [[18, 32, 38], [62, 60, 58], [77, 89, 82], [46, 27, 29], [4, 15, 18], [36, 23, 45], [42, 40, 45], [40, 19, 27]]
+[2026-02-25 01:52:46,552][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 209/? [12:04<00:00,  0.29it/s, v_num=seh9]train step 210; scene = [['3ffaf1e704427fc8'], ['87194474b33de348']]; loss = 0.020357
+Epoch 0: |                                                                                                                                                        | 210/? [12:07<00:00,  0.29it/s, v_num=seh9]context = [[78, 86, 87, 94, 96, 108, 109, 119, 122, 127, 134, 139, 143, 145, 149, 150, 151, 156, 157, 159, 161, 164, 166, 175]]target = [[132, 83, 156, 168, 127, 89, 174, 159, 123, 105, 171, 141, 98, 112, 113, 137, 150, 131, 124, 109, 97, 120, 152, 133]]
+Epoch 0: |                                                                                                                                                        | 219/? [12:37<00:00,  0.29it/s, v_num=seh9]train step 220; scene = [['87491783936d3687']]; loss = 0.030278
+Epoch 0: |                                                                                                                                                        | 220/? [12:40<00:00,  0.29it/s, v_num=seh9]context = [[2, 8, 9, 34], [5, 16, 31, 37], [21, 31, 50, 52], [107, 120, 134, 142], [77, 86, 93, 114], [34, 50, 66, 67]]target = [[13, 21, 6, 16], [14, 11, 8, 28], [27, 50, 39, 43], [116, 140, 127, 141], [96, 104, 80, 91], [36, 52, 39, 66]]
+Epoch 0: |                                                                                                                                                        | 229/? [13:11<00:00,  0.29it/s, v_num=seh9]train step 230; scene = [['34c8c62d878eca66'], ['4c691165a406de40'], ['87c9128b943f1b3f'], ['8b052fd18b3b7c20'], ['4d27fb96530fe02b'], ['ee19678152518002'], ['c49bd62c183dd925'], ['da388971863189eb']]; loss = 0.025407
+Epoch 0: |                                                                                                                                                        | 230/? [13:14<00:00,  0.29it/s, v_num=seh9]context = [[0, 1, 5, 6, 10, 13, 18, 24, 32, 34, 35, 37, 48, 58, 59, 68, 70, 73, 74, 81, 89, 90, 92, 97]]target = [[53, 56, 14, 5, 44, 96, 33, 65, 40, 24, 20, 17, 36, 19, 9, 21, 2, 7, 80, 28, 73, 27, 79, 87]]
+Epoch 0: |                                                                                                                                                        | 239/? [13:45<00:00,  0.29it/s, v_num=seh9]train step 240; scene = [['86dfadf971ff9ff5'], ['68e586537c71a833'], ['9ee2ca77349564bd'], ['9794641b7e015578'], ['d8b22b4eb5e28e71'], ['be0d02ca6abeb470']]; loss = 0.034101
+Epoch 0: |                                                                                                                                                        | 240/? [13:48<00:00,  0.29it/s, v_num=seh9]context = [[46, 52, 55, 56, 64, 67, 72, 73, 74, 77, 78, 80, 84, 90, 93, 94, 95, 101, 104, 109, 112, 113, 124, 143]]target = [[47, 57, 142, 76, 92, 60, 72, 66, 71, 115, 86, 105, 119, 70, 65, 100, 95, 101, 133, 111, 116, 62, 130, 126]]
+Epoch 0: |                                                                                                                                                        | 249/? [14:20<00:00,  0.29it/s, v_num=seh9]train step 250; scene = [['0ccfac8b4f535035']]; loss = 0.015534
+Epoch 0: |                                                                                                                                                        | 250/? [14:23<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 250; scene = ['49b8f80c849dc341'];
+target intrinsic:  tensor(0.8891, device='cuda:0') tensor(0.8894, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8873, device='cuda:0') tensor(0.8911, device='cuda:0')
+[2026-02-25 01:55:34,591][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 250/? [14:24<00:00,  0.29it/s, v_num=seh9]context = [[27, 30, 31, 33, 34, 39, 40, 44, 50, 51, 72, 76], [19, 25, 30, 32, 34, 37, 39, 43, 51, 58, 63, 68]]target = [[34, 31, 43, 37, 52, 59, 54, 48, 55, 65, 44, 71], [65, 45, 51, 58, 53, 67, 57, 22, 52, 26, 46, 61]]
+Epoch 0: |                                                                                                                                                        | 259/? [14:54<00:00,  0.29it/s, v_num=seh9]train step 260; scene = [['c0e3203eaee46f03'], ['47cd9a752d3fb63f'], ['3155b05b96e2111e']]; loss = 0.014766
+Epoch 0: |                                                                                                                                                        | 260/? [14:58<00:00,  0.29it/s, v_num=seh9]context = [[41, 43, 45, 49, 59, 60, 68, 73, 77, 83, 89, 90], [9, 12, 15, 16, 23, 25, 27, 34, 42, 51, 53, 58]]target = [[84, 73, 81, 76, 55, 42, 62, 80, 68, 83, 47, 49], [45, 34, 54, 32, 49, 37, 17, 12, 18, 11, 16, 27]]
+Epoch 0: |                                                                                                                                                        | 269/? [15:29<00:00,  0.29it/s, v_num=seh9]train step 270; scene = [['e7a85a4315af4cc1']]; loss = 0.019321
+Epoch 0: |                                                                                                                                                        | 270/? [15:33<00:00,  0.29it/s, v_num=seh9]context = [[5, 17, 20, 21, 27, 30, 31, 35, 36, 41, 47, 50, 53, 57, 71, 75, 77, 87, 89, 90, 95, 99, 100, 102]]target = [[92, 95, 40, 64, 67, 37, 31, 47, 42, 55, 51, 33, 53, 59, 68, 9, 50, 93, 46, 71, 14, 44, 45, 28]]
+Epoch 0: |                                                                                                                                                        | 279/? [16:04<00:00,  0.29it/s, v_num=seh9]train step 280; scene = [['1b778f72bbee1f27']]; loss = 0.016219
+Epoch 0: |                                                                                                                                                        | 280/? [16:08<00:00,  0.29it/s, v_num=seh9]context = [[70, 72, 76, 90, 100, 112], [34, 40, 47, 48, 60, 64], [80, 89, 97, 105, 112, 115], [0, 2, 13, 24, 28, 36]]target = [[96, 110, 76, 95, 71, 75], [58, 40, 55, 51, 49, 35], [104, 96, 81, 110, 85, 97], [11, 35, 26, 21, 10, 28]]
+Epoch 0: |                                                                                                                                                        | 289/? [16:37<00:00,  0.29it/s, v_num=seh9]train step 290; scene = [['e7a04da784bddf9b'], ['5e9846591aa4776c'], ['b1d7d8d0b587ffc4']]; loss = 0.019634
+Epoch 0: |                                                                                                                                                        | 290/? [16:41<00:00,  0.29it/s, v_num=seh9]context = [[34, 37, 60, 61, 63, 73], [18, 22, 35, 37, 39, 52], [22, 23, 25, 35, 37, 61], [13, 22, 33, 38, 42, 44]]target = [[53, 55, 41, 47, 42, 66], [38, 43, 51, 41, 40, 39], [53, 33, 46, 57, 44, 36], [27, 24, 35, 31, 23, 26]]
+Epoch 0: |                                                                                                                                                        | 299/? [17:13<00:00,  0.29it/s, v_num=seh9]train step 300; scene = [['58ae98561b032336'], ['a942015dc7100bd5'], ['cc68efcc2b123690'], ['120391d54ad880a1']]; loss = 0.020225
+Epoch 0: |                                                                                                                                                        | 300/? [17:16<00:00,  0.29it/s, v_num=seh9]context = [[43, 55, 58, 61, 63, 65, 68, 70, 79, 83, 91, 92], [96, 98, 101, 105, 108, 117, 118, 123, 132, 141, 143, 145]]target = [[82, 45, 49, 79, 50, 62, 58, 61, 59, 72, 51, 44], [131, 124, 140, 108, 137, 123, 120, 122, 141, 102, 134, 143]]
+Epoch 0: |                                                                                                                                                        | 309/? [17:47<00:00,  0.29it/s, v_num=seh9]train step 310; scene = [['7e58b6857e275547']]; loss = 0.026335
+Epoch 0: |                                                                                                                                                        | 310/? [17:51<00:00,  0.29it/s, v_num=seh9]context = [[7, 13, 26, 32, 42, 49], [106, 110, 121, 124, 145, 146], [88, 99, 102, 103, 104, 127], [73, 85, 97, 99, 102, 103]]target = [[45, 46, 15, 38, 17, 41], [122, 137, 124, 126, 132, 142], [102, 117, 123, 119, 109, 125], [94, 96, 102, 77, 84, 79]]
+Epoch 0: |                                                                                                                                                        | 319/? [18:21<00:00,  0.29it/s, v_num=seh9]train step 320; scene = [['82376678602466d7'], ['3d4b9b728d676007'], ['62dbaf64a3a79445'], ['ec36aa235dc5a597']]; loss = 0.013789
+Epoch 0: |                                                                                                                                                        | 320/? [18:25<00:00,  0.29it/s, v_num=seh9]context = [[1, 3, 10, 11, 31, 37], [50, 51, 52, 71, 80, 93], [60, 64, 82, 87, 101, 103], [40, 47, 51, 67, 68, 71]]target = [[8, 20, 11, 3, 5, 14], [54, 63, 66, 91, 79, 76], [77, 96, 81, 94, 75, 85], [66, 54, 60, 64, 43, 53]]
+Epoch 0: |                                                                                                                                                        | 329/? [18:56<00:00,  0.29it/s, v_num=seh9]train step 330; scene = [['05c6dc05aaf40f45']]; loss = 0.020309
+Epoch 0: |                                                                                                                                                        | 330/? [19:00<00:00,  0.29it/s, v_num=seh9]context = [[58, 59, 61, 68, 72, 84, 89, 90, 95, 100, 103, 107], [35, 37, 41, 49, 52, 56, 58, 59, 66, 69, 80, 84]]target = [[89, 75, 61, 103, 60, 79, 62, 93, 99, 106, 81, 78], [53, 79, 64, 76, 38, 45, 49, 83, 56, 61, 69, 43]]
+Epoch 0: |                                                                                                                                                        | 339/? [19:32<00:00,  0.29it/s, v_num=seh9]train step 340; scene = [['6fceeb6dbfb1d42d'], ['89a30f540e35e364'], ['aa2c714ead9d4071']]; loss = 0.016609
+Epoch 0: |                                                                                                                                                        | 340/? [19:35<00:00,  0.29it/s, v_num=seh9]context = [[15, 54], [43, 84], [0, 31], [7, 52], [7, 51], [21, 57], [45, 89], [17, 51], [33, 69], [61, 100], [2, 35], [27, 67]]target = [[34, 28], [44, 83], [14, 7], [26, 15], [17, 21], [37, 23], [49, 71], [30, 36], [58, 41], [77, 73], [24, 23], [34, 46]]
+Epoch 0: |                                                                                                                                                        | 349/? [20:07<00:00,  0.29it/s, v_num=seh9]train step 350; scene = [['d17d6e951b1fb862'], ['dad4c7a3ecb07b58'], ['e2f3157c10aa655b'], ['7af404ed5bc4ae26']]; loss = 0.013421
+Epoch 0: |                                                                                                                                                        | 350/? [20:10<00:00,  0.29it/s, v_num=seh9]context = [[67, 78, 104, 105], [24, 28, 36, 59], [77, 81, 104, 109], [67, 71, 103, 110], [64, 82, 90, 95], [10, 33, 46, 52]]target = [[75, 83, 70, 102], [52, 37, 25, 45], [106, 100, 80, 95], [91, 106, 72, 81], [82, 70, 75, 74], [26, 30, 28, 38]]
+Epoch 0: |                                                                                                                                                        | 359/? [20:42<00:00,  0.29it/s, v_num=seh9]train step 360; scene = [['c7d35bb824ce8724'], ['fa06053c189cf9bd']]; loss = 0.023816
+Epoch 0: |                                                                                                                                                        | 360/? [20:45<00:00,  0.29it/s, v_num=seh9]context = [[8, 12, 13, 16, 19, 24, 31, 38, 46, 49, 52, 57], [34, 36, 42, 48, 50, 54, 60, 61, 62, 67, 81, 83]]target = [[46, 45, 12, 37, 35, 17, 40, 30, 49, 53, 28, 11], [68, 36, 78, 56, 82, 55, 51, 42, 54, 40, 60, 46]]
+Epoch 0: |                                                                                                                                                        | 369/? [21:16<00:00,  0.29it/s, v_num=seh9]train step 370; scene = [['a162b27e42a276da'], ['4b1ace7056c3ef7c'], ['f46fdb94ef36d8db'], ['6545c42d1fa5df33'], ['d85bf23db2a97502'], ['cc68c11bf97c66c3'], ['560c5720aa4920fd'], ['599f80ec8db49f3f']]; loss = 0.025807
+Epoch 0: |                                                                                                                                                        | 370/? [21:20<00:00,  0.29it/s, v_num=seh9]context = [[3, 4, 10, 13, 27, 28, 32, 40], [3, 6, 11, 19, 23, 32, 35, 44], [36, 42, 43, 45, 46, 54, 59, 69]]target = [[20, 18, 33, 29, 23, 35, 15, 13], [39, 15, 34, 30, 16, 20, 4, 11], [48, 46, 42, 39, 49, 57, 51, 67]]
+Epoch 0: |                                                                                                                                                        | 379/? [21:51<00:00,  0.29it/s, v_num=seh9]train step 380; scene = [['89e29a4ab957f358'], ['8dcd060a4bdf34b9'], ['23c1f863919d9a5a']]; loss = 0.015937
+Epoch 0: |                                                                                                                                                        | 380/? [21:55<00:00,  0.29it/s, v_num=seh9]context = [[141, 142, 146, 151, 159, 162, 163, 176, 177, 185, 189, 190], [23, 26, 30, 31, 33, 34, 43, 60, 65, 69, 70, 72]]target = [[182, 158, 178, 155, 176, 156, 186, 165, 188, 159, 181, 157], [40, 29, 39, 52, 51, 65, 28, 68, 37, 70, 53, 34]]
+Epoch 0: |                                                                                                                                                        | 389/? [22:26<00:00,  0.29it/s, v_num=seh9]train step 390; scene = [['298c3bccb02eb533'], ['452625cd6b071b87'], ['14900b71ac66b7bd'], ['9b8b7ae8b1327717'], ['012ae7e6ae8eb7b1'], ['ac8497e8d2e2d395']]; loss = 0.020251
+Epoch 0: |                                                                                                                                                        | 390/? [22:28<00:00,  0.29it/s, v_num=seh9]context = [[176, 183, 192, 194, 199, 214, 215, 216, 218, 220, 224, 225], [86, 88, 90, 92, 97, 103, 109, 110, 112, 118, 126, 135]]target = [[218, 220, 197, 217, 180, 216, 214, 219, 186, 210, 212, 213], [103, 101, 94, 108, 96, 105, 118, 88, 89, 123, 127, 109]]
+Epoch 0: |                                                                                                                                                        | 399/? [23:00<00:00,  0.29it/s, v_num=seh9]train step 400; scene = [['b290b6a0afa1dac7'], ['b25963fb28a6fd6a'], ['c0f67af5cd34e8d8'], ['a7c776da58a96494'], ['3b22bca6f62b1f3e'], ['99947f86c2cc4108']]; loss = 0.020816
+Epoch 0: |                                                                                                                                                        | 400/? [23:04<00:00,  0.29it/s, v_num=seh9]context = [[5, 7, 9, 19, 26, 53], [9, 15, 18, 37, 39, 47], [41, 56, 59, 63, 64, 76], [26, 29, 49, 57, 62, 65]]target = [[43, 52, 51, 29, 40, 39], [42, 31, 45, 32, 14, 37], [45, 51, 67, 60, 65, 70], [42, 41, 63, 35, 43, 54]]
+[2026-02-25 02:04:17,312][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 409/? [23:34<00:00,  0.29it/s, v_num=seh9]train step 410; scene = [['8d8bd167d9a3ab72'], ['8c40e162a6c8015d'], ['6803b3d46e5a2c7f'], ['3eb44f0a98eaa502']]; loss = 0.019031
+Epoch 0: |                                                                                                                                                        | 410/? [23:37<00:00,  0.29it/s, v_num=seh9]context = [[59, 60, 62, 63, 68, 70, 71, 72, 81, 82, 89, 98, 106, 107, 112, 120, 121, 126, 129, 130, 142, 146, 152, 156]]target = [[70, 97, 139, 124, 103, 150, 132, 63, 127, 62, 142, 136, 90, 107, 152, 131, 108, 69, 86, 91, 128, 148, 147, 117]]
+Epoch 0: |                                                                                                                                                        | 419/? [24:07<00:00,  0.29it/s, v_num=seh9]train step 420; scene = [['f10456015e79b2b7']]; loss = 0.011766
+Epoch 0: |                                                                                                                                                        | 420/? [24:11<00:00,  0.29it/s, v_num=seh9]context = [[66, 69, 71, 73, 78, 79, 83, 86, 90, 97, 103, 110, 112, 117, 118, 120, 139, 140, 143, 144, 151, 152, 154, 163]]target = [[73, 132, 67, 122, 110, 151, 120, 89, 148, 84, 150, 126, 88, 116, 157, 138, 134, 145, 76, 105, 74, 119, 158, 121]]
+Epoch 0: |                                                                                                                                                        | 429/? [24:41<00:00,  0.29it/s, v_num=seh9]train step 430; scene = [['423716b02f72a648']]; loss = 0.012475
+Epoch 0: |                                                                                                                                                        | 430/? [24:44<00:00,  0.29it/s, v_num=seh9]context = [[110, 111, 122, 124, 125, 127, 138, 140, 142, 145, 150, 151, 152, 155, 156, 160, 170, 181, 182, 189, 194, 200, 203, 207]]target = [[165, 170, 159, 162, 180, 164, 143, 138, 203, 161, 129, 160, 168, 118, 128, 139, 111, 202, 148, 200, 204, 141, 188, 132]]
+Epoch 0: |                                                                                                                                                        | 439/? [25:15<00:00,  0.29it/s, v_num=seh9]train step 440; scene = [['a56f0d62418193a5'], ['297b64e30f27b9d2']]; loss = 0.018105
+Epoch 0: |                                                                                                                                                        | 440/? [25:19<00:00,  0.29it/s, v_num=seh9]context = [[92, 99, 106, 109, 115, 117, 120, 127, 130, 131, 139, 140, 141, 144, 148, 153, 154, 155, 169, 175, 177, 178, 179, 189]]target = [[140, 93, 165, 158, 179, 104, 95, 150, 125, 177, 122, 143, 108, 134, 139, 101, 148, 185, 117, 119, 169, 118, 159, 161]]
+Epoch 0: |                                                                                                                                                        | 449/? [25:50<00:00,  0.29it/s, v_num=seh9]train step 450; scene = [['ac2fc17a4de90b7b'], ['d8a3b176b0529293'], ['60a93ae5bb06a238'], ['53dc437d11e76ef6']]; loss = 0.017921
+Epoch 0: |                                                                                                                                                        | 450/? [25:54<00:00,  0.29it/s, v_num=seh9]context = [[3, 19, 25, 26, 28, 31, 32, 33, 37, 44, 48, 56], [225, 232, 239, 241, 255, 257, 262, 263, 264, 269, 271, 275]]target = [[20, 7, 24, 36, 23, 21, 29, 39, 50, 54, 6, 42], [268, 254, 231, 245, 248, 241, 256, 267, 265, 258, 250, 227]]
+Epoch 0: |                                                                                                                                                        | 459/? [26:24<00:00,  0.29it/s, v_num=seh9]train step 460; scene = [['02f8df567f0c3db6']]; loss = 0.097493
+Epoch 0: |                                                                                                                                                        | 460/? [26:27<00:00,  0.29it/s, v_num=seh9]context = [[18, 22, 25, 28, 35, 36, 42, 43, 48, 58, 59, 71], [65, 70, 78, 83, 85, 89, 90, 91, 93, 97, 99, 114]]target = [[37, 36, 30, 24, 32, 69, 19, 25, 52, 62, 26, 51], [86, 112, 101, 106, 71, 98, 77, 70, 87, 111, 99, 74]]
+Epoch 0: |                                                                                                                                                        | 469/? [26:57<00:00,  0.29it/s, v_num=seh9]train step 470; scene = [['a52d26a78b04aebd']]; loss = 0.012453
+Epoch 0: |                                                                                                                                                        | 470/? [27:01<00:00,  0.29it/s, v_num=seh9]context = [[27, 34, 38, 43, 48, 49, 56, 61, 74, 78, 79, 81], [0, 2, 11, 13, 16, 21, 29, 33, 40, 42, 48, 50]]target = [[58, 28, 41, 60, 49, 65, 44, 52, 36, 55, 61, 78], [20, 26, 29, 31, 36, 5, 47, 7, 3, 42, 22, 17]]
+Epoch 0: |                                                                                                                                                        | 479/? [27:32<00:00,  0.29it/s, v_num=seh9]train step 480; scene = [['3d9f44d7bdd0796d']]; loss = 0.013421
+Epoch 0: |                                                                                                                                                        | 480/? [27:35<00:00,  0.29it/s, v_num=seh9]context = [[60, 62, 75, 76, 85, 89, 104, 107], [4, 18, 26, 30, 35, 36, 39, 45], [80, 88, 100, 102, 109, 112, 115, 127]]target = [[97, 77, 98, 94, 96, 91, 105, 95], [39, 19, 29, 28, 26, 43, 41, 36], [98, 109, 94, 88, 100, 101, 87, 95]]
+Epoch 0: |                                                                                                                                                        | 489/? [28:06<00:00,  0.29it/s, v_num=seh9]train step 490; scene = [['8dd25b0a12c6a8f8'], ['6882e4590bb8ff85'], ['24f3efef10906531'], ['3bff367484a62a13'], ['62216d162b71b5b4'], ['6cb7f85c92dc6f1a'], ['dfb8f1b208a8949f'], ['0d7c1a3319b74e43']]; loss = 0.018471
+Epoch 0: |                                                                                                                                                        | 490/? [28:10<00:00,  0.29it/s, v_num=seh9]context = [[0, 1, 31, 35, 48, 55], [18, 20, 43, 47, 55, 60], [35, 58, 61, 73, 77, 80], [5, 15, 26, 29, 59, 60]]target = [[47, 35, 1, 37, 36, 28], [38, 42, 41, 36, 31, 52], [79, 42, 49, 71, 54, 72], [22, 49, 8, 41, 47, 18]]
+Epoch 0: |                                                                                                                                                        | 499/? [28:41<00:00,  0.29it/s, v_num=seh9]train step 500; scene = [['63e22bbf20853cd9'], ['bcef3076b93012b1'], ['65c3f29c43dd1e63'], ['e8d6100917c31f4c'], ['4e4314a00227ed83'], ['964a8c79fc153038'], ['3062ac95a34b9b4f'], ['3d7dfcfce85588a6']]; loss = 0.028978
+Epoch 0: |                                                                                                                                                        | 500/? [28:44<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 500; scene = ['73d6f935f31b3fd4'];
+target intrinsic:  tensor(0.8576, device='cuda:0') tensor(0.8579, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8877, device='cuda:0') tensor(0.8939, device='cuda:0')
+[2026-02-25 02:09:55,959][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 500/? [28:46<00:00,  0.29it/s, v_num=seh9]context = [[1, 3, 6, 8, 12, 32, 33, 41], [1, 19, 24, 25, 27, 29, 33, 44], [77, 79, 83, 92, 94, 103, 112, 119]]target = [[31, 8, 16, 22, 14, 27, 7, 32], [9, 33, 22, 23, 7, 19, 32, 17], [80, 101, 103, 106, 96, 97, 86, 114]]
+Epoch 0: |                                                                                                                                                        | 509/? [29:15<00:00,  0.29it/s, v_num=seh9]train step 510; scene = [['2b1f47da224557a3'], ['407f9d609bda4fb1'], ['23b99c0bdf36a8ab']]; loss = 0.017944
+Epoch 0: |                                                                                                                                                        | 510/? [29:19<00:00,  0.29it/s, v_num=seh9]context = [[184, 205, 217, 221, 222, 228], [5, 16, 19, 28, 42, 47], [0, 1, 2, 9, 11, 36], [8, 28, 31, 41, 51, 58]]target = [[214, 190, 193, 202, 199, 196], [14, 41, 8, 40, 29, 19], [28, 16, 32, 3, 33, 13], [44, 35, 34, 23, 51, 17]]
+Epoch 0: |                                                                                                                                                        | 519/? [29:50<00:00,  0.29it/s, v_num=seh9]train step 520; scene = [['9741eb65c3d6d1b0'], ['61049a4bf21ec488'], ['176ba1bba0744fe1']]; loss = 0.017753
+Epoch 0: |                                                                                                                                                        | 520/? [29:53<00:00,  0.29it/s, v_num=seh9]context = [[79, 80, 96, 113, 116, 118, 123, 124, 125, 130, 133, 134], [9, 17, 18, 23, 29, 35, 36, 39, 47, 56, 60, 65]]target = [[126, 101, 85, 92, 98, 115, 120, 119, 130, 117, 91, 105], [50, 35, 20, 21, 58, 46, 40, 11, 60, 62, 37, 15]]
+Epoch 0: |                                                                                                                                                        | 529/? [30:23<00:00,  0.29it/s, v_num=seh9]train step 530; scene = [['5a040cbc8d351ff7']]; loss = 0.027365
+Epoch 0: |                                                                                                                                                        | 530/? [30:27<00:00,  0.29it/s, v_num=seh9]context = [[20, 54, 60], [22, 26, 77], [75, 102, 125], [45, 73, 97], [197, 236, 247], [12, 26, 47], [62, 112, 113], [145, 174, 180]]target = [[37, 33, 30], [66, 56, 24], [91, 103, 88], [47, 80, 62], [209, 222, 239], [18, 45, 38], [105, 92, 110], [156, 178, 160]]
+Epoch 0: |                                                                                                                                                        | 539/? [30:58<00:00,  0.29it/s, v_num=seh9]train step 540; scene = [['5a43331e136e1666'], ['e20fa4c9c8fc8f42'], ['9d8ddcdbe1f7ac42'], ['ab2680bf91942e23']]; loss = 0.014743
+Epoch 0: |                                                                                                                                                        | 540/? [31:02<00:00,  0.29it/s, v_num=seh9]context = [[18, 21, 28, 29, 30, 31, 33, 39, 40, 43, 61, 68, 70, 72, 74, 77, 80, 83, 91, 104, 105, 108, 113, 115]]target = [[38, 62, 51, 34, 61, 58, 30, 89, 113, 56, 82, 112, 64, 60, 73, 40, 71, 41, 109, 96, 100, 97, 47, 24]]
+Epoch 0: |                                                                                                                                                        | 549/? [31:33<00:00,  0.29it/s, v_num=seh9]train step 550; scene = [['46eccfae1e09e577']]; loss = 0.013424
+Epoch 0: |                                                                                                                                                        | 550/? [31:36<00:00,  0.29it/s, v_num=seh9]context = [[23, 24, 51, 59], [63, 79, 117, 121], [31, 40, 43, 72], [210, 211, 229, 262], [108, 119, 135, 160], [33, 53, 74, 86]]target = [[34, 55, 37, 39], [92, 76, 107, 77], [36, 61, 57, 70], [233, 244, 230, 253], [150, 120, 147, 110], [85, 46, 80, 54]]
+Epoch 0: |                                                                                                                                                        | 559/? [32:08<00:00,  0.29it/s, v_num=seh9]train step 560; scene = [['afc970648a89e04a'], ['100270e080fe5b87'], ['90d87190e41314e7'], ['267277440899ef99']]; loss = 0.016702
+Epoch 0: |                                                                                                                                                        | 560/? [32:11<00:00,  0.29it/s, v_num=seh9]context = [[125, 126, 130, 136, 139, 142, 144, 149, 151, 154, 171, 176, 184, 188, 192, 193, 197, 200, 206, 207, 210, 217, 220, 222]]target = [[195, 179, 184, 167, 187, 208, 126, 166, 127, 146, 180, 134, 207, 206, 145, 181, 143, 165, 209, 218, 156, 171, 141, 190]]
+Epoch 0: |                                                                                                                                                        | 569/? [32:43<00:00,  0.29it/s, v_num=seh9]train step 570; scene = [['2bf30153b26e2060'], ['9c41c4a921687df3'], ['7f1f3578729394b9'], ['8eb8e4b1dfd6afbd']]; loss = 0.014690
+Epoch 0: |                                                                                                                                                        | 570/? [32:47<00:00,  0.29it/s, v_num=seh9]context = [[54, 56, 58, 59, 63, 74, 81, 85, 87, 90, 93, 101, 102, 103, 117, 118, 119, 127, 130, 133, 136, 140, 150, 151]]target = [[85, 91, 142, 119, 147, 80, 121, 104, 65, 105, 113, 134, 79, 90, 149, 71, 88, 102, 60, 73, 123, 64, 130, 66]]
+Epoch 0: |                                                                                                                                                        | 579/? [33:18<00:00,  0.29it/s, v_num=seh9]train step 580; scene = [['e9e24b6ec4ba54e2'], ['7ac2f82f912c0ed2'], ['d4775a73d902fab3']]; loss = 0.020872
+Epoch 0: |                                                                                                                                                        | 580/? [33:22<00:00,  0.29it/s, v_num=seh9]context = [[102, 106, 110, 111, 115, 116, 122, 131, 135, 137, 138, 146, 151, 158, 163, 165, 167, 170, 174, 176, 183, 186, 191, 199]]target = [[197, 181, 194, 189, 140, 142, 125, 162, 135, 127, 121, 129, 111, 168, 176, 179, 183, 184, 177, 108, 193, 148, 107, 134]]
+Epoch 0: |                                                                                                                                                        | 589/? [33:53<00:00,  0.29it/s, v_num=seh9]train step 590; scene = [['82b22a784b4af245'], ['f0a89e470598390d'], ['7ea719e779680555'], ['59f9c9b55158d47d']]; loss = 0.016695
+Epoch 0: |                                                                                                                                                        | 590/? [33:57<00:00,  0.29it/s, v_num=seh9]context = [[4, 14, 15, 18, 28, 32, 36, 39, 41, 42, 45, 46, 51, 59, 66, 71, 73, 75, 77, 85, 86, 90, 97, 101]]target = [[53, 55, 72, 86, 33, 97, 31, 88, 18, 25, 75, 99, 22, 81, 59, 7, 38, 90, 85, 54, 60, 32, 57, 92]]
+Epoch 0: |                                                                                                                                                        | 599/? [34:27<00:00,  0.29it/s, v_num=seh9]train step 600; scene = [['e734e244799d0bfd'], ['c62f05d391aa6ac6'], ['bc500f01a9368065'], ['8f8405a3407168ff'], ['30124191dafb3383'], ['43947a751e7c8059']]; loss = 0.027815
+Epoch 0: |                                                                                                                                                        | 600/? [34:30<00:00,  0.29it/s, v_num=seh9]context = [[33, 69], [3, 42], [37, 76], [17, 66], [4, 44], [22, 78], [41, 79], [36, 93], [97, 133], [50, 89], [19, 80], [7, 58]]target = [[63, 36], [19, 20], [41, 52], [57, 37], [29, 22], [58, 74], [67, 65], [72, 56], [122, 117], [84, 88], [22, 69], [22, 56]]
+[2026-02-25 02:15:44,850][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 609/? [35:01<00:00,  0.29it/s, v_num=seh9]train step 610; scene = [['3bca29d56784093e']]; loss = 0.017258
+Epoch 0: |                                                                                                                                                        | 610/? [35:04<00:00,  0.29it/s, v_num=seh9]context = [[29, 34, 35, 37, 42, 44, 53, 54, 63, 72, 73, 80, 82, 84, 95, 96, 101, 102, 103, 106, 109, 112, 114, 126]]target = [[65, 35, 40, 57, 117, 30, 102, 46, 41, 112, 58, 39, 52, 114, 87, 89, 77, 31, 72, 91, 53, 59, 74, 120]]
+Epoch 0: |                                                                                                                                                        | 619/? [35:36<00:00,  0.29it/s, v_num=seh9]train step 620; scene = [['873339b3cca41d9d'], ['cbb3ec1745f4d00a'], ['a8f7946996aa6d67'], ['6093fe2362714108'], ['bcec4cead9a8da81'], ['ef685fbdf3e576a7'], ['6515de6c81e7c80c'], ['547583d5f8ae1238'], ['adde543e98ec0342'], ['bc5ebbad305b5d0f'], ['5dd26c5a45c5e40c'], ['9fde0d7ec4b06f6c']]; loss = 0.030805
+Epoch 0: |                                                                                                                                                        | 620/? [35:40<00:00,  0.29it/s, v_num=seh9]context = [[165, 181, 184, 186, 191, 195, 199, 204, 205, 212, 213, 214, 216, 217, 222, 223, 232, 237, 238, 246, 248, 256, 259, 262]]target = [[219, 247, 207, 235, 190, 229, 176, 186, 218, 237, 166, 256, 252, 233, 178, 192, 234, 226, 182, 191, 203, 212, 198, 217]]
+Epoch 0: |                                                                                                                                                        | 629/? [36:11<00:00,  0.29it/s, v_num=seh9]train step 630; scene = [['b29d69f0e6f21178']]; loss = 0.021198
+Epoch 0: |                                                                                                                                                        | 630/? [36:15<00:00,  0.29it/s, v_num=seh9]context = [[129, 130, 131, 132, 137, 147, 155, 160, 162, 181, 192, 193, 194, 195, 198, 200, 203, 210, 211, 215, 216, 219, 222, 226]]target = [[172, 169, 157, 221, 143, 186, 151, 205, 159, 219, 178, 146, 191, 194, 145, 162, 190, 222, 209, 193, 216, 225, 170, 217]]
+Epoch 0: |                                                                                                                                                        | 639/? [36:46<00:00,  0.29it/s, v_num=seh9]train step 640; scene = [['2726f059d69c25f4'], ['f95349aeab4d3dc0'], ['da963e97170fd9e9'], ['43c4c89b3414a0fc'], ['38e8301a4405eca6'], ['513aa200a2b07a1e'], ['992f739dba95f93a'], ['f8532bb5eca0e29d'], ['6e47f42627792c8a'], ['27c7362a27f418ef'], ['5874223a67b1e443'], ['00a9f110ad222aa4']]; loss = 0.052910
+Epoch 0: |                                                                                                                                                        | 640/? [36:49<00:00,  0.29it/s, v_num=seh9]context = [[11, 12, 14, 19, 20, 21, 26, 34, 35, 40, 41, 43, 45, 56, 59, 66, 72, 74, 76, 81, 85, 90, 96, 108]]target = [[61, 58, 36, 99, 89, 93, 56, 34, 38, 88, 39, 81, 32, 51, 84, 62, 48, 69, 64, 82, 29, 53, 15, 72]]
+Epoch 0: |                                                                                                                                                        | 649/? [37:21<00:00,  0.29it/s, v_num=seh9]train step 650; scene = [['9254db8c00084848'], ['467e6cb4bcb7fc53']]; loss = 0.012349
+Epoch 0: |                                                                                                                                                        | 650/? [37:25<00:00,  0.29it/s, v_num=seh9]context = [[90, 96, 97, 100, 112, 114, 117, 122, 131, 136, 140, 144, 146, 150, 161, 164, 173, 176, 177, 178, 179, 180, 186, 187]]target = [[164, 120, 180, 177, 118, 142, 122, 154, 105, 172, 128, 168, 116, 147, 119, 160, 166, 101, 169, 152, 121, 157, 130, 125]]
+Epoch 0: |                                                                                                                                                        | 659/? [37:56<00:00,  0.29it/s, v_num=seh9]train step 660; scene = [['a6d8d6545b09da76'], ['eb453d1c11da72d4'], ['46d291b62f5032b2'], ['eeb55a861130a21d']]; loss = 0.014671
+Epoch 0: |                                                                                                                                                        | 660/? [37:59<00:00,  0.29it/s, v_num=seh9]context = [[22, 33, 43, 47, 51, 53, 55, 56, 61, 71, 74, 78], [18, 31, 35, 41, 42, 43, 44, 47, 48, 67, 69, 73]]target = [[36, 65, 30, 66, 31, 27, 25, 43, 69, 76, 34, 45], [22, 37, 28, 23, 20, 19, 69, 24, 44, 72, 36, 30]]
+Epoch 0: |                                                                                                                                                        | 669/? [38:31<00:00,  0.29it/s, v_num=seh9]train step 670; scene = [['823a32785927191a'], ['ca2389f5e4fcfe61'], ['174ebd189316bd92'], ['b5260341870c7aa0']]; loss = 0.014347
+Epoch 0: |                                                                                                                                                        | 670/? [38:34<00:00,  0.29it/s, v_num=seh9]context = [[1, 5, 11, 22, 27, 29, 36, 40], [47, 67, 72, 89, 91, 94, 99, 101], [7, 16, 22, 23, 32, 38, 42, 47]]target = [[34, 2, 31, 30, 39, 35, 25, 6], [93, 58, 87, 51, 66, 92, 79, 98], [20, 29, 19, 33, 24, 21, 10, 43]]
+Epoch 0: |                                                                                                                                                        | 679/? [39:05<00:00,  0.29it/s, v_num=seh9]train step 680; scene = [['b703bdb45d172fe7'], ['d5d88e90c900def8'], ['7a20ba81fb778529'], ['bf2b8a204cdeb69f'], ['6e47af576b38cfa9'], ['c86cfcc517103af2'], ['17c29c5294714185'], ['97e7e5b8b6e493dd']]; loss = 0.027625
+Epoch 0: |                                                                                                                                                        | 680/? [39:08<00:00,  0.29it/s, v_num=seh9]context = [[2, 7, 32, 33, 35, 46, 49, 57], [45, 55, 69, 70, 74, 88, 89, 99], [60, 63, 74, 78, 81, 88, 91, 100]]target = [[39, 31, 7, 24, 45, 51, 26, 46], [78, 64, 90, 96, 73, 57, 67, 77], [81, 93, 66, 69, 84, 80, 85, 88]]
+Epoch 0: |                                                                                                                                                        | 689/? [39:40<00:00,  0.29it/s, v_num=seh9]train step 690; scene = [['43b4e4c3f5ae7c81'], ['80fdc99ef1b23262'], ['3ec9b76cab1555e4']]; loss = 0.016370
+Epoch 0: |                                                                                                                                                        | 690/? [39:44<00:00,  0.29it/s, v_num=seh9]context = [[125, 130, 131, 133, 139, 142, 153, 156, 160, 161, 166, 179], [0, 6, 18, 31, 32, 36, 50, 58, 59, 64, 65, 68]]target = [[133, 164, 151, 155, 128, 139, 142, 159, 160, 129, 174, 147], [29, 47, 32, 64, 33, 56, 63, 66, 24, 65, 8, 55]]
+Epoch 0: |                                                                                                                                                        | 699/? [40:16<00:00,  0.29it/s, v_num=seh9]train step 700; scene = [['3d30b440244efcd5'], ['8301348e86826b01']]; loss = 0.022419
+Epoch 0: |                                                                                                                                                        | 700/? [40:19<00:00,  0.29it/s, v_num=seh9]context = [[112, 118, 122, 123, 125, 131, 136, 139, 152, 154, 157, 165, 167, 168, 176, 177, 178, 186, 188, 194, 199, 203, 207, 209]]target = [[185, 195, 177, 197, 145, 155, 138, 151, 192, 130, 162, 184, 173, 113, 139, 202, 178, 198, 206, 182, 114, 169, 180, 115]]
+Epoch 0: |                                                                                                                                                        | 709/? [40:50<00:00,  0.29it/s, v_num=seh9]train step 710; scene = [['43c939b11c5fed4a']]; loss = 0.046366
+Epoch 0: |                                                                                                                                                        | 710/? [40:54<00:00,  0.29it/s, v_num=seh9]context = [[0, 3, 6, 11, 14, 27, 31, 36, 43, 46, 51, 54], [6, 8, 9, 23, 26, 28, 42, 44, 47, 60, 62, 63]]target = [[50, 44, 31, 45, 12, 7, 14, 43, 33, 11, 42, 20], [7, 27, 25, 55, 31, 51, 44, 21, 56, 60, 46, 57]]
+Epoch 0: |                                                                                                                                                        | 719/? [41:24<00:00,  0.29it/s, v_num=seh9]train step 720; scene = [['cf8fc3268c3034d6'], ['0715871c63748812']]; loss = 0.012882
+Epoch 0: |                                                                                                                                                        | 720/? [41:27<00:00,  0.29it/s, v_num=seh9]context = [[96, 103, 105, 116, 126, 128, 137, 145], [22, 31, 44, 45, 51, 52, 62, 79], [2, 19, 33, 55, 57, 58, 60, 66]]target = [[137, 129, 116, 118, 142, 119, 135, 108], [43, 55, 54, 69, 45, 38, 65, 27], [30, 28, 57, 15, 43, 16, 12, 3]]
+Epoch 0: |                                                                                                                                                        | 729/? [41:58<00:00,  0.29it/s, v_num=seh9]train step 730; scene = [['076a6c5199542c91']]; loss = 0.020287
+Epoch 0: |                                                                                                                                                        | 730/? [42:01<00:00,  0.29it/s, v_num=seh9]context = [[44, 67, 78, 90, 92, 93], [95, 98, 99, 119, 138, 140], [6, 16, 44, 64, 67, 69], [61, 89, 91, 97, 113, 131]]target = [[50, 84, 81, 90, 56, 67], [116, 108, 99, 107, 115, 103], [51, 39, 45, 53, 60, 38], [70, 67, 119, 62, 104, 75]]
+Epoch 0: |                                                                                                                                                        | 739/? [42:32<00:00,  0.29it/s, v_num=seh9]train step 740; scene = [['106fa5a7ff2a8839'], ['df378b8aa82b58d8'], ['cd6fdd63dcab9099']]; loss = 0.013965
+Epoch 0: |                                                                                                                                                        | 740/? [42:35<00:00,  0.29it/s, v_num=seh9]context = [[19, 23, 25, 28, 35, 39, 40, 44, 45, 47, 55, 62, 63, 69, 73, 76, 78, 85, 87, 92, 98, 99, 113, 116]]target = [[76, 92, 111, 107, 100, 47, 40, 87, 73, 57, 58, 66, 55, 109, 46, 88, 29, 51, 89, 69, 32, 53, 83, 93]]
+Epoch 0: |                                                                                                                                                        | 749/? [43:06<00:00,  0.29it/s, v_num=seh9]train step 750; scene = [['400098a7ab312bdc']]; loss = 0.073660
+Epoch 0: |                                                                                                                                                        | 750/? [43:09<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 750; scene = ['91fda69e1cda4602'];
+target intrinsic:  tensor(0.8937, device='cuda:0') tensor(0.8939, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9137, device='cuda:0') tensor(0.9128, device='cuda:0')
+[2026-02-25 02:24:20,918][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 750/? [43:11<00:00,  0.29it/s, v_num=seh9]context = [[105, 125, 144, 147, 160, 173], [31, 58, 61, 63, 68, 90], [11, 20, 27, 40, 42, 50], [127, 138, 144, 154, 173, 183]]target = [[167, 135, 112, 157, 107, 116], [68, 76, 58, 64, 52, 88], [18, 34, 30, 32, 27, 12], [180, 143, 138, 141, 154, 149]]
+Epoch 0: |                                                                                                                                                        | 759/? [43:41<00:00,  0.29it/s, v_num=seh9]train step 760; scene = [['882c2c3ccb2f06c3'], ['e5f6c40c0454e1aa'], ['d10eef0641b38981'], ['dd549a909b62ec24'], ['8d45251c07bebb7b'], ['8eb9e437797482f6']]; loss = 0.022826
+Epoch 0: |                                                                                                                                                        | 760/? [43:45<00:00,  0.29it/s, v_num=seh9]context = [[3, 13, 14, 21, 22, 23, 36, 44, 47, 53, 55, 56], [1, 9, 14, 15, 22, 23, 40, 41, 43, 44, 52, 63]]target = [[25, 46, 19, 38, 22, 31, 36, 8, 39, 17, 48, 28], [12, 29, 62, 39, 37, 55, 28, 56, 41, 60, 34, 48]]
+Epoch 0: |                                                                                                                                                        | 769/? [44:16<00:00,  0.29it/s, v_num=seh9]train step 770; scene = [['944e92ff3fea78eb'], ['f946d94d8adc2178']]; loss = 0.028902
+Epoch 0: |                                                                                                                                                        | 770/? [44:20<00:00,  0.29it/s, v_num=seh9]context = [[10, 17, 28, 40, 62, 64], [175, 183, 185, 199, 201, 239], [18, 21, 30, 44, 70, 84], [72, 78, 84, 94, 113, 117]]target = [[60, 55, 35, 49, 20, 17], [219, 207, 191, 196, 200, 212], [81, 66, 62, 36, 69, 70], [115, 83, 88, 116, 101, 113]]
+Epoch 0: |                                                                                                                                                        | 779/? [44:51<00:00,  0.29it/s, v_num=seh9]train step 780; scene = [['ea037bd5bd5dd020']]; loss = 0.008695
+Epoch 0: |                                                                                                                                                        | 780/? [44:55<00:00,  0.29it/s, v_num=seh9]context = [[27, 38, 49, 70], [6, 22, 76, 79], [5, 10, 24, 58], [56, 60, 78, 107], [63, 78, 89, 109], [22, 43, 73, 96]]target = [[66, 56, 58, 40], [34, 61, 42, 55], [57, 51, 26, 38], [76, 59, 103, 90], [105, 89, 100, 84], [93, 61, 63, 47]]
+Epoch 0: |                                                                                                                                                        | 789/? [45:26<00:00,  0.29it/s, v_num=seh9]train step 790; scene = [['a611f23daa8ebe85'], ['e6f06611137751ed'], ['f95e0bbb315250f4'], ['f9e4a4a27b9f0530'], ['65b40c140ed34d82'], ['643a5274a1acd8fe'], ['b1c505350a76d200'], ['db602b5c46a6037e'], ['5b7c70d9fd79a963'], ['a9cd1a8fc1fa2269'], ['111a2975a86f6e89'], ['42ac9272440cdbce']]; loss = 0.031850
+Epoch 0: |                                                                                                                                                        | 790/? [45:30<00:00,  0.29it/s, v_num=seh9]context = [[2, 8, 10, 13, 16, 18, 22, 24, 29, 32, 39, 44, 50, 54, 60, 70, 76, 79, 90, 93, 94, 95, 97, 99]]target = [[82, 57, 4, 35, 18, 6, 81, 47, 43, 97, 63, 20, 13, 56, 75, 39, 92, 51, 66, 71, 34, 86, 96, 64]]
+Epoch 0: |                                                                                                                                                        | 799/? [46:01<00:00,  0.29it/s, v_num=seh9]train step 800; scene = [['2698ac35434969d2'], ['75c89a32bc3a65d0'], ['f67633dab609f48a'], ['0c23c5d9c2010333']]; loss = 0.021363
+Epoch 0: |                                                                                                                                                        | 800/? [46:05<00:00,  0.29it/s, v_num=seh9]context = [[23, 33, 34, 41, 47, 50, 63, 67], [28, 33, 46, 64, 66, 76, 86, 90], [106, 119, 120, 134, 151, 156, 157, 158]]target = [[52, 58, 55, 61, 62, 36, 29, 31], [77, 40, 63, 79, 32, 57, 70, 42], [129, 154, 156, 119, 120, 124, 110, 157]]
+[2026-02-25 02:27:18,595][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                        | 809/? [46:36<00:00,  0.29it/s, v_num=seh9]train step 810; scene = [['a12ec328b1ce8ed3']]; loss = 0.011696
+Epoch 0: |                                                                                                                                                        | 810/? [46:39<00:00,  0.29it/s, v_num=seh9]context = [[6, 7, 8, 12, 16, 17, 22, 33, 36, 42, 44, 47, 56, 64, 66, 74, 75, 84, 86, 90, 92, 98, 100, 103]]target = [[31, 53, 66, 12, 16, 52, 25, 8, 35, 62, 42, 43, 81, 34, 49, 79, 82, 101, 84, 59, 83, 14, 29, 69]]
+Epoch 0: |                                                                                                                                                        | 819/? [47:10<00:00,  0.29it/s, v_num=seh9]train step 820; scene = [['13e915d66f8c9bf3'], ['f1dab83bdccd9a64'], ['afba4de2c78eb96f'], ['f97a862ad842d7b4']]; loss = 0.017552
+Epoch 0: |                                                                                                                                                        | 820/? [47:13<00:00,  0.29it/s, v_num=seh9]context = [[1, 56], [10, 74], [41, 84], [7, 49], [28, 87], [21, 65], [1, 57], [0, 49], [25, 95], [16, 90], [58, 104], [114, 157]]target = [[7, 24], [63, 72], [74, 43], [24, 28], [37, 45], [52, 35], [5, 27], [9, 41], [35, 58], [73, 82], [96, 89], [152, 148]]
+Epoch 0: |                                                                                                                                                        | 829/? [47:45<00:00,  0.29it/s, v_num=seh9]train step 830; scene = [['38d4b64a2c62a10c']]; loss = 0.012793
+Epoch 0: |                                                                                                                                                        | 830/? [47:49<00:00,  0.29it/s, v_num=seh9]context = [[1, 8, 34, 37, 41, 65, 66, 70, 71, 72, 74, 75], [39, 40, 42, 43, 56, 62, 66, 92, 98, 99, 101, 102]]target = [[34, 71, 31, 37, 55, 42, 72, 70, 49, 28, 18, 58], [93, 96, 61, 60, 73, 58, 76, 80, 47, 97, 98, 49]]
+Epoch 0: |                                                                                                                                                        | 839/? [48:20<00:00,  0.29it/s, v_num=seh9]train step 840; scene = [['6e46379a7ad5ea92'], ['2abf4fde25e3836a'], ['a86ad11e2813e855']]; loss = 0.013896
+Epoch 0: |                                                                                                                                                        | 840/? [48:24<00:00,  0.29it/s, v_num=seh9]context = [[1, 8, 22, 30, 31, 42, 53, 60, 62, 66, 69, 72], [76, 79, 85, 89, 90, 92, 101, 102, 104, 113, 127, 129]]target = [[50, 65, 54, 30, 11, 7, 51, 26, 33, 28, 19, 9], [113, 116, 84, 92, 115, 89, 104, 90, 83, 78, 98, 122]]
+Epoch 0: |                                                                                                                                                        | 849/? [48:55<00:00,  0.29it/s, v_num=seh9]train step 850; scene = [['95522bc4e658c4bb'], ['a54a8ec49d0c4c52'], ['6c292d2668c5df52']]; loss = 0.019966
+Epoch 0: |                                                                                                                                                        | 850/? [48:58<00:00,  0.29it/s, v_num=seh9]context = [[210, 276], [10, 52], [100, 172], [26, 85], [15, 79], [0, 43], [9, 65], [79, 122], [7, 50], [1, 76], [10, 86], [7, 58]]target = [[271, 260], [39, 32], [117, 155], [46, 71], [24, 72], [20, 4], [49, 30], [90, 110], [19, 16], [36, 35], [81, 85], [57, 17]]
+Epoch 0: |                                                                                                                                                        | 859/? [49:28<00:00,  0.29it/s, v_num=seh9]train step 860; scene = [['056b5030399a131f'], ['066424b7f19bf6e5']]; loss = 0.010934
+Epoch 0: |                                                                                                                                                        | 860/? [49:31<00:00,  0.29it/s, v_num=seh9]context = [[18, 30, 31, 45, 47, 59, 64, 66], [0, 13, 19, 25, 31, 54, 59, 69], [5, 7, 8, 24, 28, 43, 54, 57]]target = [[19, 22, 45, 35, 25, 39, 36, 33], [2, 62, 51, 18, 68, 19, 3, 41], [52, 19, 23, 34, 28, 55, 33, 26]]
+Epoch 0: |                                                                                                                                                        | 869/? [50:03<00:00,  0.29it/s, v_num=seh9]train step 870; scene = [['2f2587735ce5a386'], ['c0f128f783625ed0']]; loss = 0.012082
+Epoch 0: |                                                                                                                                                        | 870/? [50:06<00:00,  0.29it/s, v_num=seh9]context = [[0, 14, 25, 44], [4, 17, 71, 83], [3, 8, 47, 52], [75, 84, 126, 153], [7, 9, 43, 66], [74, 102, 104, 134]]target = [[11, 38, 22, 37], [54, 23, 45, 28], [12, 50, 5, 17], [90, 135, 84, 122], [35, 31, 11, 54], [76, 104, 116, 94]]
+Epoch 0: |                                                                                                                                                        | 879/? [50:38<00:00,  0.29it/s, v_num=seh9]train step 880; scene = [['fdd01ef3e4a926df'], ['430d79082d999336'], ['3140f7bea5597af4'], ['4b4fbc022aa36b37'], ['b903f35fbb538e4d'], ['39d9f692bfb58d80'], ['e9905e5bf1c49ce7'], ['f44cc142d9796ff7']]; loss = 0.031563
+Epoch 0: |                                                                                                                                                        | 880/? [50:42<00:00,  0.29it/s, v_num=seh9]context = [[67, 119, 129, 145], [46, 82, 98, 100], [69, 88, 89, 136], [45, 74, 106, 124], [2, 30, 43, 55], [29, 57, 72, 108]]target = [[84, 122, 110, 68], [83, 61, 84, 52], [100, 83, 74, 103], [90, 62, 77, 118], [43, 36, 50, 48], [100, 77, 92, 58]]
+Epoch 0: |                                                                                                                                                        | 889/? [51:13<00:00,  0.29it/s, v_num=seh9]train step 890; scene = [['58b43b9b210b78b8']]; loss = 0.024633
+Epoch 0: |                                                                                                                                                        | 890/? [51:17<00:00,  0.29it/s, v_num=seh9]context = [[1, 7, 40, 62], [7, 14, 23, 64], [112, 129, 170, 191], [10, 11, 35, 83], [11, 12, 63, 84], [4, 25, 59, 67]]target = [[20, 43, 34, 55], [40, 29, 43, 18], [190, 173, 175, 134], [55, 73, 69, 12], [76, 45, 65, 55], [54, 15, 25, 22]]
+Epoch 0: |                                                                                                                                                        | 899/? [51:47<00:00,  0.29it/s, v_num=seh9]train step 900; scene = [['db02cd4ba6a027da'], ['7b4630c7ece2e8ab'], ['7ff18d239739a030']]; loss = 0.011895
+Epoch 0: |                                                                                                                                                        | 900/? [51:51<00:00,  0.29it/s, v_num=seh9]context = [[31, 35, 86], [20, 35, 73], [66, 69, 139], [2, 5, 58], [26, 84, 88], [94, 153, 158], [1, 4, 51], [75, 128, 131]]target = [[44, 67, 39], [23, 55, 38], [135, 89, 103], [33, 24, 6], [72, 58, 70], [119, 112, 101], [47, 13, 37], [102, 111, 88]]
+Epoch 0: |                                                                                                                                                        | 909/? [52:22<00:00,  0.29it/s, v_num=seh9]train step 910; scene = [['e0222b577fdfde97'], ['245ccba767bcd121'], ['51587e352ab35ba1'], ['9b4d466924c40d8b']]; loss = 0.021657
+Epoch 0: |                                                                                                                                                        | 910/? [52:26<00:00,  0.29it/s, v_num=seh9]context = [[4, 7, 12, 13, 16, 22, 25, 30, 34, 46, 52, 57], [9, 19, 26, 31, 40, 41, 44, 46, 51, 57, 61, 63]]target = [[36, 53, 17, 24, 22, 20, 12, 50, 51, 39, 41, 33], [27, 41, 11, 20, 28, 46, 48, 32, 44, 12, 49, 31]]
+Epoch 0: |                                                                                                                                                        | 919/? [52:58<00:00,  0.29it/s, v_num=seh9]train step 920; scene = [['42512ca51e222e21'], ['d2739b0ce00e63cf'], ['7a92426dbce9e920']]; loss = 0.013897
+Epoch 0: |                                                                                                                                                        | 920/? [53:01<00:00,  0.29it/s, v_num=seh9]context = [[4, 25, 48, 58, 65, 78], [0, 8, 11, 34, 41, 45], [13, 16, 17, 22, 33, 57], [125, 135, 140, 149, 171, 202]]target = [[62, 56, 77, 28, 24, 42], [9, 11, 16, 40, 22, 23], [31, 47, 42, 54, 32, 33], [143, 195, 128, 146, 127, 151]]
+Epoch 0: |                                                                                                                                                        | 929/? [53:32<00:00,  0.29it/s, v_num=seh9]train step 930; scene = [['c8c034439fbe43f2'], ['4b1409aed425c619'], ['bac018643bed4fe8'], ['6920a4f8f496ef7c'], ['bfc897cfc8733b10'], ['56b98fa7036d9f70'], ['cb2506c57773a6bd'], ['c7f66b7e91a9e97e']]; loss = 0.021355
+Epoch 0: |                                                                                                                                                        | 930/? [53:35<00:00,  0.29it/s, v_num=seh9]context = [[15, 23, 27, 58, 64, 80], [38, 53, 54, 66, 79, 100], [0, 19, 24, 34, 55, 65], [5, 6, 16, 28, 30, 58]]target = [[66, 61, 41, 52, 33, 59], [90, 72, 39, 42, 82, 89], [36, 23, 10, 24, 11, 51], [27, 12, 40, 21, 42, 29]]
+Epoch 0: |                                                                                                                                                        | 939/? [54:07<00:00,  0.29it/s, v_num=seh9]train step 940; scene = [['4ae003ef13ad4ffd'], ['ef8f0fd78fcb44e0']]; loss = 0.009495
+Epoch 0: |                                                                                                                                                        | 940/? [54:09<00:00,  0.29it/s, v_num=seh9]context = [[33, 35, 45, 50, 69, 70, 87, 93], [40, 53, 63, 70, 78, 96, 97, 99], [31, 37, 39, 48, 57, 61, 71, 80]]target = [[42, 58, 77, 64, 63, 39, 82, 79], [71, 59, 76, 92, 84, 95, 61, 79], [74, 32, 44, 61, 72, 64, 73, 77]]
+Epoch 0: |                                                                                                                                                        | 949/? [54:40<00:00,  0.29it/s, v_num=seh9]train step 950; scene = [['cfe7d2764a367f81']]; loss = 0.024328
+Epoch 0: |                                                                                                                                                        | 950/? [54:43<00:00,  0.29it/s, v_num=seh9]context = [[160, 166, 180, 184, 202, 212, 241, 243], [212, 213, 216, 229, 230, 247, 271, 275], [13, 14, 27, 33, 47, 62, 64, 71]]target = [[172, 241, 230, 237, 235, 164, 207, 179], [242, 224, 219, 231, 244, 260, 227, 230], [55, 69, 22, 60, 66, 37, 38, 41]]
+Epoch 0: |                                                                                                                                                        | 959/? [55:14<00:00,  0.29it/s, v_num=seh9]train step 960; scene = [['ef6401c117a2701a'], ['7f9480301fa3e38b'], ['a2ae856e2faf3097'], ['7b4006c02dddd695'], ['5b33a7567dabea55'], ['be1b98bfac2059b2']]; loss = 0.020075
+Epoch 0: |                                                                                                                                                        | 960/? [55:18<00:00,  0.29it/s, v_num=seh9]context = [[139, 146, 152, 173, 214, 224], [16, 23, 33, 48, 66, 87], [20, 26, 54, 60, 63, 103], [42, 56, 61, 71, 74, 127]]target = [[174, 145, 187, 203, 152, 159], [49, 20, 43, 26, 55, 31], [80, 30, 56, 41, 21, 81], [51, 73, 78, 70, 97, 48]]
+Epoch 0: |                                                                                                                                                        | 969/? [55:49<00:00,  0.29it/s, v_num=seh9]train step 970; scene = [['77f1b53620811084'], ['3c077d5c8af0bd36']]; loss = 0.024646
+Epoch 0: |                                                                                                                                                        | 970/? [55:52<00:00,  0.29it/s, v_num=seh9]context = [[3, 15, 17, 31, 41, 45, 63, 70, 71, 76, 89, 90], [15, 16, 29, 31, 35, 52, 69, 73, 83, 86, 94, 98]]target = [[79, 61, 68, 22, 81, 12, 29, 74, 8, 4, 69, 35], [59, 86, 45, 77, 36, 25, 51, 16, 23, 91, 82, 24]]
+Epoch 0: |                                                                                                                                                        | 979/? [56:24<00:00,  0.29it/s, v_num=seh9]train step 980; scene = [['2ff32d3c912d6d4b'], ['b7ffe15e4ef99466'], ['477b227f30b26d5f']]; loss = 0.014626
+Epoch 0: |                                                                                                                                                        | 980/? [56:27<00:00,  0.29it/s, v_num=seh9]context = [[2, 4, 10, 15, 19, 31, 32, 35, 36, 48, 54, 59, 71, 73, 75, 78, 81, 84, 89, 90, 91, 92, 96, 99]]target = [[78, 75, 60, 26, 20, 87, 76, 85, 16, 58, 46, 42, 44, 6, 50, 15, 43, 62, 5, 96, 89, 66, 83, 92]]
+Epoch 0: |                                                                                                                                                        | 989/? [56:58<00:00,  0.29it/s, v_num=seh9]train step 990; scene = [['d78079ff1a0f045d'], ['3b42fa1245f6b00b'], ['b66059464cc436c0']]; loss = 0.018827
+Epoch 0: |                                                                                                                                                        | 990/? [57:02<00:00,  0.29it/s, v_num=seh9]context = [[168, 175, 176, 185, 186, 188, 197, 208, 211, 223, 233, 245], [21, 34, 37, 41, 42, 45, 48, 51, 58, 62, 65, 70]]target = [[202, 190, 206, 174, 221, 205, 196, 211, 232, 219, 238, 178], [50, 45, 51, 29, 63, 54, 23, 52, 26, 41, 43, 61]]
+Epoch 0: |                                                                                                                                                        | 999/? [57:33<00:00,  0.29it/s, v_num=seh9]train step 1000; scene = [['c2a108b0ff836b87'], ['395b6ac27ddcc9d1'], ['e80ac219ae75ce5e'], ['2ac1cf1adda42447'], ['2de40cefd70bc708'], ['25d17315f4472e27'], ['658d3f56767f0396'], ['b953af75bafccce8']]; loss = 0.017500
+Epoch 0: |                                                                                                                                                       | 1000/? [57:36<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1000; scene = ['647f2049bf4cb3f3'];
+target intrinsic:  tensor(0.8998, device='cuda:0') tensor(0.9001, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8922, device='cuda:0') tensor(0.8940, device='cuda:0')
+[2026-02-25 02:38:48,014][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                       | 1000/? [57:38<00:00,  0.29it/s, v_num=seh9]context = [[6, 9, 21, 32, 39, 40, 49, 54, 61, 62, 65, 68, 71, 75, 76, 78, 85, 87, 88, 91, 96, 99, 102, 103]]target = [[56, 19, 49, 28, 80, 26, 35, 23, 82, 39, 13, 15, 17, 44, 7, 58, 21, 36, 45, 40, 91, 53, 74, 70]]
+[2026-02-25 02:38:52,181][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                       | 1009/? [58:11<00:00,  0.29it/s, v_num=seh9]train step 1010; scene = [['2beac37f687333f5'], ['8030e59b90ab02d6'], ['6e58885ab481a69e'], ['2f3f9d1ca227e778'], ['8ebbf8173feadc66'], ['6755d1366bce2ef0']]; loss = 0.032081
+Epoch 0: |                                                                                                                                                       | 1010/? [58:14<00:00,  0.29it/s, v_num=seh9]context = [[88, 90, 94, 95, 98, 99, 103, 122, 125, 128, 143, 144, 147, 151, 152, 155, 161, 164, 165, 166, 173, 181, 182, 185]]target = [[163, 135, 183, 143, 105, 90, 169, 97, 151, 127, 140, 138, 91, 104, 157, 102, 146, 152, 145, 129, 100, 162, 99, 130]]
+Epoch 0: |                                                                                                                                                       | 1019/? [58:45<00:00,  0.29it/s, v_num=seh9]train step 1020; scene = [['dd657030a3c637d2']]; loss = 0.024582
+Epoch 0: |                                                                                                                                                       | 1020/? [58:49<00:00,  0.29it/s, v_num=seh9]context = [[77, 80, 91, 96, 98, 105, 108, 112, 130, 131, 134, 136], [10, 18, 20, 26, 34, 36, 65, 75, 78, 80, 81, 82]]target = [[102, 105, 104, 88, 119, 127, 110, 121, 101, 80, 84, 122], [65, 17, 15, 23, 21, 52, 48, 50, 38, 68, 42, 47]]
+Epoch 0: |                                                                                                                                                       | 1029/? [59:21<00:00,  0.29it/s, v_num=seh9]train step 1030; scene = [['b6a629f2e3cbf2c3'], ['04d96fd21593f793'], ['3dc32c403d742301']]; loss = 0.019946
+Epoch 0: |                                                                                                                                                       | 1030/? [59:24<00:00,  0.29it/s, v_num=seh9]context = [[58, 61, 64, 72, 73, 77, 79, 85, 95, 101, 103, 112], [13, 14, 21, 27, 36, 38, 49, 53, 54, 69, 71, 79]]target = [[78, 77, 97, 71, 100, 101, 99, 68, 73, 105, 87, 88], [53, 36, 27, 57, 47, 64, 28, 39, 49, 58, 33, 35]]
+Epoch 0: |                                                                                                                                                       | 1039/? [59:56<00:00,  0.29it/s, v_num=seh9]train step 1040; scene = [['94aa5fd022f4400b'], ['21be2154f3c073c1'], ['103ff72c7bde3074'], ['db811a2460c4f9b5'], ['27159dbf7b134cd9'], ['db61533e680f5d74']]; loss = 0.021292
+Epoch 0: |                                                                                                                                                       | 1040/? [59:59<00:00,  0.29it/s, v_num=seh9]context = [[1, 8, 13, 22, 24, 25, 31, 39, 43, 44, 47, 58, 60, 64, 65, 70, 76, 77, 85, 87, 90, 91, 94, 98]]target = [[30, 38, 34, 60, 10, 39, 94, 80, 35, 59, 20, 52, 6, 97, 61, 72, 40, 63, 93, 46, 41, 7, 55, 23]]
+Epoch 0: |                                                                                                                                                     | 1049/? [1:00:31<00:00,  0.29it/s, v_num=seh9]train step 1050; scene = [['94ddacdba86c4757'], ['2650dfe1a0b7976a'], ['7273382ad589d7ac'], ['dcfb0ddcac6008e6'], ['b382af5f342061fa'], ['d325bd35d81548c3']]; loss = 0.012752
+Epoch 0: |                                                                                                                                                     | 1050/? [1:00:34<00:00,  0.29it/s, v_num=seh9]context = [[68, 69, 80, 82, 84, 86, 90, 98, 104, 108, 117, 120, 122, 126, 136, 138, 144, 149, 150, 154, 157, 159, 160, 165]]target = [[138, 111, 147, 162, 96, 129, 95, 97, 103, 117, 90, 164, 78, 130, 155, 119, 123, 75, 107, 80, 82, 143, 72, 150]]
+Epoch 0: |                                                                                                                                                     | 1059/? [1:01:05<00:00,  0.29it/s, v_num=seh9]train step 1060; scene = [['1fdf19d7c4402d40'], ['57520a34789bd5d8'], ['5431b4af9712cad6']]; loss = 0.010836
+Epoch 0: |                                                                                                                                                     | 1060/? [1:01:09<00:00,  0.29it/s, v_num=seh9]context = [[23, 28, 29, 30, 35, 38, 41, 47, 56, 60, 61, 66, 83, 84, 86, 91, 95, 96, 102, 107, 109, 110, 111, 120]]target = [[49, 44, 112, 36, 32, 68, 78, 38, 75, 113, 34, 50, 91, 54, 116, 82, 33, 87, 24, 117, 48, 98, 43, 96]]
+Epoch 0: |                                                                                                                                                     | 1069/? [1:01:39<00:00,  0.29it/s, v_num=seh9]train step 1070; scene = [['16fd029febc69afa'], ['901e47c506f0d2c1']]; loss = 0.009375
+Epoch 0: |                                                                                                                                                     | 1070/? [1:01:43<00:00,  0.29it/s, v_num=seh9]context = [[1, 10, 55, 65], [52, 104, 105, 106], [124, 138, 171, 197], [1, 25, 73, 76], [9, 48, 62, 99], [129, 166, 172, 195]]target = [[32, 14, 59, 35], [87, 80, 69, 98], [186, 172, 133, 129], [53, 21, 42, 7], [57, 81, 40, 91], [161, 159, 155, 175]]
+Epoch 0: |                                                                                                                                                     | 1079/? [1:02:14<00:00,  0.29it/s, v_num=seh9]train step 1080; scene = [['99e7a4ff1897a94d'], ['d0e7e477ff1174d4'], ['b21723140f5d4a70'], ['5102042496230ec1'], ['3c5825296fdc2a5f'], ['f8532bb5eca0e29d'], ['c58abf74527c5ce1'], ['d7decc9863ca94ef']]; loss = 0.023664
+Epoch 0: |                                                                                                                                                     | 1080/? [1:02:18<00:00,  0.29it/s, v_num=seh9]context = [[6, 7, 10, 11, 13, 17, 28, 31, 32, 41, 46, 50, 51, 52, 57, 64, 65, 74, 80, 83, 86, 93, 101, 103]]target = [[44, 96, 81, 43, 67, 92, 30, 36, 83, 74, 79, 78, 59, 41, 85, 10, 16, 64, 52, 12, 24, 58, 35, 54]]
+Epoch 0: |                                                                                                                                                     | 1089/? [1:02:49<00:00,  0.29it/s, v_num=seh9]train step 1090; scene = [['9724d02dd7954ece'], ['0cf733dbbb0e017f'], ['436ccd6836a9c74e'], ['d4cc5c4887f72d3e'], ['d3711afbbda1b025'], ['18bc67e83f83fd74'], ['2517e14bc498b4f8'], ['8edf724cd98c75c0']]; loss = 0.025556
+Epoch 0: |                                                                                                                                                     | 1090/? [1:02:53<00:00,  0.29it/s, v_num=seh9]context = [[5, 9, 16, 17, 18, 19, 24, 27, 32, 42, 64, 71], [19, 22, 24, 36, 40, 51, 52, 57, 60, 62, 65, 70]]target = [[9, 13, 11, 40, 43, 34, 6, 52, 18, 61, 56, 49], [51, 48, 39, 43, 58, 68, 25, 44, 21, 22, 29, 62]]
+Epoch 0: |                                                                                                                                                     | 1099/? [1:03:24<00:00,  0.29it/s, v_num=seh9]train step 1100; scene = [['6cee8362ee57a4c7']]; loss = 0.014239
+Epoch 0: |                                                                                                                                                     | 1100/? [1:03:27<00:00,  0.29it/s, v_num=seh9]context = [[140, 161, 164, 170, 177, 205, 207, 217], [11, 12, 17, 21, 27, 44, 63, 72], [0, 7, 10, 29, 45, 54, 55, 62]]target = [[167, 185, 195, 197, 177, 196, 168, 192], [17, 42, 22, 47, 65, 21, 50, 56], [3, 58, 22, 9, 25, 24, 37, 52]]
+Epoch 0: |                                                                                                                                                     | 1109/? [1:03:58<00:00,  0.29it/s, v_num=seh9]train step 1110; scene = [['7b2c118f021e6902'], ['05ed7f84c7d34ed9'], ['ced7fe64b7867119'], ['e1d9628095f76a9f']]; loss = 0.018727
+Epoch 0: |                                                                                                                                                     | 1110/? [1:04:01<00:00,  0.29it/s, v_num=seh9]context = [[4, 11, 12, 16, 20, 24, 25, 29, 33, 35, 44, 46, 47, 53, 57, 58, 62, 63, 83, 84, 90, 94, 98, 101]]target = [[98, 46, 60, 18, 67, 28, 23, 16, 37, 34, 53, 71, 80, 57, 88, 64, 54, 36, 42, 78, 17, 89, 27, 45]]
+Epoch 0: |                                                                                                                                                     | 1119/? [1:04:32<00:00,  0.29it/s, v_num=seh9]train step 1120; scene = [['1b449c2b2ba10206']]; loss = 0.043976
+Epoch 0: |                                                                                                                                                     | 1120/? [1:04:36<00:00,  0.29it/s, v_num=seh9]context = [[161, 199, 216, 230], [200, 229, 244, 248], [86, 110, 132, 147], [21, 31, 50, 70], [7, 56, 60, 71], [0, 56, 73, 84]]target = [[218, 207, 225, 224], [236, 220, 241, 212], [100, 131, 108, 89], [31, 35, 58, 32], [13, 59, 40, 22], [37, 35, 29, 70]]
+Epoch 0: |                                                                                                                                                     | 1129/? [1:05:06<00:00,  0.29it/s, v_num=seh9]train step 1130; scene = [['373f042a22f7d6ff'], ['4fd852226711ffa9'], ['013e509c64d311ba'], ['2dc9c96992a50777']]; loss = 0.015268
+Epoch 0: |                                                                                                                                                     | 1130/? [1:05:10<00:00,  0.29it/s, v_num=seh9]context = [[105, 136, 160], [2, 89, 90], [42, 113, 117], [16, 53, 66], [82, 110, 133], [4, 17, 62], [111, 147, 181], [25, 27, 79]]target = [[121, 151, 140], [28, 11, 50], [97, 75, 99], [64, 24, 29], [116, 105, 120], [37, 38, 59], [147, 136, 154], [43, 68, 33]]
+Epoch 0: |                                                                                                                                                     | 1139/? [1:05:40<00:00,  0.29it/s, v_num=seh9]train step 1140; scene = [['04719d552f4d9d25'], ['066f14037c220103'], ['2d6dc5466ac4ed93'], ['4f9716bb3dc7feec']]; loss = 0.012820
+Epoch 0: |                                                                                                                                                     | 1140/? [1:05:44<00:00,  0.29it/s, v_num=seh9]context = [[23, 33, 34, 35, 51, 58, 70, 71, 89, 91, 95, 106], [10, 16, 24, 26, 30, 40, 45, 46, 48, 53, 55, 64]]target = [[53, 100, 54, 79, 60, 98, 36, 78, 105, 28, 52, 42], [59, 31, 46, 41, 63, 58, 22, 43, 17, 24, 13, 15]]
+Epoch 0: |                                                                                                                                                     | 1149/? [1:06:16<00:00,  0.29it/s, v_num=seh9]train step 1150; scene = [['308ec88c1ad9b768'], ['b3c43704ee2efd4f'], ['7239b2a7554c75fe'], ['5645a008715acf0a']]; loss = 0.016435
+Epoch 0: |                                                                                                                                                     | 1150/? [1:06:19<00:00,  0.29it/s, v_num=seh9]context = [[10, 20, 28, 29, 34, 35, 40, 42, 50, 51, 54, 59], [18, 19, 26, 31, 34, 36, 37, 39, 46, 52, 69, 76]]target = [[29, 38, 24, 15, 11, 54, 21, 26, 28, 36, 19, 37], [71, 38, 24, 72, 27, 29, 30, 19, 60, 58, 53, 40]]
+Epoch 0: |                                                                                                                                                     | 1159/? [1:06:50<00:00,  0.29it/s, v_num=seh9]train step 1160; scene = [['016349b94babf6d4'], ['d72595b1e4d300b8'], ['b072d5e75e95e9cd'], ['3dc75de99ef09a0c'], ['8cfd52fbbac85c10'], ['742eaf7317415c35']]; loss = 0.036689
+Epoch 0: |                                                                                                                                                     | 1160/? [1:06:53<00:00,  0.29it/s, v_num=seh9]context = [[87, 88, 95, 99, 102, 112, 125, 156, 159, 160, 162, 167], [3, 4, 6, 8, 22, 25, 27, 30, 33, 37, 48, 58]]target = [[104, 129, 155, 116, 114, 147, 98, 143, 108, 162, 166, 131], [20, 40, 11, 42, 7, 16, 21, 56, 27, 50, 28, 6]]
+Epoch 0: |                                                                                                                                                     | 1169/? [1:07:25<00:00,  0.29it/s, v_num=seh9]train step 1170; scene = [['f75b1a55b42e01fc'], ['fc33d4814d21b0ff']]; loss = 0.007836
+Epoch 0: |                                                                                                                                                     | 1170/? [1:07:29<00:00,  0.29it/s, v_num=seh9]context = [[10, 31, 58], [30, 58, 84], [178, 228, 258], [17, 48, 85], [56, 96, 104], [99, 112, 153], [28, 44, 91], [4, 16, 60]]target = [[37, 30, 54], [35, 45, 78], [184, 229, 232], [38, 66, 72], [69, 81, 101], [143, 127, 121], [33, 89, 50], [9, 18, 29]]
+Epoch 0: |                                                                                                                                                     | 1179/? [1:07:59<00:00,  0.29it/s, v_num=seh9]train step 1180; scene = [['93b36a54151e085e']]; loss = 0.009968
+Epoch 0: |                                                                                                                                                     | 1180/? [1:08:03<00:00,  0.29it/s, v_num=seh9]context = [[35, 69, 77, 82], [22, 66, 77, 92], [3, 23, 51, 59], [0, 11, 35, 67], [179, 193, 225, 245], [60, 64, 90, 110]]target = [[59, 78, 50, 41], [29, 66, 30, 42], [35, 30, 12, 5], [29, 46, 1, 16], [222, 211, 185, 205], [61, 78, 84, 71]]
+Epoch 0: |                                                                                                                                                     | 1189/? [1:08:35<00:00,  0.29it/s, v_num=seh9]train step 1190; scene = [['cc20d092a599c493']]; loss = 0.008323
+Epoch 0: |                                                                                                                                                     | 1190/? [1:08:38<00:00,  0.29it/s, v_num=seh9]context = [[151, 152, 163, 164, 166, 176, 186, 189, 190, 191, 196, 197, 201, 203, 206, 211, 218, 220, 226, 233, 235, 242, 244, 248]]target = [[183, 188, 157, 243, 213, 164, 235, 182, 234, 225, 247, 221, 194, 212, 218, 224, 153, 198, 214, 197, 184, 166, 210, 191]]
+Epoch 0: |                                                                                                                                                     | 1199/? [1:09:09<00:00,  0.29it/s, v_num=seh9]train step 1200; scene = [['ad3c210c5cdc595f'], ['fca33c5a37b6c7bf']]; loss = 0.017562
+Epoch 0: |                                                                                                                                                     | 1200/? [1:09:12<00:00,  0.29it/s, v_num=seh9]context = [[0, 8, 13, 14, 21, 61, 83, 88], [4, 25, 31, 32, 42, 47, 55, 56], [60, 65, 69, 74, 92, 105, 114, 133]]target = [[39, 31, 85, 21, 17, 38, 22, 29], [18, 41, 37, 6, 22, 52, 39, 31], [77, 123, 120, 62, 65, 68, 78, 67]]
+[2026-02-25 02:50:26,889][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1209/? [1:09:44<00:00,  0.29it/s, v_num=seh9]train step 1210; scene = [['41986a305a19ca1d'], ['0b707735ebc3eeed']]; loss = 0.014840
+Epoch 0: |                                                                                                                                                     | 1210/? [1:09:47<00:00,  0.29it/s, v_num=seh9]context = [[135, 145, 149, 150, 152, 153, 157, 158, 160, 163, 167, 174, 178, 183, 185, 187, 192, 193, 195, 198, 204, 212, 230, 232]]target = [[157, 224, 203, 149, 225, 192, 197, 217, 139, 142, 211, 218, 215, 156, 173, 155, 188, 136, 160, 180, 222, 178, 137, 204]]
+Epoch 0: |                                                                                                                                                     | 1219/? [1:10:19<00:00,  0.29it/s, v_num=seh9]train step 1220; scene = [['5795c143b1adc665'], ['d51d569c27d0b2b5'], ['f51a6709d0525ad3']]; loss = 0.011346
+Epoch 0: |                                                                                                                                                     | 1220/? [1:10:23<00:00,  0.29it/s, v_num=seh9]context = [[0, 11, 45, 56], [3, 17, 48, 54], [59, 78, 108, 111], [15, 30, 46, 71], [3, 39, 51, 53], [21, 46, 71, 99]]target = [[9, 30, 19, 54], [18, 27, 48, 49], [110, 101, 89, 90], [33, 58, 57, 70], [27, 41, 36, 43], [75, 35, 48, 49]]
+Epoch 0: |                                                                                                                                                     | 1229/? [1:10:54<00:00,  0.29it/s, v_num=seh9]train step 1230; scene = [['3d2496a06b64ab52'], ['95a572d344c5c54f']]; loss = 0.008285
+Epoch 0: |                                                                                                                                                     | 1230/? [1:10:58<00:00,  0.29it/s, v_num=seh9]context = [[12, 21, 27, 73, 83, 95, 98, 99], [15, 27, 37, 59, 68, 78, 80, 98], [181, 182, 200, 210, 221, 236, 250, 257]]target = [[62, 45, 80, 83, 92, 37, 72, 56], [74, 95, 23, 71, 43, 89, 19, 56], [193, 224, 203, 192, 249, 255, 202, 207]]
+Epoch 0: |                                                                                                                                                     | 1239/? [1:11:28<00:00,  0.29it/s, v_num=seh9]train step 1240; scene = [['b07961433f13f70e'], ['89a3b0eed03cf522']]; loss = 0.009711
+Epoch 0: |                                                                                                                                                     | 1240/? [1:11:32<00:00,  0.29it/s, v_num=seh9]context = [[200, 206, 207, 208, 209, 216, 218, 219, 226, 233, 237, 261], [16, 21, 25, 30, 32, 61, 67, 73, 76, 94, 97, 99]]target = [[252, 211, 215, 220, 225, 235, 239, 203, 229, 217, 231, 230], [64, 42, 62, 98, 88, 28, 63, 66, 45, 90, 36, 89]]
+Epoch 0: |                                                                                                                                                     | 1249/? [1:12:04<00:00,  0.29it/s, v_num=seh9]train step 1250; scene = [['8bd8e3f6166b3bf2']]; loss = 0.035505
+Epoch 0: |                                                                                                                                                     | 1250/? [1:12:07<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1250; scene = ['70b0a33083333dc9'];
+target intrinsic:  tensor(0.8872, device='cuda:0') tensor(0.8874, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8738, device='cuda:0') tensor(0.8725, device='cuda:0')
+[2026-02-25 02:53:18,593][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1250/? [1:12:08<00:00,  0.29it/s, v_num=seh9]context = [[49, 53, 64, 77, 88, 90, 103, 106], [33, 57, 60, 68, 70, 77, 78, 79], [67, 85, 87, 108, 109, 112, 131, 136]]target = [[98, 95, 87, 69, 100, 50, 82, 61], [36, 74, 51, 57, 62, 35, 77, 39], [119, 68, 85, 82, 77, 113, 96, 73]]
+Epoch 0: |                                                                                                                                                     | 1259/? [1:12:39<00:00,  0.29it/s, v_num=seh9]train step 1260; scene = [['8875b49387b58221']]; loss = 0.010358
+Epoch 0: |                                                                                                                                                     | 1260/? [1:12:42<00:00,  0.29it/s, v_num=seh9]context = [[62, 63, 66, 67, 71, 78, 82, 83, 86, 92, 95, 99, 102, 106, 108, 110, 119, 125, 126, 140, 146, 147, 152, 159]]target = [[88, 73, 95, 75, 125, 147, 90, 157, 91, 79, 104, 120, 132, 98, 107, 111, 118, 129, 154, 112, 114, 156, 134, 78]]
+Epoch 0: |                                                                                                                                                     | 1269/? [1:13:13<00:00,  0.29it/s, v_num=seh9]train step 1270; scene = [['336b3140d9d8bebd']]; loss = 0.012858
+Epoch 0: |                                                                                                                                                     | 1270/? [1:13:16<00:00,  0.29it/s, v_num=seh9]context = [[7, 10, 11, 24, 28, 33, 35, 47, 51, 57, 63, 64], [151, 162, 169, 174, 176, 179, 180, 181, 187, 197, 213, 216]]target = [[25, 26, 24, 54, 8, 18, 55, 38, 15, 58, 41, 61], [179, 193, 206, 204, 202, 191, 187, 166, 189, 182, 154, 196]]
+Epoch 0: |                                                                                                                                                     | 1279/? [1:13:47<00:00,  0.29it/s, v_num=seh9]train step 1280; scene = [['06ed4433d2640109'], ['6ca39802dcef328e']]; loss = 0.015196
+Epoch 0: |                                                                                                                                                     | 1280/? [1:13:50<00:00,  0.29it/s, v_num=seh9]context = [[215, 218, 240, 241, 254, 272], [47, 63, 108, 113, 114, 116], [77, 91, 108, 114, 118, 151], [19, 22, 24, 36, 72, 74]]target = [[242, 267, 217, 259, 226, 220], [48, 63, 79, 78, 100, 108], [81, 147, 143, 129, 82, 112], [53, 71, 20, 36, 21, 41]]
+Epoch 0: |                                                                                                                                                     | 1289/? [1:14:22<00:00,  0.29it/s, v_num=seh9]train step 1290; scene = [['9138f683b523d708'], ['c1289a667237cd36'], ['9c1dfe99d620e067'], ['77c03290d901ec15'], ['37f2b83c868084df'], ['06ea674963ad018e'], ['4797e813ea4a5ed2'], ['aa5175d18a929cb8']]; loss = 0.016519
+Epoch 0: |                                                                                                                                                     | 1290/? [1:14:25<00:00,  0.29it/s, v_num=seh9]context = [[198, 205, 214, 215, 235, 236, 245, 246], [21, 24, 28, 52, 75, 78, 83, 108], [8, 18, 20, 49, 60, 80, 83, 86]]target = [[222, 235, 234, 215, 219, 240, 218, 200], [24, 35, 28, 23, 46, 32, 90, 84], [33, 38, 57, 55, 20, 52, 58, 26]]
+Epoch 0: |                                                                                                                                                     | 1299/? [1:14:55<00:00,  0.29it/s, v_num=seh9]train step 1300; scene = [['6b2e49b1e748eb08'], ['20a321699aab25aa'], ['d15f97346b6c753c']]; loss = 0.021752
+Epoch 0: |                                                                                                                                                     | 1300/? [1:14:58<00:00,  0.29it/s, v_num=seh9]context = [[104, 106, 107, 112, 113, 116, 119, 120, 128, 137, 138, 146, 148, 150, 152, 153, 158, 164, 168, 173, 177, 193, 198, 201]]target = [[117, 169, 142, 191, 108, 125, 146, 177, 126, 116, 155, 161, 135, 132, 110, 122, 160, 199, 162, 190, 150, 168, 127, 140]]
+Epoch 0: |                                                                                                                                                     | 1309/? [1:15:30<00:00,  0.29it/s, v_num=seh9]train step 1310; scene = [['a4db4f86a3f6a3c9']]; loss = 0.021121
+Epoch 0: |                                                                                                                                                     | 1310/? [1:15:34<00:00,  0.29it/s, v_num=seh9]context = [[95, 103, 109, 123, 131, 140, 149, 153], [2, 3, 10, 15, 31, 53, 71, 75], [97, 98, 114, 118, 121, 126, 137, 143]]target = [[142, 152, 122, 137, 115, 143, 109, 107], [64, 11, 71, 42, 74, 41, 39, 24], [117, 99, 135, 133, 98, 116, 127, 129]]
+Epoch 0: |                                                                                                                                                     | 1319/? [1:16:05<00:00,  0.29it/s, v_num=seh9]train step 1320; scene = [['651489c4eabb5ef9']]; loss = 0.014759
+Epoch 0: |                                                                                                                                                     | 1320/? [1:16:08<00:00,  0.29it/s, v_num=seh9]context = [[5, 7, 14, 25, 27, 30, 34, 37, 39, 43, 47, 51, 55, 59, 63, 69, 78, 81, 87, 90, 93, 96, 99, 102]]target = [[40, 74, 11, 26, 92, 6, 61, 96, 12, 78, 19, 93, 54, 9, 33, 21, 66, 62, 32, 52, 60, 99, 41, 83]]
+Epoch 0: |                                                                                                                                                     | 1329/? [1:16:40<00:00,  0.29it/s, v_num=seh9]train step 1330; scene = [['36047ec1694f9d49'], ['1781ab25cd748560'], ['23ac87fd90e9a51a']]; loss = 0.033708
+Epoch 0: |                                                                                                                                                     | 1330/? [1:16:44<00:00,  0.29it/s, v_num=seh9]context = [[105, 138, 140, 169, 183, 184], [36, 62, 69, 77, 87, 91], [24, 43, 49, 54, 95, 97], [69, 84, 88, 104, 106, 123]]target = [[167, 150, 143, 112, 123, 113], [81, 61, 85, 38, 57, 84], [52, 41, 36, 89, 91, 63], [76, 78, 83, 74, 96, 86]]
+Epoch 0: |                                                                                                                                                     | 1339/? [1:17:14<00:00,  0.29it/s, v_num=seh9]train step 1340; scene = [['1eb41a7aa81df97b']]; loss = 0.009225
+Epoch 0: |                                                                                                                                                     | 1340/? [1:17:18<00:00,  0.29it/s, v_num=seh9]context = [[32, 46, 61, 65, 77, 96, 101, 102], [18, 23, 25, 40, 57, 58, 71, 80], [1, 6, 15, 35, 42, 52, 67, 75]]target = [[50, 85, 100, 73, 48, 89, 44, 42], [19, 36, 69, 52, 72, 70, 22, 67], [14, 13, 26, 23, 47, 53, 17, 34]]
+Epoch 0: |                                                                                                                                                     | 1349/? [1:17:49<00:00,  0.29it/s, v_num=seh9]train step 1350; scene = [['aa293d98738dc39f'], ['40dc627bd4ddc389'], ['fb45a86f6154e126'], ['baa5664b5ea5348e'], ['bd814603b3e38fb1'], ['252a4d522d463417'], ['68df2105110fbafe'], ['19dbf45ccabba3ff']]; loss = 0.014541
+Epoch 0: |                                                                                                                                                     | 1350/? [1:17:53<00:00,  0.29it/s, v_num=seh9]context = [[10, 13, 14, 22, 24, 27, 29, 31, 34, 38, 50, 59, 60, 61, 62, 65, 68, 71, 81, 87, 93, 103, 104, 107]]target = [[59, 64, 73, 49, 15, 84, 12, 70, 47, 33, 65, 97, 101, 75, 25, 27, 28, 42, 76, 17, 98, 86, 88, 68]]
+Epoch 0: |                                                                                                                                                     | 1359/? [1:18:24<00:00,  0.29it/s, v_num=seh9]train step 1360; scene = [['2d78bdcafda0e615'], ['fb526fa7f45a3ae4']]; loss = 0.013928
+Epoch 0: |                                                                                                                                                     | 1360/? [1:18:28<00:00,  0.29it/s, v_num=seh9]context = [[32, 53, 57, 60, 98, 103, 104, 119], [10, 16, 17, 25, 30, 34, 58, 65], [6, 40, 41, 43, 46, 52, 58, 63]]target = [[112, 55, 85, 93, 52, 106, 102, 92], [57, 58, 33, 47, 63, 64, 50, 59], [56, 30, 40, 10, 29, 50, 37, 59]]
+Epoch 0: |                                                                                                                                                     | 1369/? [1:18:59<00:00,  0.29it/s, v_num=seh9]train step 1370; scene = [['b7aeefdcf8c006ec']]; loss = 0.009936
+Epoch 0: |                                                                                                                                                     | 1370/? [1:19:03<00:00,  0.29it/s, v_num=seh9]context = [[42, 53, 56, 57, 63, 70, 77, 87], [21, 30, 36, 54, 60, 73, 84, 91], [121, 130, 134, 141, 151, 160, 161, 166]]target = [[46, 85, 73, 76, 49, 52, 75, 60], [59, 83, 46, 71, 74, 23, 82, 49], [156, 130, 139, 123, 161, 158, 145, 135]]
+Epoch 0: |                                                                                                                                                     | 1379/? [1:19:34<00:00,  0.29it/s, v_num=seh9]train step 1380; scene = [['db25ce184054217a']]; loss = 0.008563
+Epoch 0: |                                                                                                                                                     | 1380/? [1:19:37<00:00,  0.29it/s, v_num=seh9]context = [[20, 21, 23, 34, 36, 43, 46, 47, 49, 60, 61, 65, 68, 83, 84, 86, 87, 91, 93, 94, 106, 112, 113, 117]]target = [[22, 102, 111, 79, 52, 59, 108, 32, 88, 90, 68, 39, 113, 24, 71, 85, 89, 67, 62, 30, 95, 92, 44, 86]]
+Epoch 0: |                                                                                                                                                     | 1389/? [1:20:09<00:00,  0.29it/s, v_num=seh9]train step 1390; scene = [['73341b79eaf41f6a'], ['7f4353922e24e719'], ['572f15c53545baaf']]; loss = 0.010275
+Epoch 0: |                                                                                                                                                     | 1390/? [1:20:13<00:00,  0.29it/s, v_num=seh9]context = [[64, 65, 69, 73, 74, 77, 81, 89, 96, 99, 101, 103, 104, 108, 114, 116, 118, 129, 132, 137, 138, 146, 150, 161]]target = [[68, 159, 160, 90, 82, 79, 78, 110, 155, 108, 115, 65, 121, 153, 67, 156, 92, 139, 114, 73, 80, 107, 109, 135]]
+Epoch 0: |                                                                                                                                                     | 1399/? [1:20:45<00:00,  0.29it/s, v_num=seh9]train step 1400; scene = [['7d634ffc2785a9f4'], ['acd293363db06c1c']]; loss = 0.013200
+Epoch 0: |                                                                                                                                                     | 1400/? [1:20:48<00:00,  0.29it/s, v_num=seh9]context = [[48, 50, 53, 54, 75, 87, 89, 115], [5, 20, 24, 31, 38, 72, 78, 85], [11, 18, 24, 53, 66, 76, 88, 91]]target = [[77, 79, 96, 75, 102, 87, 62, 54], [14, 61, 24, 79, 43, 11, 13, 77], [38, 25, 29, 20, 69, 43, 60, 28]]
+[2026-02-25 03:02:02,967][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1409/? [1:21:20<00:00,  0.29it/s, v_num=seh9]train step 1410; scene = [['3ec980a78e9467e0'], ['6744c1cf4e4dee91']]; loss = 0.009663
+Epoch 0: |                                                                                                                                                     | 1410/? [1:21:23<00:00,  0.29it/s, v_num=seh9]context = [[1, 2, 6, 9, 17, 27, 43, 80], [40, 43, 54, 57, 67, 73, 97, 103], [89, 91, 110, 112, 125, 131, 140, 147]]target = [[39, 53, 4, 2, 75, 68, 31, 51], [102, 56, 62, 45, 92, 90, 60, 75], [127, 128, 144, 94, 130, 124, 113, 112]]
+Epoch 0: |                                                                                                                                                     | 1419/? [1:21:55<00:00,  0.29it/s, v_num=seh9]train step 1420; scene = [['9ac8241a27b0ec6d']]; loss = 0.015695
+Epoch 0: |                                                                                                                                                     | 1420/? [1:21:59<00:00,  0.29it/s, v_num=seh9]context = [[29, 36, 38, 41, 42, 43, 48, 51, 56, 59, 62, 69, 78, 87, 90, 92, 94, 97, 99, 101, 105, 106, 124, 126]]target = [[41, 97, 113, 39, 116, 48, 107, 102, 36, 121, 54, 50, 115, 61, 94, 123, 57, 82, 86, 30, 101, 93, 117, 32]]
+Epoch 0: |                                                                                                                                                     | 1429/? [1:22:30<00:00,  0.29it/s, v_num=seh9]train step 1430; scene = [['54d3668b6c7ed4b8']]; loss = 0.009761
+Epoch 0: |                                                                                                                                                     | 1430/? [1:22:34<00:00,  0.29it/s, v_num=seh9]context = [[12, 18, 32, 39, 56, 72, 76, 92], [70, 74, 86, 100, 115, 118, 125, 128], [22, 54, 56, 65, 68, 70, 72, 77]]target = [[29, 33, 27, 22, 90, 81, 54, 63], [123, 120, 72, 92, 74, 101, 104, 125], [49, 56, 58, 69, 31, 37, 32, 73]]
+Epoch 0: |                                                                                                                                                     | 1439/? [1:23:05<00:00,  0.29it/s, v_num=seh9]train step 1440; scene = [['070e6ca07b5f898c'], ['6c383c3e7ece2df7'], ['fbc5d8715d6debec']]; loss = 0.011603
+Epoch 0: |                                                                                                                                                     | 1440/? [1:23:09<00:00,  0.29it/s, v_num=seh9]context = [[1, 7, 12, 17, 24, 26, 27, 29, 32, 34, 38, 42, 45, 55, 63, 67, 69, 71, 72, 75, 88, 91, 95, 98]]target = [[62, 12, 35, 63, 55, 46, 47, 38, 82, 22, 4, 89, 75, 25, 9, 54, 70, 60, 76, 90, 28, 2, 67, 18]]
+Epoch 0: |                                                                                                                                                     | 1449/? [1:23:39<00:00,  0.29it/s, v_num=seh9]train step 1450; scene = [['61aef773e015ee1b'], ['895fe07be01436f1']]; loss = 0.014014
+Epoch 0: |                                                                                                                                                     | 1450/? [1:23:42<00:00,  0.29it/s, v_num=seh9]context = [[52, 60, 116], [92, 107, 146], [81, 134, 149], [11, 15, 85], [60, 67, 136], [32, 78, 121], [9, 24, 78], [25, 52, 77]]target = [[63, 103, 95], [107, 94, 138], [131, 129, 114], [73, 36, 42], [73, 65, 123], [61, 83, 107], [50, 38, 34], [42, 62, 74]]
+Epoch 0: |                                                                                                                                                     | 1459/? [1:24:13<00:00,  0.29it/s, v_num=seh9]train step 1460; scene = [['913bc26ba7d3afc5']]; loss = 0.023541
+Epoch 0: |                                                                                                                                                     | 1460/? [1:24:16<00:00,  0.29it/s, v_num=seh9]context = [[53, 54, 55, 58, 67, 68, 74, 76, 81, 82, 85, 86, 97, 104, 109, 112, 114, 115, 119, 127, 141, 142, 144, 150]]target = [[71, 143, 95, 81, 66, 54, 140, 125, 141, 94, 131, 146, 86, 124, 129, 144, 92, 111, 91, 110, 138, 93, 109, 128]]
+Epoch 0: |                                                                                                                                                     | 1469/? [1:24:47<00:00,  0.29it/s, v_num=seh9]train step 1470; scene = [['f73db02fdfe72073'], ['1bb8fe74783c5161'], ['7fa967710229854e'], ['4ec72717091e92c7']]; loss = 0.009061
+Epoch 0: |                                                                                                                                                     | 1470/? [1:24:51<00:00,  0.29it/s, v_num=seh9]context = [[115, 121, 127, 134, 143, 144, 149, 150, 151, 161, 165, 197], [1, 3, 4, 11, 29, 31, 37, 49, 58, 72, 80, 87]]target = [[143, 184, 154, 183, 132, 151, 157, 144, 168, 130, 169, 153], [77, 4, 82, 21, 18, 76, 29, 10, 71, 15, 56, 50]]
+Epoch 0: |                                                                                                                                                     | 1479/? [1:25:21<00:00,  0.29it/s, v_num=seh9]train step 1480; scene = [['8b91f281b1c3feec'], ['801309986851fc16']]; loss = 0.011668
+Epoch 0: |                                                                                                                                                     | 1480/? [1:25:25<00:00,  0.29it/s, v_num=seh9]context = [[27, 30, 33, 36, 57, 58, 59, 61, 71, 75, 82, 93], [78, 79, 82, 87, 98, 99, 108, 110, 123, 125, 132, 136]]target = [[47, 64, 90, 74, 32, 44, 79, 86, 81, 88, 30, 82], [81, 104, 82, 119, 130, 113, 97, 87, 105, 125, 93, 102]]
+Epoch 0: |                                                                                                                                                     | 1489/? [1:25:57<00:00,  0.29it/s, v_num=seh9]train step 1490; scene = [['bbc6cab96f65f530'], ['62755beb1ebd7813']]; loss = 0.013427
+Epoch 0: |                                                                                                                                                     | 1490/? [1:26:00<00:00,  0.29it/s, v_num=seh9]context = [[17, 37, 94], [50, 65, 101], [89, 93, 143], [1, 27, 61], [17, 66, 81], [1, 50, 52], [23, 70, 78], [1, 17, 72]]target = [[35, 92, 33], [68, 63, 97], [111, 133, 120], [50, 42, 29], [55, 59, 60], [22, 10, 15], [48, 46, 64], [19, 14, 34]]
+Epoch 0: |                                                                                                                                                     | 1499/? [1:26:32<00:00,  0.29it/s, v_num=seh9]train step 1500; scene = [['35a4da1b96cb6e60'], ['a5bcdef3441b37cd']]; loss = 0.014048
+Epoch 0: |                                                                                                                                                     | 1500/? [1:26:35<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1500; scene = ['45592a7f307bccd0'];
+target intrinsic:  tensor(0.8508, device='cuda:0') tensor(0.8510, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8800, device='cuda:0') tensor(0.8808, device='cuda:0')
+[2026-02-25 03:07:58,736][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1500/? [1:26:48<00:00,  0.29it/s, v_num=seh9]context = [[121, 160, 183], [50, 62, 95], [5, 48, 57], [27, 29, 90], [23, 107, 111], [10, 25, 55], [1, 50, 89], [17, 60, 67]]target = [[128, 139, 182], [60, 93, 89], [15, 25, 17], [79, 84, 54], [80, 51, 60], [12, 13, 28], [80, 51, 57], [47, 36, 52]]
+Epoch 0: |                                                                                                                                                     | 1509/? [1:27:18<00:00,  0.29it/s, v_num=seh9]train step 1510; scene = [['7415ffa5c457e330'], ['abdfc86976ee11de'], ['3e30d1a14cad58d5'], ['a7b928c84f551702'], ['9de970f9a14770a9'], ['4558408811e71246']]; loss = 0.021795
+Epoch 0: |                                                                                                                                                     | 1510/? [1:27:22<00:00,  0.29it/s, v_num=seh9]context = [[3, 6, 11, 16, 24, 30, 37, 40, 41, 51, 55, 58, 61, 66, 71, 73, 77, 79, 86, 87, 95, 97, 98, 100]]target = [[31, 62, 24, 19, 4, 88, 98, 60, 6, 55, 52, 82, 36, 74, 69, 46, 37, 22, 23, 58, 87, 8, 38, 89]]
+Epoch 0: |                                                                                                                                                     | 1519/? [1:27:54<00:00,  0.29it/s, v_num=seh9]train step 1520; scene = [['17bfe61015a33214'], ['aecaf8906a0a9bb1']]; loss = 0.016057
+Epoch 0: |                                                                                                                                                     | 1520/? [1:27:57<00:00,  0.29it/s, v_num=seh9]context = [[3, 8, 9, 10, 13, 21, 28, 30, 35, 38, 42, 44, 46, 52, 57, 62, 63, 75, 84, 85, 90, 93, 94, 100]]target = [[48, 12, 4, 84, 73, 17, 71, 22, 96, 43, 94, 40, 27, 39, 66, 63, 59, 50, 26, 56, 37, 76, 68, 64]]
+Epoch 0: |                                                                                                                                                     | 1529/? [1:28:28<00:00,  0.29it/s, v_num=seh9]train step 1530; scene = [['aa790649db11f096']]; loss = 0.019485
+Epoch 0: |                                                                                                                                                     | 1530/? [1:28:31<00:00,  0.29it/s, v_num=seh9]context = [[166, 172, 173, 176, 179, 184, 196, 211], [30, 31, 35, 40, 51, 67, 89, 117], [11, 18, 21, 35, 41, 44, 70, 74]]target = [[180, 174, 203, 206, 179, 177, 192, 197], [55, 85, 98, 79, 62, 53, 108, 70], [12, 22, 56, 32, 65, 17, 21, 49]]
+Epoch 0: |                                                                                                                                                     | 1539/? [1:29:02<00:00,  0.29it/s, v_num=seh9]train step 1540; scene = [['1661797137c6af53']]; loss = 0.024900
+Epoch 0: |                                                                                                                                                     | 1540/? [1:29:06<00:00,  0.29it/s, v_num=seh9]context = [[144, 209, 210], [50, 103, 119], [99, 111, 146], [163, 198, 217], [51, 65, 124], [8, 58, 95], [44, 97, 113], [2, 44, 88]]target = [[187, 173, 188], [83, 59, 103], [145, 117, 108], [201, 207, 194], [121, 53, 112], [24, 23, 53], [78, 69, 83], [38, 29, 82]]
+Epoch 0: |                                                                                                                                                     | 1549/? [1:29:37<00:00,  0.29it/s, v_num=seh9]train step 1550; scene = [['db24baad7f78f763']]; loss = 0.010661
+Epoch 0: |                                                                                                                                                     | 1550/? [1:29:41<00:00,  0.29it/s, v_num=seh9]context = [[109, 112, 117, 133, 138, 147, 154, 156], [9, 12, 23, 58, 60, 62, 88, 94], [0, 3, 5, 13, 16, 22, 48, 59]]target = [[145, 149, 124, 121, 152, 130, 122, 153], [74, 49, 44, 28, 36, 68, 19, 26], [26, 53, 14, 17, 13, 31, 55, 29]]
+Epoch 0: |                                                                                                                                                     | 1559/? [1:30:11<00:00,  0.29it/s, v_num=seh9]train step 1560; scene = [['8d89803ec0446d0e'], ['e86d5510c0fb2a93'], ['6549544819976e02'], ['a7bc2227bd3a3006'], ['a6e57377594a7a40'], ['4c76d28a44171812']]; loss = 0.034922
+Epoch 0: |                                                                                                                                                     | 1560/? [1:30:15<00:00,  0.29it/s, v_num=seh9]context = [[14, 24, 28, 30, 34, 38, 46, 48, 51, 52, 55, 57, 64, 71, 73, 75, 81, 82, 83, 89, 98, 107, 108, 111]]target = [[107, 43, 55, 51, 29, 22, 64, 58, 105, 85, 56, 28, 65, 32, 36, 53, 102, 21, 78, 80, 39, 76, 74, 87]]
+Epoch 0: |                                                                                                                                                     | 1569/? [1:30:47<00:00,  0.29it/s, v_num=seh9]train step 1570; scene = [['df29eae05ee5e69f'], ['af8f48c2701a86d7'], ['46f7affec5786351']]; loss = 0.015198
+Epoch 0: |                                                                                                                                                     | 1570/? [1:30:50<00:00,  0.29it/s, v_num=seh9]context = [[2, 8, 10, 19, 23, 28, 29, 32, 34, 42, 45, 46, 48, 53, 54, 56, 58, 64, 66, 69, 74, 80, 90, 99]]target = [[53, 5, 17, 87, 13, 86, 54, 47, 65, 90, 62, 55, 57, 60, 51, 35, 66, 81, 93, 4, 25, 9, 74, 31]]
+Epoch 0: |                                                                                                                                                     | 1579/? [1:31:22<00:00,  0.29it/s, v_num=seh9]train step 1580; scene = [['09fef1964d67a64e']]; loss = 0.013293
+Epoch 0: |                                                                                                                                                     | 1580/? [1:31:26<00:00,  0.29it/s, v_num=seh9]context = [[24, 32, 64, 72], [45, 88, 92, 104], [15, 29, 51, 89], [29, 63, 69, 82], [14, 59, 81, 103], [10, 28, 48, 64]]target = [[48, 33, 32, 64], [75, 85, 98, 51], [43, 72, 58, 69], [67, 32, 69, 63], [83, 95, 61, 39], [36, 54, 12, 58]]
+Epoch 0: |                                                                                                                                                     | 1589/? [1:31:57<00:00,  0.29it/s, v_num=seh9]train step 1590; scene = [['d3917d0a1eda2a1f'], ['8df7ef966383d5ba']]; loss = 0.017259
+Epoch 0: |                                                                                                                                                     | 1590/? [1:32:01<00:00,  0.29it/s, v_num=seh9]context = [[20, 23, 36, 43, 44, 51, 68, 70, 71, 74, 77, 84], [6, 8, 19, 38, 39, 48, 55, 56, 63, 64, 70, 71]]target = [[71, 26, 69, 38, 28, 44, 64, 56, 43, 50, 51, 61], [18, 15, 51, 36, 8, 25, 70, 23, 30, 63, 31, 55]]
+Epoch 0: |                                                                                                                                                     | 1599/? [1:32:32<00:00,  0.29it/s, v_num=seh9]train step 1600; scene = [['8bb4464a97a1190f'], ['72d6108a977fb29c'], ['f2cc31144c64dc6a']]; loss = 0.013397
+Epoch 0: |                                                                                                                                                     | 1600/? [1:32:35<00:00,  0.29it/s, v_num=seh9]context = [[29, 37, 44, 66, 75, 86], [101, 117, 120, 140, 161, 169], [38, 53, 65, 70, 73, 85], [55, 67, 71, 76, 95, 102]]target = [[39, 62, 68, 36, 69, 34], [118, 115, 120, 157, 107, 102], [76, 39, 45, 64, 50, 68], [56, 78, 60, 57, 59, 95]]
+[2026-02-25 03:13:49,838][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1609/? [1:33:07<00:00,  0.29it/s, v_num=seh9]train step 1610; scene = [['93598ff73741eb47'], ['9c3f970e0901e956'], ['a005c2781ac3e795']]; loss = 0.010612
+Epoch 0: |                                                                                                                                                     | 1610/? [1:33:10<00:00,  0.29it/s, v_num=seh9]context = [[22, 23, 31, 33, 35, 40, 41, 43, 44, 53, 62, 63, 67, 72, 84, 94, 95, 97, 102, 106, 109, 113, 116, 119]]target = [[92, 105, 54, 104, 40, 89, 87, 102, 96, 57, 99, 72, 48, 109, 55, 47, 75, 33, 91, 39, 66, 29, 23, 56]]
+Epoch 0: |                                                                                                                                                     | 1619/? [1:33:42<00:00,  0.29it/s, v_num=seh9]train step 1620; scene = [['81426f2270c3b08e'], ['f7d839262e5cc647']]; loss = 0.073225
+Epoch 0: |                                                                                                                                                     | 1620/? [1:33:45<00:00,  0.29it/s, v_num=seh9]context = [[52, 54, 57, 63, 70, 72, 81, 83, 86, 99, 103, 106], [19, 33, 38, 47, 48, 51, 53, 57, 58, 67, 68, 73]]target = [[64, 79, 60, 103, 102, 88, 70, 97, 56, 95, 82, 86], [72, 61, 54, 32, 44, 50, 69, 56, 37, 70, 28, 68]]
+Epoch 0: |                                                                                                                                                     | 1629/? [1:34:16<00:00,  0.29it/s, v_num=seh9]train step 1630; scene = [['4fa0b5f07c785cd6'], ['a3c9ffcea941f481']]; loss = 0.018238
+Epoch 0: |                                                                                                                                                     | 1630/? [1:34:20<00:00,  0.29it/s, v_num=seh9]context = [[91, 102, 107, 122, 123, 129, 132, 145], [2, 16, 17, 21, 53, 54, 57, 58], [0, 15, 21, 23, 26, 45, 53, 57]]target = [[119, 141, 113, 107, 127, 132, 133, 110], [38, 32, 45, 46, 12, 51, 52, 49], [37, 6, 43, 20, 28, 36, 29, 15]]
+Epoch 0: |                                                                                                                                                     | 1639/? [1:34:49<00:00,  0.29it/s, v_num=seh9]train step 1640; scene = [['a802441c288b598f'], ['8c8616ef6d449859'], ['51386d82afab8d56'], ['a5c1c97ecddf796a'], ['c26c2e3257c2d464'], ['ca35b3630b4f28ec'], ['6c10df89a6e8c8c1'], ['1a5022549d2718ae']]; loss = 0.022597
+Epoch 0: |                                                                                                                                                     | 1640/? [1:34:52<00:00,  0.29it/s, v_num=seh9]context = [[97, 98, 101, 102, 108, 115, 128, 139, 142, 143, 144, 156], [3, 5, 6, 7, 13, 20, 33, 45, 48, 49, 68, 77]]target = [[149, 98, 122, 102, 128, 120, 105, 138, 147, 110, 101, 117], [66, 25, 54, 52, 47, 38, 55, 27, 69, 59, 44, 48]]
+Epoch 0: |                                                                                                                                                     | 1649/? [1:35:23<00:00,  0.29it/s, v_num=seh9]train step 1650; scene = [['193e4b6a051b246b']]; loss = 0.008263
+Epoch 0: |                                                                                                                                                     | 1650/? [1:35:27<00:00,  0.29it/s, v_num=seh9]context = [[83, 89, 99, 100, 105, 107, 108, 118, 119, 120, 129, 134, 137, 138, 139, 140, 144, 145, 163, 164, 168, 174, 179, 180]]target = [[141, 110, 134, 98, 164, 172, 152, 162, 135, 136, 87, 178, 114, 123, 103, 95, 115, 158, 175, 119, 153, 94, 86, 166]]
+Epoch 0: |                                                                                                                                                     | 1659/? [1:35:58<00:00,  0.29it/s, v_num=seh9]train step 1660; scene = [['891648f976ef8493'], ['635af1b06fef3489'], ['f682d1c5b8431f46']]; loss = 0.024243
+Epoch 0: |                                                                                                                                                     | 1660/? [1:36:02<00:00,  0.29it/s, v_num=seh9]context = [[144, 176, 199, 202], [27, 50, 82, 102], [93, 128, 147, 166], [11, 15, 17, 76], [22, 32, 73, 77], [10, 27, 45, 82]]target = [[147, 178, 157, 197], [78, 79, 88, 80], [137, 159, 98, 119], [14, 22, 32, 43], [76, 32, 33, 27], [56, 60, 26, 13]]
+Epoch 0: |                                                                                                                                                     | 1669/? [1:36:34<00:00,  0.29it/s, v_num=seh9]train step 1670; scene = [['e7222b23037f83f7']]; loss = 0.015136
+Epoch 0: |                                                                                                                                                     | 1670/? [1:36:37<00:00,  0.29it/s, v_num=seh9]context = [[14, 16, 22, 26, 32, 57, 62, 64], [29, 33, 56, 59, 65, 69, 73, 76], [24, 29, 31, 51, 52, 69, 71, 76]]target = [[51, 47, 56, 55, 63, 50, 46, 21], [59, 65, 69, 45, 32, 52, 51, 33], [40, 51, 38, 67, 43, 61, 44, 69]]
+Epoch 0: |                                                                                                                                                     | 1679/? [1:37:09<00:00,  0.29it/s, v_num=seh9]train step 1680; scene = [['9e80f805fd6c7b93'], ['4f37cf41bcd303b6'], ['7cbf9883cb0c72d8'], ['bee292458a6d4570'], ['0ced4934341f8375'], ['b6fd5085cd94fb12'], ['c513a3c2f59aa548'], ['6cfdd5147e9d0d35'], ['60ac2ed64f533db5'], ['5edda72d021a1bc3'], ['ad127c962ce28e63'], ['2a2e3da0444d2ecd']]; loss = 0.020726
+Epoch 0: |                                                                                                                                                     | 1680/? [1:37:12<00:00,  0.29it/s, v_num=seh9]context = [[32, 35, 43, 45, 47, 59, 64, 67, 70, 74, 79, 85, 87, 89, 90, 102, 103, 110, 112, 116, 120, 125, 128, 129]]target = [[68, 125, 115, 78, 113, 48, 34, 58, 109, 85, 81, 72, 121, 77, 82, 98, 64, 111, 63, 65, 110, 37, 60, 52]]
+Epoch 0: |                                                                                                                                                     | 1689/? [1:37:44<00:00,  0.29it/s, v_num=seh9]train step 1690; scene = [['c537c888e9db11fb'], ['b008df0261874734'], ['63341a860ea3a43a'], ['276825499ea5dd5c'], ['de34678ce0e6b0bf'], ['2b026a3e98536fb5'], ['9d8eb52936b59070'], ['03acfb465cc3a17d']]; loss = 0.016417
+Epoch 0: |                                                                                                                                                     | 1690/? [1:37:47<00:00,  0.29it/s, v_num=seh9]context = [[50, 64, 68, 70, 78, 85, 91, 99, 101, 103, 115, 118], [178, 183, 184, 189, 190, 208, 210, 216, 221, 223, 240, 241]]target = [[101, 115, 91, 81, 57, 104, 90, 74, 103, 72, 82, 71], [180, 188, 235, 224, 240, 216, 206, 192, 237, 230, 214, 200]]
+Epoch 0: |                                                                                                                                                     | 1699/? [1:38:19<00:00,  0.29it/s, v_num=seh9]train step 1700; scene = [['f7e2aedf86e18a26']]; loss = 0.005774
+Epoch 0: |                                                                                                                                                     | 1700/? [1:38:22<00:00,  0.29it/s, v_num=seh9]context = [[37, 38, 42, 75, 80, 86, 100, 117], [127, 141, 150, 155, 156, 161, 169, 172], [0, 39, 45, 51, 65, 72, 73, 75]]target = [[40, 84, 114, 96, 57, 43, 38, 56], [131, 151, 128, 149, 148, 153, 132, 136], [24, 30, 46, 28, 25, 73, 31, 49]]
+Epoch 0: |                                                                                                                                                     | 1709/? [1:38:54<00:00,  0.29it/s, v_num=seh9]train step 1710; scene = [['d23860636bd4f6b9'], ['b0c6597c77c51a8c'], ['276c268acac9eca8'], ['2702020b31c28210'], ['ddc9f94d295eec5d'], ['d52a9764036894ac'], ['94034faf0ea1937a'], ['488f101dac781951']]; loss = 0.069144
+Epoch 0: |                                                                                                                                                     | 1710/? [1:38:58<00:00,  0.29it/s, v_num=seh9]context = [[89, 105, 115, 118, 121, 124, 125, 128, 129, 136, 157, 168], [32, 35, 38, 54, 56, 65, 68, 73, 86, 94, 98, 101]]target = [[167, 124, 131, 121, 102, 94, 93, 114, 90, 122, 106, 145], [44, 56, 57, 94, 33, 40, 68, 70, 90, 91, 82, 43]]
+Epoch 0: |                                                                                                                                                     | 1719/? [1:39:28<00:00,  0.29it/s, v_num=seh9]train step 1720; scene = [['af71d7a4edf61c31']]; loss = 0.012235
+Epoch 0: |                                                                                                                                                     | 1720/? [1:39:32<00:00,  0.29it/s, v_num=seh9]context = [[197, 215, 217, 233, 243, 247, 267, 274], [0, 7, 30, 47, 55, 69, 78, 89], [130, 136, 157, 165, 181, 182, 192, 201]]target = [[249, 263, 267, 222, 214, 208, 204, 256], [87, 47, 55, 24, 71, 58, 6, 35], [191, 143, 199, 172, 156, 139, 148, 176]]
+Epoch 0: |                                                                                                                                                     | 1729/? [1:40:03<00:00,  0.29it/s, v_num=seh9]train step 1730; scene = [['59ae8e2d2504d760']]; loss = 0.012658
+Epoch 0: |                                                                                                                                                     | 1730/? [1:40:07<00:00,  0.29it/s, v_num=seh9]context = [[12, 20, 26, 27, 35, 53, 55, 59, 60, 62, 66, 67], [12, 14, 18, 26, 31, 36, 41, 42, 48, 55, 63, 65]]target = [[24, 19, 58, 35, 48, 41, 63, 56, 62, 29, 28, 42], [42, 35, 18, 13, 39, 38, 20, 56, 52, 64, 19, 53]]
+Epoch 0: |                                                                                                                                                     | 1739/? [1:40:37<00:00,  0.29it/s, v_num=seh9]train step 1740; scene = [['af4565fb713ed79f'], ['473a917e0ee91a80']]; loss = 0.007735
+Epoch 0: |                                                                                                                                                     | 1740/? [1:40:40<00:00,  0.29it/s, v_num=seh9]context = [[20, 41, 54, 59, 61, 68], [37, 42, 48, 54, 61, 85], [116, 118, 130, 160, 161, 166], [37, 40, 68, 91, 92, 124]]target = [[31, 61, 50, 66, 46, 43], [68, 39, 48, 47, 40, 49], [157, 154, 128, 160, 141, 158], [60, 70, 98, 82, 39, 87]]
+Epoch 0: |                                                                                                                                                     | 1749/? [1:41:12<00:00,  0.29it/s, v_num=seh9]train step 1750; scene = [['bbaa2254cc65ceac'], ['8677be0fd77ef664']]; loss = 0.008902
+Epoch 0: |                                                                                                                                                     | 1750/? [1:41:14<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 1750; scene = ['3b273cb40c55db95'];
+target intrinsic:  tensor(1.0504, device='cuda:0') tensor(1.0506, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(1.0675, device='cuda:0') tensor(1.0597, device='cuda:0')
+[2026-02-25 03:22:26,000][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1750/? [1:41:16<00:00,  0.29it/s, v_num=seh9]context = [[5, 7, 12, 14, 17, 24, 29, 38, 49, 51, 57, 60], [28, 34, 39, 42, 46, 47, 50, 54, 56, 58, 71, 78]]target = [[10, 54, 15, 44, 52, 56, 6, 20, 35, 27, 30, 46], [55, 44, 50, 53, 49, 54, 62, 46, 35, 48, 61, 70]]
+Epoch 0: |                                                                                                                                                     | 1759/? [1:41:46<00:00,  0.29it/s, v_num=seh9]train step 1760; scene = [['380aa8c6d47a59a7'], ['1643b6a11a819f37'], ['c2162040f6db4d2f']]; loss = 0.011306
+Epoch 0: |                                                                                                                                                     | 1760/? [1:41:49<00:00,  0.29it/s, v_num=seh9]context = [[99, 104, 109, 112, 114, 119, 122, 130, 133, 137, 152, 161], [33, 40, 56, 59, 63, 67, 76, 84, 91, 94, 96, 97]]target = [[147, 120, 135, 124, 155, 157, 150, 138, 104, 142, 159, 109], [68, 35, 65, 51, 58, 93, 89, 49, 94, 64, 63, 72]]
+Epoch 0: |                                                                                                                                                     | 1769/? [1:42:20<00:00,  0.29it/s, v_num=seh9]train step 1770; scene = [['f63d2df8871ce70c'], ['066424b7f19bf6e5'], ['cc966a9b2af2232f']]; loss = 0.011400
+Epoch 0: |                                                                                                                                                     | 1770/? [1:42:23<00:00,  0.29it/s, v_num=seh9]context = [[7, 31, 84, 91], [5, 23, 52, 53], [45, 63, 89, 104], [70, 73, 125, 139], [7, 25, 48, 52], [95, 106, 138, 142]]target = [[36, 62, 87, 73], [16, 40, 27, 47], [64, 65, 74, 68], [98, 123, 97, 137], [19, 14, 33, 50], [125, 120, 102, 108]]
+Epoch 0: |                                                                                                                                                     | 1779/? [1:42:55<00:00,  0.29it/s, v_num=seh9]train step 1780; scene = [['c8f7ce4900a24c98'], ['966391dedc9b04e1']]; loss = 0.008602
+Epoch 0: |                                                                                                                                                     | 1780/? [1:42:58<00:00,  0.29it/s, v_num=seh9]context = [[137, 146, 150, 157, 165, 169, 173, 179, 183, 189, 194, 199], [12, 14, 17, 25, 29, 30, 33, 36, 45, 66, 77, 93]]target = [[197, 142, 143, 186, 170, 146, 147, 196, 185, 181, 179, 154], [23, 49, 46, 72, 31, 71, 17, 26, 92, 50, 55, 56]]
+Epoch 0: |                                                                                                                                                     | 1789/? [1:43:29<00:00,  0.29it/s, v_num=seh9]train step 1790; scene = [['7e934c4d5fc1df83'], ['48d1d822c47b5d0c'], ['afe6b05d0554a880'], ['32b6c71b6f77de55']]; loss = 0.021911
+Epoch 0: |                                                                                                                                                     | 1790/? [1:43:33<00:00,  0.29it/s, v_num=seh9]context = [[0, 1, 2, 3, 4, 9, 11, 26, 28, 32, 33, 36, 39, 41, 42, 47, 52, 54, 55, 67, 71, 87, 93, 97]]target = [[40, 57, 2, 6, 92, 63, 4, 13, 65, 71, 84, 66, 38, 16, 78, 55, 10, 79, 26, 34, 74, 46, 51, 14]]
+Epoch 0: |                                                                                                                                                     | 1799/? [1:44:05<00:00,  0.29it/s, v_num=seh9]train step 1800; scene = [['7e7eef981c97198b'], ['dc98908eb1cd3761'], ['2ff53fd35bc18dca']]; loss = 0.016925
+Epoch 0: |                                                                                                                                                     | 1800/? [1:44:08<00:00,  0.29it/s, v_num=seh9]context = [[17, 18, 24, 26, 35, 46, 49, 50, 60, 62, 69, 70, 72, 74, 76, 84, 85, 86, 89, 90, 104, 105, 110, 114]]target = [[97, 77, 86, 73, 24, 100, 21, 52, 78, 54, 30, 72, 41, 67, 84, 74, 94, 20, 82, 46, 65, 35, 61, 37]]
+[2026-02-25 03:25:22,744][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 1809/? [1:44:40<00:00,  0.29it/s, v_num=seh9]train step 1810; scene = [['cd7187c5ac3ae71e'], ['8b9ad2078e503344'], ['93416bc86e84f4c5']]; loss = 0.023126
+Epoch 0: |                                                                                                                                                     | 1810/? [1:44:44<00:00,  0.29it/s, v_num=seh9]context = [[27, 28, 29, 30, 34, 39, 43, 47, 49, 55, 78, 81, 92, 94, 96, 99, 103, 104, 106, 107, 109, 111, 117, 124]]target = [[99, 84, 100, 68, 103, 70, 49, 36, 120, 65, 78, 92, 46, 58, 80, 74, 121, 57, 95, 39, 28, 59, 67, 62]]
+Epoch 0: |                                                                                                                                                     | 1819/? [1:45:15<00:00,  0.29it/s, v_num=seh9]train step 1820; scene = [['54e0c39079f7a30b']]; loss = 0.010378
+Epoch 0: |                                                                                                                                                     | 1820/? [1:45:18<00:00,  0.29it/s, v_num=seh9]context = [[47, 53, 56, 59, 61, 65, 68, 74, 76, 81, 86, 91, 92, 93, 94, 98, 106, 114, 123, 128, 129, 137, 140, 144]]target = [[120, 49, 106, 72, 133, 71, 48, 113, 138, 122, 116, 66, 134, 110, 103, 89, 76, 84, 69, 115, 61, 112, 123, 136]]
+Epoch 0: |                                                                                                                                                     | 1829/? [1:45:49<00:00,  0.29it/s, v_num=seh9]train step 1830; scene = [['1a0b5e0311249585'], ['f5d05305398238d4']]; loss = 0.027166
+Epoch 0: |                                                                                                                                                     | 1830/? [1:45:52<00:00,  0.29it/s, v_num=seh9]context = [[97, 106, 108, 114, 142, 151], [3, 15, 16, 51, 63, 70], [21, 29, 45, 46, 68, 75], [181, 182, 194, 246, 265, 268]]target = [[112, 127, 141, 134, 133, 124], [51, 26, 50, 60, 59, 30], [68, 73, 28, 22, 23, 67], [237, 206, 224, 188, 267, 187]]
+Epoch 0: |                                                                                                                                                     | 1839/? [1:46:23<00:00,  0.29it/s, v_num=seh9]train step 1840; scene = [['f30398f000a252c0'], ['2c1b49fe037d6723'], ['4a0f95a3db913b56'], ['e0e41af517c97c32'], ['8d758914077e5926'], ['7887a8b69eef3f6c'], ['11c69796af56d901'], ['9cdbe57a44e11638'], ['f98d6c32648cdc8b'], ['81f8c3d21285d1cc'], ['7391f166ec0c4169'], ['da1f9f2859b59142']]; loss = 0.029092
+Epoch 0: |                                                                                                                                                     | 1840/? [1:46:27<00:00,  0.29it/s, v_num=seh9]context = [[36, 49, 51, 53, 69, 74, 75, 80, 90, 94, 101, 113], [68, 78, 86, 87, 89, 90, 95, 113, 127, 133, 136, 143]]target = [[64, 81, 86, 101, 67, 50, 62, 37, 107, 73, 89, 71], [101, 81, 100, 114, 86, 103, 131, 72, 113, 95, 137, 102]]
+Epoch 0: |                                                                                                                                                     | 1849/? [1:46:58<00:00,  0.29it/s, v_num=seh9]train step 1850; scene = [['509210a9f0f87e8f']]; loss = 0.008247
+Epoch 0: |                                                                                                                                                     | 1850/? [1:47:02<00:00,  0.29it/s, v_num=seh9]context = [[54, 90, 135], [1, 13, 72], [13, 62, 103], [45, 77, 94], [93, 117, 167], [30, 42, 104], [44, 89, 103], [1, 3, 55]]target = [[65, 105, 131], [46, 9, 41], [102, 34, 68], [65, 91, 46], [132, 134, 125], [87, 81, 45], [65, 50, 54], [54, 13, 36]]
+Epoch 0: |                                                                                                                                                     | 1859/? [1:47:33<00:00,  0.29it/s, v_num=seh9]train step 1860; scene = [['4486fd0a58b1b39b']]; loss = 0.011390
+Epoch 0: |                                                                                                                                                     | 1860/? [1:47:37<00:00,  0.29it/s, v_num=seh9]context = [[4, 8, 15, 23, 29, 38, 46, 50], [140, 156, 174, 199, 200, 209, 215, 220], [2, 30, 36, 42, 45, 49, 62, 65]]target = [[30, 17, 16, 23, 6, 14, 37, 25], [153, 175, 149, 211, 210, 151, 148, 157], [55, 33, 11, 3, 59, 42, 25, 7]]
+Epoch 0: |                                                                                                                                                     | 1869/? [1:48:08<00:00,  0.29it/s, v_num=seh9]train step 1870; scene = [['ae9963ff8cd6c4de']]; loss = 0.012821
+Epoch 0: |                                                                                                                                                     | 1870/? [1:48:11<00:00,  0.29it/s, v_num=seh9]context = [[133, 135, 136, 144, 163, 166, 167, 172, 175, 176, 181, 189, 191, 192, 194, 195, 197, 201, 202, 204, 216, 224, 227, 230]]target = [[171, 154, 201, 193, 166, 210, 167, 150, 222, 136, 155, 159, 164, 148, 208, 188, 137, 221, 181, 216, 152, 160, 161, 226]]
+Epoch 0: |                                                                                                                                                     | 1879/? [1:48:42<00:00,  0.29it/s, v_num=seh9]train step 1880; scene = [['97eab841f52c3532']]; loss = 0.008098
+Epoch 0: |                                                                                                                                                     | 1880/? [1:48:46<00:00,  0.29it/s, v_num=seh9]context = [[9, 19, 24, 30, 37, 41, 45, 50, 51, 54, 55, 61, 65, 67, 69, 72, 75, 77, 78, 81, 87, 90, 96, 106]]target = [[95, 39, 90, 41, 67, 94, 89, 84, 66, 81, 29, 45, 40, 104, 20, 34, 74, 26, 17, 103, 102, 33, 68, 22]]
+Epoch 0: |                                                                                                                                                     | 1889/? [1:49:17<00:00,  0.29it/s, v_num=seh9]train step 1890; scene = [['c58d526cb2894a17'], ['d3a0a89d951a6101'], ['e8e48c278d3f4624']]; loss = 0.024537
+Epoch 0: |                                                                                                                                                     | 1890/? [1:49:21<00:00,  0.29it/s, v_num=seh9]context = [[79, 80, 116, 122, 127, 128, 129, 131, 139, 141, 144, 151], [12, 13, 21, 24, 26, 32, 44, 45, 53, 62, 68, 71]]target = [[94, 147, 101, 142, 150, 136, 104, 80, 140, 125, 144, 96], [31, 26, 46, 41, 39, 18, 53, 24, 23, 13, 63, 21]]
+Epoch 0: |                                                                                                                                                     | 1899/? [1:49:52<00:00,  0.29it/s, v_num=seh9]train step 1900; scene = [['7a03c6c69a883bac'], ['9467c3431e0e586c'], ['eb91b53e213504b8'], ['9e496cf3ceac8708'], ['6b01b60bd401ef3e'], ['db6ce23f93a9a0d3'], ['252ab646bcd51b88'], ['752724fcd4a2061f']]; loss = 0.025752
+Epoch 0: |                                                                                                                                                     | 1900/? [1:49:56<00:00,  0.29it/s, v_num=seh9]context = [[72, 78, 84, 85, 87, 92, 95, 99, 116, 118, 119, 123], [28, 29, 30, 32, 33, 35, 45, 62, 68, 76, 80, 100]]target = [[97, 77, 113, 99, 74, 92, 81, 90, 114, 80, 83, 119], [77, 85, 58, 45, 82, 49, 97, 80, 99, 93, 79, 53]]
+Epoch 0: |                                                                                                                                                     | 1909/? [1:50:27<00:00,  0.29it/s, v_num=seh9]train step 1910; scene = [['5f14e0e70310b263'], ['5c9877200c432073'], ['c0580f743d41ee56']]; loss = 0.008825
+Epoch 0: |                                                                                                                                                     | 1910/? [1:50:30<00:00,  0.29it/s, v_num=seh9]context = [[6, 8, 9, 11, 12, 13, 26, 32, 34, 46, 59, 67], [67, 70, 83, 85, 90, 94, 118, 122, 132, 133, 142, 146]]target = [[26, 13, 54, 52, 35, 55, 19, 56, 53, 59, 30, 36], [116, 87, 82, 79, 104, 89, 76, 126, 128, 131, 123, 68]]
+Epoch 0: |                                                                                                                                                     | 1919/? [1:51:02<00:00,  0.29it/s, v_num=seh9]train step 1920; scene = [['242ee2c8b1208167']]; loss = 0.010234
+Epoch 0: |                                                                                                                                                     | 1920/? [1:51:05<00:00,  0.29it/s, v_num=seh9]context = [[20, 30, 33, 40, 44, 45, 47, 57, 67, 70, 77, 83, 84, 88, 93, 95, 98, 102, 103, 108, 111, 112, 116, 117]]target = [[46, 105, 95, 29, 114, 78, 35, 69, 77, 44, 112, 40, 60, 74, 21, 107, 87, 38, 104, 22, 43, 27, 52, 34]]
+Epoch 0: |                                                                                                                                                     | 1929/? [1:51:37<00:00,  0.29it/s, v_num=seh9]train step 1930; scene = [['4e8e3034d0aa307b'], ['da65394ba9286a06'], ['640dafc4b5a3d491'], ['b25a0f4ffca51d79'], ['534ddfd7e12e4188'], ['db19784431c29cae']]; loss = 0.027209
+Epoch 0: |                                                                                                                                                     | 1930/? [1:51:41<00:00,  0.29it/s, v_num=seh9]context = [[2, 5, 8, 23, 32, 47, 61, 62], [0, 20, 21, 29, 32, 33, 43, 50], [27, 33, 34, 37, 51, 62, 68, 109]]target = [[13, 49, 20, 39, 23, 4, 42, 36], [14, 6, 25, 39, 20, 44, 43, 47], [105, 37, 104, 72, 75, 55, 34, 64]]
+Epoch 0: |                                                                                                                                                     | 1939/? [1:52:12<00:00,  0.29it/s, v_num=seh9]train step 1940; scene = [['861260140edeed05']]; loss = 0.026124
+Epoch 0: |                                                                                                                                                     | 1940/? [1:52:16<00:00,  0.29it/s, v_num=seh9]context = [[3, 43, 46, 55, 57, 83], [3, 24, 25, 40, 47, 54], [0, 12, 31, 36, 70, 73], [18, 31, 40, 50, 64, 69]]target = [[74, 48, 79, 6, 51, 82], [9, 10, 33, 23, 21, 36], [8, 65, 61, 53, 46, 66], [23, 19, 40, 22, 55, 47]]
+Epoch 0: |                                                                                                                                                     | 1949/? [1:52:47<00:00,  0.29it/s, v_num=seh9]train step 1950; scene = [['d19b2171a8edd830'], ['677817c638b61cbe'], ['ae1ce26f27f55ee0'], ['a54c4e390105d712'], ['9a5201a21290371f'], ['39ed8d2efe760b94'], ['789a1b00d73a2072'], ['9ad88af7e2443b55'], ['bb8df0200c9fa103'], ['e6cd96ceddb9b665'], ['27474a9b81b5b915'], ['20b2bb20840251c6']]; loss = 0.029989
+Epoch 0: |                                                                                                                                                     | 1950/? [1:52:51<00:00,  0.29it/s, v_num=seh9]context = [[7, 23, 28, 35, 43, 62], [0, 3, 14, 15, 23, 78], [27, 33, 46, 47, 82, 101], [28, 41, 42, 80, 100, 103]]target = [[55, 17, 13, 47, 33, 28], [1, 74, 3, 5, 20, 7], [63, 74, 85, 73, 41, 90], [35, 36, 99, 78, 88, 57]]
+Epoch 0: |                                                                                                                                                     | 1959/? [1:53:22<00:00,  0.29it/s, v_num=seh9]train step 1960; scene = [['c176654607fd9cce']]; loss = 0.018165
+Epoch 0: |                                                                                                                                                     | 1960/? [1:53:25<00:00,  0.29it/s, v_num=seh9]context = [[2, 8, 22, 32, 42, 44, 45, 47, 55, 57, 64, 67], [2, 29, 34, 35, 38, 42, 46, 47, 51, 57, 64, 69]]target = [[8, 32, 23, 65, 53, 19, 59, 14, 42, 11, 40, 9], [37, 24, 10, 46, 49, 23, 3, 68, 55, 44, 14, 26]]
+Epoch 0: |                                                                                                                                                     | 1969/? [1:53:55<00:00,  0.29it/s, v_num=seh9]train step 1970; scene = [['48cfc82a1cc3ee30'], ['1c92d53dec160984'], ['68f7ee0def2ecdb3'], ['c6b3169c8747d1fd']]; loss = 0.010688
+Epoch 0: |                                                                                                                                                     | 1970/? [1:53:58<00:00,  0.29it/s, v_num=seh9]context = [[189, 224, 239], [50, 66, 96], [125, 146, 182], [43, 50, 126], [118, 128, 179], [26, 77, 87], [4, 42, 68], [56, 65, 108]]target = [[236, 195, 197], [94, 79, 78], [170, 167, 143], [120, 92, 49], [169, 142, 172], [62, 75, 80], [43, 35, 40], [98, 95, 70]]
+Epoch 0: |                                                                                                                                                     | 1979/? [1:54:30<00:00,  0.29it/s, v_num=seh9]train step 1980; scene = [['01cbdc2687833ad4'], ['ad9a0c2b517fc375'], ['b6d6a1ec65fa5263'], ['64199c57dfb3dc4e'], ['6d0232cfe9cc7adc'], ['e3b385876279e0fc'], ['56c780409feaf822'], ['c841844509970531']]; loss = 0.027574
+Epoch 0: |                                                                                                                                                     | 1980/? [1:54:33<00:00,  0.29it/s, v_num=seh9]context = [[84, 88, 91, 109, 117, 140, 144, 149], [162, 178, 182, 184, 189, 210, 235, 240], [7, 18, 23, 32, 42, 68, 75, 78]]target = [[134, 147, 122, 135, 121, 142, 132, 129], [210, 186, 227, 232, 218, 193, 195, 220], [56, 72, 60, 75, 15, 27, 18, 29]]
+Epoch 0: |                                                                                                                                                     | 1989/? [1:55:04<00:00,  0.29it/s, v_num=seh9]train step 1990; scene = [['c67ada3a05b5cac2']]; loss = 0.014514
+Epoch 0: |                                                                                                                                                     | 1990/? [1:55:07<00:00,  0.29it/s, v_num=seh9]context = [[22, 30, 31, 35, 52, 59, 65, 69], [19, 37, 47, 48, 53, 64, 87, 102], [8, 19, 20, 21, 27, 52, 53, 54]]target = [[47, 45, 37, 25, 67, 29, 54, 59], [61, 20, 72, 71, 53, 36, 98, 39], [19, 39, 11, 52, 44, 13, 16, 21]]
+Epoch 0: |                                                                                                                                                     | 1999/? [1:55:38<00:00,  0.29it/s, v_num=seh9]train step 2000; scene = [['af11d275959fa201'], ['770a3d6cc6eb482b'], ['e2f2a27bfce53270']]; loss = 0.009533
+Epoch 0: |                                                                                                                                                     | 2000/? [1:55:42<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 2000; scene = ['be75142d4652fe3e'];
+target intrinsic:  tensor(0.9402, device='cuda:0') tensor(0.9404, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8958, device='cuda:0') tensor(0.8937, device='cuda:0')
+[2026-02-25 03:36:53,441][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2000/? [1:55:43<00:00,  0.29it/s, v_num=seh9]context = [[9, 26, 82], [130, 132, 192], [35, 48, 110], [83, 145, 172], [27, 95, 113], [94, 101, 144], [23, 43, 75], [8, 57, 96]]target = [[51, 71, 22], [186, 183, 166], [71, 76, 77], [84, 95, 134], [49, 43, 91], [116, 113, 131], [32, 54, 45], [65, 25, 14]]
+[2026-02-25 03:36:56,656][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2009/? [1:56:13<00:00,  0.29it/s, v_num=seh9]train step 2010; scene = [['9f2fffeb41e0834f'], ['e718b9c40932334a'], ['8b9d2f4e94b006f6']]; loss = 0.012716
+Epoch 0: |                                                                                                                                                     | 2010/? [1:56:16<00:00,  0.29it/s, v_num=seh9]context = [[12, 38, 73, 93], [0, 7, 44, 46], [70, 79, 112, 160], [12, 48, 49, 81], [75, 78, 104, 122], [16, 67, 78, 83]]target = [[81, 53, 22, 16], [39, 32, 23, 13], [110, 107, 96, 155], [62, 26, 59, 17], [117, 100, 101, 82], [42, 34, 57, 29]]
+Epoch 0: |                                                                                                                                                     | 2019/? [1:56:48<00:00,  0.29it/s, v_num=seh9]train step 2020; scene = [['4b4f4736ef565c77'], ['0be2e01e27a683b0'], ['3c1aa6e11ad2bb0c'], ['90998c5f495333d2']]; loss = 0.013938
+Epoch 0: |                                                                                                                                                     | 2020/? [1:56:51<00:00,  0.29it/s, v_num=seh9]context = [[28, 29, 33, 46, 48, 49, 50, 56, 67, 71, 85, 86, 94, 98, 99, 104, 105, 109, 111, 116, 119, 120, 124, 125]]target = [[105, 112, 57, 45, 65, 70, 86, 82, 46, 73, 40, 94, 35, 121, 72, 120, 110, 118, 117, 56, 39, 90, 101, 98]]
+Epoch 0: |                                                                                                                                                     | 2029/? [1:57:22<00:00,  0.29it/s, v_num=seh9]train step 2030; scene = [['aa5f0f9be3bd3b2a'], ['7e8bc02377c148f9'], ['12a521cf027c35f6'], ['d8940a999b9fc9d8']]; loss = 0.014420
+Epoch 0: |                                                                                                                                                     | 2030/? [1:57:26<00:00,  0.29it/s, v_num=seh9]context = [[40, 44, 53, 57, 59, 67, 72, 76, 81, 83, 88, 91], [139, 140, 148, 154, 159, 162, 165, 171, 173, 179, 184, 193]]target = [[51, 68, 56, 70, 65, 55, 49, 48, 52, 67, 54, 69], [181, 180, 169, 185, 140, 161, 182, 192, 144, 143, 142, 148]]
+Epoch 0: |                                                                                                                                                     | 2039/? [1:57:57<00:00,  0.29it/s, v_num=seh9]train step 2040; scene = [['f6d34faab262bd11'], ['b3eb7d9489b53fa6'], ['3562ce0ddbf1d45e'], ['56b9fc4f422e8019'], ['245ec87f817cf1e4'], ['417bfbed2fe3e456'], ['5dedc6ab8fbe23d7'], ['480231b83d000bf7']]; loss = 0.015183
+Epoch 0: |                                                                                                                                                     | 2040/? [1:58:01<00:00,  0.29it/s, v_num=seh9]context = [[143, 153, 158, 159, 162, 163, 165, 167, 173, 185, 189, 198, 200, 205, 219, 221, 222, 223, 227, 230, 231, 235, 236, 240]]target = [[176, 193, 207, 188, 229, 218, 204, 236, 238, 239, 185, 152, 196, 197, 223, 150, 186, 172, 233, 162, 227, 170, 232, 160]]
+Epoch 0: |                                                                                                                                                     | 2049/? [1:58:30<00:00,  0.29it/s, v_num=seh9]train step 2050; scene = [['a660d0f19f982875'], ['09f2ad8e87f8f42f'], ['3530096ad3087c6c'], ['3c79b442b329338e']]; loss = 0.012646
+Epoch 0: |                                                                                                                                                     | 2050/? [1:58:34<00:00,  0.29it/s, v_num=seh9]context = [[47, 51, 53, 60, 61, 68, 70, 92, 98, 105, 107, 114], [3, 15, 16, 23, 27, 28, 29, 31, 38, 46, 58, 61]]target = [[108, 92, 82, 67, 68, 93, 103, 52, 107, 94, 72, 113], [52, 4, 33, 35, 54, 55, 11, 5, 18, 7, 42, 45]]
+Epoch 0: |                                                                                                                                                     | 2059/? [1:59:06<00:00,  0.29it/s, v_num=seh9]train step 2060; scene = [['a2e2c2e66db9a05c']]; loss = 0.010343
+Epoch 0: |                                                                                                                                                     | 2060/? [1:59:09<00:00,  0.29it/s, v_num=seh9]context = [[3, 31, 34, 41, 47, 57, 60, 62, 75, 76, 81, 89], [30, 40, 43, 44, 52, 69, 74, 80, 87, 89, 100, 102]]target = [[75, 37, 83, 16, 78, 52, 19, 67, 86, 57, 72, 69], [80, 42, 38, 49, 90, 50, 46, 63, 97, 70, 71, 94]]
+Epoch 0: |                                                                                                                                                     | 2069/? [1:59:39<00:00,  0.29it/s, v_num=seh9]train step 2070; scene = [['a16a0940ea913a28']]; loss = 0.009636
+Epoch 0: |                                                                                                                                                     | 2070/? [1:59:43<00:00,  0.29it/s, v_num=seh9]context = [[105, 108, 109, 111, 116, 117, 120, 122, 132, 133, 135, 139, 145, 147, 154, 155, 164, 170, 182, 184, 187, 188, 196, 202]]target = [[162, 107, 121, 131, 127, 173, 161, 134, 159, 199, 182, 133, 128, 146, 115, 189, 144, 168, 157, 178, 148, 152, 196, 117]]
+Epoch 0: |                                                                                                                                                     | 2079/? [2:00:14<00:00,  0.29it/s, v_num=seh9]train step 2080; scene = [['06ed257e33ae67f5']]; loss = 0.007093
+Epoch 0: |                                                                                                                                                     | 2080/? [2:00:18<00:00,  0.29it/s, v_num=seh9]context = [[99, 102, 114, 128, 136, 149, 150, 180], [59, 61, 86, 103, 114, 123, 125, 130], [0, 4, 7, 8, 14, 27, 36, 45]]target = [[158, 155, 114, 148, 120, 102, 178, 171], [121, 92, 93, 106, 61, 72, 91, 68], [8, 33, 39, 11, 22, 13, 38, 7]]
+Epoch 0: |                                                                                                                                                     | 2089/? [2:00:50<00:00,  0.29it/s, v_num=seh9]train step 2090; scene = [['ca0ef4335afc5644'], ['31ff05dbf772323c'], ['2f3a17c1963f8747'], ['ba57ad8ee0354f53'], ['222ef9ba2f4b2abb'], ['e98d838758c7b400'], ['c5536f755d325407'], ['204c24d2d8cc0ef9']]; loss = 0.024821
+Epoch 0: |                                                                                                                                                     | 2090/? [2:00:53<00:00,  0.29it/s, v_num=seh9]context = [[141, 178, 179, 181, 199, 203, 207, 214, 217, 218, 220, 221], [9, 14, 18, 27, 30, 33, 34, 40, 43, 52, 55, 61]]target = [[199, 202, 218, 214, 142, 197, 187, 180, 155, 144, 165, 154], [47, 21, 13, 57, 59, 52, 34, 23, 40, 31, 12, 49]]
+Epoch 0: |                                                                                                                                                     | 2099/? [2:01:25<00:00,  0.29it/s, v_num=seh9]train step 2100; scene = [['83222e191a9467b0']]; loss = 0.014521
+Epoch 0: |                                                                                                                                                     | 2100/? [2:01:28<00:00,  0.29it/s, v_num=seh9]context = [[3, 27, 56, 60], [11, 39, 51, 61], [13, 14, 58, 73], [33, 78, 94, 95], [29, 75, 81, 94], [65, 97, 98, 111]]target = [[43, 28, 40, 32], [23, 39, 45, 37], [70, 36, 20, 49], [89, 92, 76, 68], [85, 39, 55, 32], [88, 81, 104, 71]]
+Epoch 0: |                                                                                                                                                     | 2109/? [2:01:59<00:00,  0.29it/s, v_num=seh9]train step 2110; scene = [['1d4ffe48542eb14b'], ['851dd950cdbcea17'], ['235b7fe59778e1e0'], ['013ec74a4fde6737']]; loss = 0.016686
+Epoch 0: |                                                                                                                                                     | 2110/? [2:02:02<00:00,  0.29it/s, v_num=seh9]context = [[21, 30, 31, 33, 34, 35, 36, 37, 43, 45, 46, 48, 54, 55, 56, 70, 71, 76, 81, 90, 94, 100, 104, 118]]target = [[73, 52, 77, 71, 36, 69, 80, 108, 43, 81, 117, 40, 44, 115, 64, 74, 109, 32, 113, 79, 45, 85, 104, 70]]
+Epoch 0: |                                                                                                                                                     | 2119/? [2:02:34<00:00,  0.29it/s, v_num=seh9]train step 2120; scene = [['434bd83c81eae45d'], ['52e13717e8b9393d'], ['90e4f7b0d3147f4b']]; loss = 0.015578
+Epoch 0: |                                                                                                                                                     | 2120/? [2:02:37<00:00,  0.29it/s, v_num=seh9]context = [[197, 204, 207, 208, 226, 229, 232, 235, 247, 255, 266, 268], [7, 9, 31, 32, 41, 42, 46, 48, 50, 65, 77, 79]]target = [[258, 238, 222, 245, 227, 265, 246, 242, 253, 211, 229, 202], [72, 25, 51, 45, 63, 39, 68, 46, 67, 15, 21, 69]]
+Epoch 0: |                                                                                                                                                     | 2129/? [2:03:10<00:00,  0.29it/s, v_num=seh9]train step 2130; scene = [['50ad33ea0f572d25']]; loss = 0.006811
+Epoch 0: |                                                                                                                                                     | 2130/? [2:03:13<00:00,  0.29it/s, v_num=seh9]context = [[5, 30, 45, 61], [44, 45, 73, 106], [7, 27, 42, 85], [112, 122, 170, 196], [14, 29, 32, 67], [69, 71, 126, 129]]target = [[19, 47, 42, 58], [85, 65, 105, 72], [39, 71, 24, 13], [161, 172, 165, 126], [40, 18, 16, 34], [107, 109, 74, 84]]
+Epoch 0: |                                                                                                                                                     | 2139/? [2:03:44<00:00,  0.29it/s, v_num=seh9]train step 2140; scene = [['b4caf9dbe386c507'], ['7f441f51d44ecaf7'], ['db84572032e69dc0'], ['4a18f49c0e887115'], ['5799a001baf71ce5'], ['7430821f0f20c61a'], ['63232e078598c9c8'], ['f291b18d16c3b10c']]; loss = 0.021084
+Epoch 0: |                                                                                                                                                     | 2140/? [2:03:47<00:00,  0.29it/s, v_num=seh9]context = [[27, 34, 53, 54, 56, 63, 72, 84], [8, 9, 11, 15, 24, 34, 50, 61], [18, 38, 43, 45, 73, 79, 93, 101]]target = [[77, 63, 61, 46, 64, 35, 73, 31], [53, 25, 59, 13, 58, 34, 37, 56], [97, 22, 71, 90, 91, 84, 50, 72]]
+Epoch 0: |                                                                                                                                                     | 2149/? [2:04:18<00:00,  0.29it/s, v_num=seh9]train step 2150; scene = [['916b86f95631b480'], ['a2e4b97b2c9a0ae6'], ['ed22594386b563a1']]; loss = 0.009704
+Epoch 0: |                                                                                                                                                     | 2150/? [2:04:22<00:00,  0.29it/s, v_num=seh9]context = [[74, 91, 116, 149], [40, 50, 66, 115], [158, 196, 225, 245], [11, 33, 48, 65], [19, 40, 79, 98], [33, 63, 77, 80]]target = [[88, 108, 102, 117], [104, 113, 80, 96], [176, 227, 159, 162], [14, 31, 46, 63], [60, 96, 49, 70], [37, 64, 46, 48]]
+Epoch 0: |                                                                                                                                                     | 2159/? [2:04:53<00:00,  0.29it/s, v_num=seh9]train step 2160; scene = [['037b77f7cc0565e4'], ['648fc6db158f6e55'], ['465fa8314b741006']]; loss = 0.009971
+Epoch 0: |                                                                                                                                                     | 2160/? [2:04:57<00:00,  0.29it/s, v_num=seh9]context = [[48, 71, 110], [66, 105, 134], [16, 74, 75], [11, 51, 63], [23, 70, 83], [70, 79, 140], [83, 105, 130], [28, 38, 84]]target = [[89, 72, 55], [67, 99, 95], [36, 35, 17], [19, 26, 32], [45, 67, 72], [134, 91, 139], [113, 123, 112], [70, 62, 78]]
+Epoch 0: |                                                                                                                                                     | 2169/? [2:05:29<00:00,  0.29it/s, v_num=seh9]train step 2170; scene = [['a56e2c19e91b75a5'], ['d82aac6bf67b7d17'], ['d46599d6e4a2b451']]; loss = 0.010081
+Epoch 0: |                                                                                                                                                     | 2170/? [2:05:32<00:00,  0.29it/s, v_num=seh9]context = [[12, 15, 26, 27, 32, 40, 57, 64, 70, 71, 74, 78], [0, 8, 15, 21, 30, 41, 49, 50, 77, 81, 85, 88]]target = [[63, 74, 18, 46, 77, 35, 76, 68, 31, 32, 48, 65], [66, 18, 4, 85, 83, 31, 63, 24, 34, 19, 58, 23]]
+Epoch 0: |                                                                                                                                                     | 2179/? [2:06:02<00:00,  0.29it/s, v_num=seh9]train step 2180; scene = [['c31a1cb20a151f9f'], ['e45fd910f1e78bb9'], ['1241275438f193c1'], ['008ec1473e4ce029'], ['69937c2efa2c110b'], ['e0f2adb45811eca4'], ['1ceacc053b551f8d'], ['4be7b2dbfbb85134']]; loss = 0.017517
+Epoch 0: |                                                                                                                                                     | 2180/? [2:06:06<00:00,  0.29it/s, v_num=seh9]context = [[31, 39, 43, 48, 54, 56, 58, 61, 62, 66, 72, 73, 74, 75, 80, 81, 88, 93, 95, 101, 104, 115, 125, 128]]target = [[72, 114, 58, 78, 68, 125, 55, 53, 50, 77, 123, 61, 90, 115, 33, 83, 127, 102, 120, 57, 41, 98, 79, 95]]
+Epoch 0: |                                                                                                                                                     | 2189/? [2:06:37<00:00,  0.29it/s, v_num=seh9]train step 2190; scene = [['aff3adf77641ddcb'], ['58cdd230b4f5b590']]; loss = 0.014316
+Epoch 0: |                                                                                                                                                     | 2190/? [2:06:41<00:00,  0.29it/s, v_num=seh9]context = [[0, 6, 8, 12, 14, 18, 25, 28, 37, 41, 45, 46, 48, 53, 54, 58, 68, 74, 76, 85, 93, 94, 96, 97]]target = [[9, 35, 80, 72, 4, 81, 22, 86, 10, 64, 19, 93, 34, 12, 50, 82, 38, 3, 53, 66, 8, 76, 71, 90]]
+Epoch 0: |                                                                                                                                                     | 2199/? [2:07:12<00:00,  0.29it/s, v_num=seh9]train step 2200; scene = [['7d2402b99eed8e47']]; loss = 0.020502
+Epoch 0: |                                                                                                                                                     | 2200/? [2:07:16<00:00,  0.29it/s, v_num=seh9]context = [[106, 109, 113, 117, 130, 132, 142, 162, 164, 171, 175, 183], [0, 8, 10, 12, 19, 20, 22, 23, 30, 39, 58, 74]]target = [[121, 161, 131, 125, 182, 166, 158, 132, 113, 118, 139, 140], [12, 46, 34, 26, 72, 1, 27, 36, 13, 64, 17, 49]]
+[2026-02-25 03:48:29,358][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2209/? [2:07:47<00:00,  0.29it/s, v_num=seh9]train step 2210; scene = [['0296e319701adeab']]; loss = 0.006938
+Epoch 0: |                                                                                                                                                     | 2210/? [2:07:51<00:00,  0.29it/s, v_num=seh9]context = [[77, 79, 91, 107, 136, 157], [2, 15, 19, 26, 69, 78], [6, 9, 42, 47, 56, 60], [18, 59, 62, 68, 75, 86]]target = [[84, 112, 128, 103, 87, 96], [40, 38, 45, 3, 30, 43], [37, 25, 44, 13, 50, 12], [81, 51, 56, 31, 41, 21]]
+Epoch 0: |                                                                                                                                                     | 2219/? [2:08:23<00:00,  0.29it/s, v_num=seh9]train step 2220; scene = [['f0f9dc20febf63c6'], ['247393e72009dab0'], ['85b95906da927213'], ['8b80a0150e20f2f2'], ['069c4f8d7a3225b4'], ['e92f5355816a5aef'], ['73089f2cde9f9cd0'], ['8812c0981c6f30a8']]; loss = 0.014641
+Epoch 0: |                                                                                                                                                     | 2220/? [2:08:26<00:00,  0.29it/s, v_num=seh9]context = [[146, 166, 170, 186, 196, 205, 219, 233], [3, 25, 36, 44, 46, 48, 55, 63], [12, 20, 33, 55, 68, 74, 97, 98]]target = [[169, 183, 158, 191, 198, 205, 227, 171], [48, 33, 26, 39, 28, 21, 58, 61], [63, 83, 48, 97, 20, 80, 66, 85]]
+Epoch 0: |                                                                                                                                                     | 2229/? [2:08:58<00:00,  0.29it/s, v_num=seh9]train step 2230; scene = [['873b6566b29196f3'], ['18061c7463438d89'], ['3055585671a8da3f'], ['d6a1f3e13c45df99']]; loss = 0.016572
+Epoch 0: |                                                                                                                                                     | 2230/? [2:09:01<00:00,  0.29it/s, v_num=seh9]context = [[106, 110, 132, 133, 135, 147, 162, 165], [17, 23, 50, 60, 61, 64, 66, 71], [68, 70, 78, 90, 113, 128, 132, 146]]target = [[137, 109, 143, 111, 121, 152, 139, 107], [51, 57, 48, 20, 18, 59, 24, 70], [92, 98, 89, 79, 125, 145, 134, 91]]
+Epoch 0: |                                                                                                                                                     | 2239/? [2:09:33<00:00,  0.29it/s, v_num=seh9]train step 2240; scene = [['513826433660cb1c'], ['53d5559c250a0f44'], ['f44cc142d9796ff7'], ['f199ea57262f903a'], ['bf00a8f83ba6fb09'], ['6b618bf721772750']]; loss = 0.031263
+Epoch 0: |                                                                                                                                                     | 2240/? [2:09:37<00:00,  0.29it/s, v_num=seh9]context = [[33, 41, 43, 61, 63, 102, 107, 116], [215, 228, 229, 235, 240, 266, 269, 274], [19, 36, 38, 42, 61, 89, 94, 97]]target = [[114, 45, 55, 110, 100, 79, 63, 66], [251, 235, 227, 267, 219, 257, 269, 258], [72, 29, 43, 69, 55, 88, 58, 76]]
+Epoch 0: |                                                                                                                                                     | 2249/? [2:10:09<00:00,  0.29it/s, v_num=seh9]train step 2250; scene = [['75cfeb226b47a5a1'], ['ff70172dd4ad43ec'], ['496e2618d0fb9556']]; loss = 0.013006
+Epoch 0: |                                                                                                                                                     | 2250/? [2:10:12<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                                                                                           | 0/? [00:00<?, ?it/s]validation step 2250; scene = ['651a7f83ed093001'];
+target intrinsic:  tensor(0.8796, device='cuda:0') tensor(0.8798, device='cuda:0')                                                                                                      | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9024, device='cuda:0') tensor(0.9020, device='cuda:0')
+[2026-02-25 03:51:23,880][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                                                                                                     | 2250/? [2:10:13<00:00,  0.29it/s, v_num=seh9]context = [[11, 13, 18, 19, 21, 29, 35, 39, 41, 44, 59, 61, 65, 66, 69, 70, 73, 84, 86, 91, 94, 95, 100, 108]]target = [[51, 83, 56, 90, 33, 46, 48, 26, 84, 28, 80, 105, 77, 69, 42, 41, 63, 76, 15, 64, 13, 87, 27, 47]]
+Epoch 0: |                                                                                                                                                     | 2259/? [2:10:45<00:00,  0.29it/s, v_num=seh9]train step 2260; scene = [['7277bcf45f9b2f21'], ['a4ccb53922783c60'], ['83bf7f4e33c13400']]; loss = 0.010474
+Epoch 0: |                                                                                                                                                     | 2260/? [2:10:48<00:00,  0.29it/s, v_num=seh9]context = [[27, 37, 42, 64, 74, 83, 100, 107], [13, 15, 21, 23, 24, 37, 39, 65], [5, 11, 13, 17, 28, 32, 51, 71]]target = [[96, 45, 92, 58, 29, 79, 90, 97], [28, 50, 36, 23, 49, 62, 37, 64], [58, 23, 41, 15, 64, 27, 12, 7]]
+Epoch 0: |                                                                                                                                                     | 2269/? [2:11:20<00:00,  0.29it/s, v_num=seh9]train step 2270; scene = [['68530d6f86185a8c']]; loss = 0.015295
+Epoch 0: |                                                                                                                                                     | 2270/? [2:11:23<00:00,  0.29it/s, v_num=seh9]context = [[167, 189, 197, 200, 204, 216, 219, 220, 225, 227, 234, 235], [8, 12, 16, 22, 23, 31, 33, 35, 39, 41, 42, 61]]target = [[184, 220, 193, 233, 225, 170, 215, 180, 186, 227, 207, 206], [27, 40, 18, 35, 38, 37, 45, 24, 22, 25, 59, 36]]
+Epoch 0: |                                                                                                                                                     | 2279/? [2:11:55<00:00,  0.29it/s, v_num=seh9]train step 2280; scene = [['f1b09696ccbe7ed0']]; loss = 0.015964
+Epoch 0: |                                                                                                                                                     | 2280/? [2:11:58<00:00,  0.29it/s, v_num=seh9]context = [[9, 26, 30, 50, 58, 73, 83, 84], [45, 46, 76, 80, 81, 88, 93, 103], [5, 26, 29, 33, 37, 50, 52, 75]]target = [[43, 51, 31, 78, 76, 27, 15, 74], [55, 57, 94, 70, 99, 46, 81, 53], [10, 47, 74, 20, 27, 61, 56, 14]]
+Epoch 0: |                                                                                                                                                     | 2289/? [2:12:29<00:00,  0.29it/s, v_num=seh9]train step 2290; scene = [['077a53a763c9b806'], ['a5d23080c6b858b0']]; loss = 0.017475
+Epoch 0: |                                                                                                                                                     | 2290/? [2:12:32<00:00,  0.29it/s, v_num=seh9]context = [[23, 31, 35, 42, 46, 48, 49, 51, 56, 61, 62, 65, 75, 77, 79, 84, 85, 86, 93, 100, 103, 106, 118, 120]]target = [[42, 67, 85, 32, 113, 77, 84, 44, 103, 81, 73, 27, 75, 35, 41, 87, 61, 53, 48, 37, 99, 110, 24, 86]]
+Epoch 0: |                                                                                                                                                     | 2299/? [2:13:04<00:00,  0.29it/s, v_num=seh9]train step 2300; scene = [['1f6b10c1bfeab825'], ['fefab354938651ee'], ['773592c80a3cbf02']]; loss = 0.011146
+Epoch 0: |                                                                                                                                                     | 2300/? [2:13:07<00:00,  0.29it/s, v_num=seh9]context = [[11, 29, 34, 37, 43, 44, 45, 67, 80, 81, 85, 87], [25, 33, 39, 42, 53, 54, 66, 67, 69, 70, 76, 78]]target = [[58, 78, 21, 29, 35, 65, 76, 43, 64, 12, 42, 15], [57, 34, 75, 45, 35, 60, 77, 67, 76, 62, 50, 36]]
+Epoch 0: |                                                                                                                                                     | 2309/? [2:13:37<00:00,  0.29it/s, v_num=seh9]train step 2310; scene = [['6349ab522cc11c42'], ['96bab6e15eb77d4e'], ['3d5b0940a28bb67d'], ['1d58a359467ff24f'], ['60c37b519a01205d'], ['6aeff4800dc767c8'], ['56a0da3c4605e004'], ['3c52bbe20a514f48'], ['f4e639c0c0c58d2f'], ['ebab00ac8428f90a'], ['57a8005cadd17369'], ['ac5159b7b9d8ab53']]; loss = 0.058609
+Epoch 0: |                                                                                                                                                     | 2310/? [2:13:41<00:00,  0.29it/s, v_num=seh9]context = [[73, 83, 87, 89, 112, 133], [121, 130, 190, 192, 197, 201], [140, 172, 174, 179, 185, 190], [22, 37, 52, 58, 63, 78]]target = [[129, 119, 117, 112, 74, 118], [186, 161, 184, 197, 195, 156], [177, 159, 184, 147, 143, 163], [39, 53, 41, 25, 42, 77]]
+Epoch 0: |                                                                                                                                                     | 2319/? [2:14:12<00:00,  0.29it/s, v_num=seh9]train step 2320; scene = [['ce50e60b3d231911'], ['b43b9870fd778691'], ['7f52ee494a9f69f1'], ['c48ef8fab1483416']]; loss = 0.009144
+Epoch 0: |                                                                                                                                                     | 2320/? [2:14:16<00:00,  0.29it/s, v_num=seh9]context = [[48, 49, 54, 64, 69, 71, 76, 86, 89, 92, 93, 95, 97, 105, 108, 109, 113, 119, 120, 123, 125, 126, 133, 145]]target = [[134, 94, 61, 54, 51, 119, 72, 68, 100, 50, 79, 104, 85, 52, 128, 89, 103, 116, 70, 57, 115, 130, 66, 62]]
+Epoch 0: |                                                                                                                                                     | 2329/? [2:14:48<00:00,  0.29it/s, v_num=seh9]train step 2330; scene = [['52b984bea1f78ad6'], ['f5d47cbf373ef48c']]; loss = 0.012140
+Epoch 0: |                                                                                                                                                     | 2330/? [2:14:51<00:00,  0.29it/s, v_num=seh9]context = [[12, 33, 40, 47, 53, 54, 63, 66], [55, 74, 105, 106, 114, 115, 117, 137], [41, 46, 55, 59, 63, 73, 85, 97]]target = [[14, 56, 35, 30, 42, 57, 40, 53], [118, 82, 89, 122, 59, 87, 79, 92], [93, 70, 85, 86, 79, 73, 88, 84]]
+Epoch 0: |                                                                                                                                                     | 2339/? [2:15:22<00:00,  0.29it/s, v_num=seh9]train step 2340; scene = [['f353f0555d221d44'], ['871ab234bc798889'], ['58a049a2126413d5'], ['6f29c6e5920ebf58'], ['6b549d67d37b3d6a'], ['4c768903212b223d'], ['0e7f80803adb73f5'], ['e39974ca56f48929'], ['7e8beb820f189792'], ['da3a6901f12a6e59'], ['de1c9674255ba611'], ['af1aad7fedbf1fc7']]; loss = 0.019785
+Epoch 0: |                                                                                                                                                     | 2340/? [2:15:26<00:00,  0.29it/s, v_num=seh9]context = [[0, 8, 17, 20, 27, 44, 47, 49], [30, 43, 45, 48, 72, 79, 90, 99], [33, 35, 45, 47, 54, 61, 66, 85]]target = [[34, 28, 7, 10, 9, 32, 2, 3], [57, 45, 73, 41, 78, 39, 53, 51], [45, 81, 42, 47, 50, 59, 44, 62]]
+Epoch 0: |                                                                                                                                                     | 2349/? [2:15:57<00:00,  0.29it/s, v_num=seh9]train step 2350; scene = [['77542628c9955b43'], ['b6f9cfe435a0fde7']]; loss = 0.009671
+Epoch 0: |                                                                                                                                                     | 2350/? [2:16:01<00:00,  0.29it/s, v_num=seh9]context = [[19, 22, 24, 28, 31, 35, 36, 37, 38, 43, 48, 54, 58, 59, 60, 62, 71, 74, 79, 80, 82, 88, 94, 116]]target = [[80, 48, 21, 102, 57, 91, 39, 74, 89, 67, 60, 87, 35, 27, 43, 85, 77, 61, 76, 45, 41, 23, 33, 81]]
+Epoch 0: |                                                                                                                                                     | 2359/? [2:16:33<00:00,  0.29it/s, v_num=seh9]train step 2360; scene = [['a3dee13fa0216d57'], ['f77ef6219492d962'], ['5bfcfd9def7a2b51'], ['fd1737c9d5f1cb37'], ['62e8f32c902fe339'], ['fb32daf720b64da7'], ['f46d9b174fbfd403'], ['3828ad130b1c93b4']]; loss = 0.017326
+Epoch 0: |                                                                                                                                                     | 2360/? [2:16:36<00:00,  0.29it/s, v_num=seh9]context = [[13, 14, 15, 20, 28, 33, 43, 70, 72, 73, 76, 78], [126, 134, 137, 165, 167, 168, 174, 187, 193, 197, 202, 212]]target = [[52, 44, 31, 72, 30, 50, 70, 62, 16, 76, 69, 15], [192, 186, 199, 161, 138, 194, 145, 183, 191, 147, 140, 170]]
+Epoch 0: |                                                                                                                                                     | 2369/? [2:17:08<00:00,  0.29it/s, v_num=seh9]train step 2370; scene = [['0a13c96ef5627832']]; loss = 0.013553
+Epoch 0: |                                                                                                                                                     | 2370/? [2:17:11<00:00,  0.29it/s, v_num=seh9]context = [[89, 91, 103, 138], [5, 7, 31, 67], [29, 47, 58, 102], [16, 45, 49, 61], [3, 13, 40, 51], [3, 17, 52, 54]]target = [[113, 98, 128, 99], [40, 56, 10, 52], [89, 55, 53, 97], [26, 39, 28, 41], [24, 23, 40, 12], [20, 48, 30, 35]]
+Epoch 0: |                                                                                                                                                     | 2379/? [2:17:43<00:00,  0.29it/s, v_num=seh9]train step 2380; scene = [['43e59b21a3805c73']]; loss = 0.016151
+Epoch 0: |                                                                                                                                                     | 2380/? [2:17:47<00:00,  0.29it/s, v_num=seh9]context = [[1, 7, 38, 48, 73, 75, 87, 91], [3, 6, 10, 20, 22, 40, 47, 48], [136, 152, 165, 167, 168, 169, 189, 195]]target = [[66, 3, 70, 59, 50, 65, 2, 86], [26, 22, 24, 35, 21, 37, 10, 6], [153, 190, 167, 166, 192, 172, 143, 183]]
+Epoch 0: |                                                                          | 2389/? [2:18:18<00:00,  0.29it/s, v_num=seh9]train step 2390; scene = [['1559e264c5f4c481'], ['6db9c9051381e481'], ['89fc915d613ff37f'], ['f5f5c6090e031091']]; loss = 0.011272
+Epoch 0: |                                                                          | 2390/? [2:18:22<00:00,  0.29it/s, v_num=seh9]context = [[138, 141, 142, 146, 151, 160, 166, 181, 187, 191, 199, 202], [5, 9, 17, 18, 29, 35, 39, 40, 47, 52, 70, 79]]target = [[196, 162, 177, 163, 155, 181, 161, 146, 178, 199, 190, 154], [58, 48, 23, 36, 42, 18, 13, 60, 49, 69, 63, 71]]
+Epoch 0: |                                                                          | 2399/? [2:18:53<00:00,  0.29it/s, v_num=seh9]train step 2400; scene = [['c411d27c23509ccb'], ['521aa01f117a1f3d'], ['b65dd487d8a9df81'], ['a30e87ca123b7dcc']]; loss = 0.011145
+Epoch 0: |                                                                          | 2400/? [2:18:55<00:00,  0.29it/s, v_num=seh9]context = [[132, 142, 149, 151, 153, 158, 162, 166, 179, 180, 184, 187, 190, 191, 192, 200, 203, 208, 212, 218, 220, 221, 222, 229]]target = [[203, 210, 195, 163, 183, 187, 226, 212, 179, 142, 177, 216, 190, 139, 213, 169, 161, 154, 135, 196, 151, 224, 137, 145]]
+[2026-02-25 04:00:10,015][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2409/? [2:19:27<00:00,  0.29it/s, v_num=seh9]train step 2410; scene = [['d35650707515c2cf'], ['3c60078735491108'], ['7f6de4b0a9fc2031']]; loss = 0.011827
+Epoch 0: |                                                                          | 2410/? [2:19:31<00:00,  0.29it/s, v_num=seh9]context = [[20, 21, 32, 37, 38, 39, 42, 43, 61, 65, 66, 67, 68, 72, 74, 75, 78, 83, 86, 102, 103, 108, 116, 117]]target = [[27, 41, 58, 112, 75, 82, 94, 61, 96, 116, 30, 56, 69, 109, 50, 64, 44, 86, 79, 104, 59, 68, 46, 93]]
+Epoch 0: |                                                                          | 2419/? [2:20:02<00:00,  0.29it/s, v_num=seh9]train step 2420; scene = [['894851ce05febce5']]; loss = 0.010643
+Epoch 0: |                                                                          | 2420/? [2:20:06<00:00,  0.29it/s, v_num=seh9]context = [[41, 58, 59, 65, 68, 71, 80, 129], [176, 180, 184, 192, 209, 215, 216, 230], [105, 107, 111, 113, 115, 153, 154, 161]]target = [[42, 95, 110, 106, 113, 82, 47, 46], [205, 198, 190, 183, 203, 206, 201, 180], [134, 142, 114, 111, 150, 135, 120, 115]]
+Epoch 0: |                                                                          | 2429/? [2:20:37<00:00,  0.29it/s, v_num=seh9]train step 2430; scene = [['8a6f6565aa136bfd']]; loss = 0.007884
+Epoch 0: |                                                                          | 2430/? [2:20:41<00:00,  0.29it/s, v_num=seh9]context = [[16, 25, 29, 30, 46, 63, 68, 69], [20, 24, 48, 50, 55, 64, 80, 81], [212, 220, 236, 237, 241, 250, 257, 261]]target = [[37, 67, 36, 34, 27, 33, 23, 35], [41, 44, 56, 25, 30, 21, 52, 75], [227, 220, 237, 215, 247, 239, 260, 236]]
+Epoch 0: |                                                                          | 2439/? [2:21:11<00:00,  0.29it/s, v_num=seh9]train step 2440; scene = [['d2bf3cc61876b88a'], ['41d454a6a4ccc3c6'], ['50c8a233bd82a613'], ['1b8238f8057a7975'], ['5e89c7650c97cdf1'], ['62b4841181d253e7'], ['49328b9c29331060'], ['77c6dd845fd239ec']]; loss = 0.018345
+Epoch 0: |                                                                          | 2440/? [2:21:14<00:00,  0.29it/s, v_num=seh9]context = [[12, 13, 20, 27, 30, 31, 39, 40, 41, 43, 44, 53, 56, 57, 60, 66, 75, 79, 82, 83, 86, 100, 107, 109]]target = [[106, 96, 48, 79, 36, 108, 37, 25, 43, 88, 38, 84, 83, 21, 76, 58, 33, 94, 73, 78, 30, 53, 64, 99]]
+Epoch 0: |                                                                          | 2449/? [2:21:46<00:00,  0.29it/s, v_num=seh9]train step 2450; scene = [['a8abd4eeea08e1ad']]; loss = 0.010394
+Epoch 0: |                                                                          | 2450/? [2:21:50<00:00,  0.29it/s, v_num=seh9]context = [[15, 21, 22, 26, 29, 34, 40, 41, 46, 49, 51, 52, 55, 62, 63, 64, 65, 68, 70, 77, 78, 84, 110, 112]]target = [[83, 49, 68, 55, 107, 44, 81, 102, 97, 89, 86, 19, 32, 21, 52, 62, 95, 64, 61, 88, 59, 101, 22, 18]]
+Epoch 0: |                                                                          | 2459/? [2:22:21<00:00,  0.29it/s, v_num=seh9]train step 2460; scene = [['9d3f87d62d49025d'], ['c152ab577a837dac']]; loss = 0.015436
+Epoch 0: |                                                                          | 2460/? [2:22:25<00:00,  0.29it/s, v_num=seh9]context = [[176, 177, 180, 212, 215, 228], [18, 27, 52, 75, 98, 101], [17, 38, 52, 81, 85, 104], [7, 36, 82, 90, 92, 95]]target = [[200, 210, 201, 198, 191, 216], [40, 53, 56, 91, 23, 48], [67, 37, 35, 81, 39, 41], [73, 92, 61, 29, 46, 39]]
+Epoch 0: |                                                                          | 2469/? [2:22:56<00:00,  0.29it/s, v_num=seh9]train step 2470; scene = [['5614d74255c9c07c'], ['abe30f78eedbc519'], ['8ef9ff3189c85eee'], ['ff66444e1858620a'], ['4976860ffb8e6b02'], ['4af1f08d7b5d1523']]; loss = 0.016927
+Epoch 0: |                                                                          | 2470/? [2:23:00<00:00,  0.29it/s, v_num=seh9]context = [[12, 14, 18, 30, 40, 48, 49, 67], [57, 60, 67, 84, 88, 97, 121, 124], [143, 160, 169, 181, 185, 190, 198, 212]]target = [[15, 13, 65, 59, 38, 63, 48, 20], [72, 87, 80, 69, 114, 93, 81, 99], [168, 172, 179, 187, 173, 149, 180, 189]]
+Epoch 0: |                                                                          | 2479/? [2:23:31<00:00,  0.29it/s, v_num=seh9]train step 2480; scene = [['1a402532663d05ad'], ['36b937cc3684eddb'], ['08da23838ee6e23b'], ['ef29eabcdae21636']]; loss = 0.014801
+Epoch 0: |                                                                          | 2480/? [2:23:34<00:00,  0.29it/s, v_num=seh9]context = [[29, 84, 88], [169, 172, 215], [20, 54, 69], [25, 69, 84], [211, 272, 277], [48, 95, 133], [112, 176, 185], [85, 102, 149]]target = [[59, 75, 32], [185, 192, 191], [34, 39, 56], [75, 51, 31], [256, 233, 240], [66, 98, 119], [122, 149, 170], [144, 138, 104]]
+Epoch 0: |                                                                          | 2489/? [2:24:04<00:00,  0.29it/s, v_num=seh9]train step 2490; scene = [['efbdf67e0c80c27b'], ['982da008194b287c'], ['4eb66b49aeb1d641'], ['9c500c3d949c224e']]; loss = 0.016549
+Epoch 0: |                                                                          | 2490/? [2:24:08<00:00,  0.29it/s, v_num=seh9]context = [[0, 5, 14, 16, 23, 25, 26, 28, 30, 34, 42, 49], [46, 60, 65, 67, 74, 75, 86, 88, 89, 90, 93, 117]]target = [[29, 26, 30, 16, 31, 11, 28, 36, 45, 18, 21, 40], [56, 104, 112, 73, 102, 57, 68, 52, 74, 91, 109, 61]]
+Epoch 0: |                                                                          | 2499/? [2:24:39<00:00,  0.29it/s, v_num=seh9]train step 2500; scene = [['a48e4e90b76cc3a3'], ['715e8695976cdb61'], ['40b14e9ca06271a0'], ['f46f73c6b994a630'], ['4d32322df2d07217'], ['faa760d7d2c034ec']]; loss = 0.015343
+Epoch 0: |                                                                          | 2500/? [2:24:41<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 2500; scene = ['97ef4323919c5e8a'];
+target intrinsic:  tensor(0.8889, device='cuda:0') tensor(0.8892, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9520, device='cuda:0') tensor(0.9468, device='cuda:0')
+[2026-02-25 04:05:52,901][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2500/? [2:24:42<00:00,  0.29it/s, v_num=seh9]context = [[135, 137, 148, 151, 152, 157, 185, 186, 192, 198, 200, 204, 208, 212, 214, 215, 217, 220, 221, 222, 223, 225, 226, 232]]target = [[228, 186, 188, 214, 163, 148, 155, 141, 146, 149, 164, 198, 194, 205, 145, 176, 151, 213, 229, 174, 201, 199, 181, 224]]
+Epoch 0: |                                                                          | 2509/? [2:25:14<00:00,  0.29it/s, v_num=seh9]train step 2510; scene = [['62bfd0a1ecd6a33c'], ['86fa33e82678f2b4'], ['a7e5c2b2c8fe38eb'], ['313c2ce83a17b22d']]; loss = 0.008339
+Epoch 0: |                                                                          | 2510/? [2:25:17<00:00,  0.29it/s, v_num=seh9]context = [[2, 3, 27, 39, 52, 60, 66, 88], [6, 22, 31, 32, 38, 39, 46, 53], [32, 41, 51, 59, 63, 65, 77, 101]]target = [[60, 18, 51, 6, 77, 10, 8, 87], [9, 46, 34, 15, 43, 31, 44, 35], [78, 55, 48, 100, 77, 84, 57, 82]]
+Epoch 0: |                                                                          | 2519/? [2:25:49<00:00,  0.29it/s, v_num=seh9]train step 2520; scene = [['3fc603b4c4531c11']]; loss = 0.046726
+Epoch 0: |                                                                          | 2520/? [2:25:53<00:00,  0.29it/s, v_num=seh9]context = [[22, 24, 38, 48, 62, 78, 83, 95], [5, 11, 30, 31, 37, 58, 70, 72], [23, 32, 33, 59, 94, 97, 101, 104]]target = [[94, 35, 69, 29, 28, 40, 72, 37], [42, 57, 40, 56, 18, 34, 11, 37], [97, 70, 59, 44, 51, 52, 34, 35]]
+Epoch 0: |                                                                          | 2529/? [2:26:24<00:00,  0.29it/s, v_num=seh9]train step 2530; scene = [['7dc5c394263df267'], ['b113c611f9029c49']]; loss = 0.013636
+Epoch 0: |                                                                          | 2530/? [2:26:27<00:00,  0.29it/s, v_num=seh9]context = [[144, 162, 163, 167, 185, 191], [44, 48, 58, 102, 106, 107], [23, 30, 71, 84, 86, 89], [1, 13, 23, 28, 54, 83]]target = [[176, 153, 187, 165, 152, 157], [76, 101, 48, 62, 78, 67], [40, 37, 41, 33, 34, 73], [51, 39, 8, 24, 70, 79]]
+Epoch 0: |                                                                          | 2539/? [2:27:00<00:00,  0.29it/s, v_num=seh9]train step 2540; scene = [['85b281f0e77a632f'], ['66ae84fb53b12759']]; loss = 0.012805
+Epoch 0: |                                                                          | 2540/? [2:27:03<00:00,  0.29it/s, v_num=seh9]context = [[11, 17, 21, 24, 33, 34, 50, 63, 68, 69, 72, 88], [5, 13, 17, 19, 22, 24, 31, 33, 35, 36, 53, 54]]target = [[54, 39, 52, 20, 16, 23, 77, 65, 87, 43, 66, 78], [9, 33, 36, 44, 52, 11, 6, 51, 14, 29, 22, 30]]
+Epoch 0: |                                                                          | 2549/? [2:27:34<00:00,  0.29it/s, v_num=seh9]train step 2550; scene = [['b35b0b431fab2105']]; loss = 0.013285
+Epoch 0: |                                                                          | 2550/? [2:27:38<00:00,  0.29it/s, v_num=seh9]context = [[96, 100, 101, 114, 118, 121, 122, 123, 124, 136, 140, 142, 149, 150, 151, 159, 160, 162, 163, 167, 169, 173, 190, 193]]target = [[165, 189, 177, 184, 168, 112, 131, 182, 138, 164, 101, 111, 192, 121, 128, 120, 132, 125, 135, 167, 163, 178, 142, 170]]
+Epoch 0: |                                                                          | 2559/? [2:28:10<00:00,  0.29it/s, v_num=seh9]train step 2560; scene = [['987ba4e94414a901'], ['4a74477406314bef'], ['e1bfe1e13278c747'], ['917b6ab7c8384c3b'], ['7eb0dffcddc1722c'], ['0fa6bc2796cbba41'], ['a039939946c82ed7'], ['478e07af7840c08a'], ['395fd5b1237500c6'], ['50468b60ba969ac2'], ['ee7cb189dff51d83'], ['aed762fe049b1f86']]; loss = 0.021388
+Epoch 0: |                                                                          | 2560/? [2:28:13<00:00,  0.29it/s, v_num=seh9]context = [[27, 31, 41, 55, 59, 67, 70, 73], [47, 52, 65, 80, 90, 92, 94, 100], [51, 66, 72, 99, 109, 113, 122, 124]]target = [[70, 71, 35, 43, 64, 67, 53, 52], [53, 94, 82, 67, 55, 63, 85, 88], [84, 116, 57, 61, 104, 88, 80, 79]]
+Epoch 0: |                                                                          | 2569/? [2:28:45<00:00,  0.29it/s, v_num=seh9]train step 2570; scene = [['b7873e1ebdb5721f']]; loss = 0.016877
+Epoch 0: |                                                                          | 2570/? [2:28:49<00:00,  0.29it/s, v_num=seh9]context = [[4, 6, 12, 16, 31, 37, 39, 40, 47, 50, 52, 63], [3, 4, 14, 22, 23, 35, 37, 38, 40, 51, 64, 74]]target = [[41, 62, 34, 35, 5, 31, 18, 24, 47, 55, 29, 10], [52, 40, 60, 63, 19, 11, 36, 66, 4, 16, 42, 18]]
+Epoch 0: |                                                                          | 2579/? [2:29:20<00:00,  0.29it/s, v_num=seh9]train step 2580; scene = [['70b45bbd1147fbf0'], ['1b2bfb2e03827c26'], ['9ade5b1dc78259b9'], ['d79b084da2f40032']]; loss = 0.016024
+Epoch 0: |                                                                          | 2580/? [2:29:23<00:00,  0.29it/s, v_num=seh9]context = [[0, 68], [30, 89], [53, 123], [12, 87], [0, 80], [135, 188], [11, 71], [9, 74], [1, 88], [4, 58], [4, 54], [7, 94]]target = [[53, 1], [83, 49], [118, 84], [63, 81], [25, 71], [143, 139], [66, 33], [56, 18], [46, 35], [48, 8], [23, 11], [53, 32]]
+Epoch 0: |                                                                          | 2589/? [2:29:54<00:00,  0.29it/s, v_num=seh9]train step 2590; scene = [['21a170c902c43a97'], ['0e57d2d8655afebb'], ['362a71463cb49249'], ['02843207f75c20d3']]; loss = 0.011718
+Epoch 0: |                                                                          | 2590/? [2:29:58<00:00,  0.29it/s, v_num=seh9]context = [[119, 139, 141, 145, 167, 171, 173, 177], [0, 3, 9, 22, 36, 42, 71, 81], [11, 16, 22, 26, 29, 30, 41, 58]]target = [[128, 160, 123, 142, 139, 157, 151, 164], [10, 34, 9, 21, 4, 65, 15, 7], [25, 14, 22, 51, 33, 45, 15, 21]]
+Epoch 0: |                                                                          | 2599/? [2:30:29<00:00,  0.29it/s, v_num=seh9]train step 2600; scene = [['31c1855b8ad30220']]; loss = 0.010673
+Epoch 0: |                                                                          | 2600/? [2:30:33<00:00,  0.29it/s, v_num=seh9]context = [[5, 10, 18, 19, 22, 28, 29, 34, 44, 51, 54, 60], [81, 82, 84, 90, 93, 95, 99, 109, 112, 124, 130, 131]]target = [[51, 48, 26, 14, 22, 54, 50, 23, 46, 6, 52, 53], [127, 85, 93, 121, 94, 88, 90, 101, 82, 104, 109, 108]]
+[2026-02-25 04:11:47,470][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2609/? [2:31:04<00:00,  0.29it/s, v_num=seh9]train step 2610; scene = [['fef928bdd59753c7'], ['d715711625070e85'], ['3b7c67b6632b1c74']]; loss = 0.010456
+Epoch 0: |                                                                          | 2610/? [2:31:07<00:00,  0.29it/s, v_num=seh9]context = [[18, 20, 39, 41, 42, 43, 51, 55, 59, 61, 62, 67], [6, 14, 16, 35, 47, 53, 64, 68, 72, 77, 82, 85]]target = [[61, 23, 64, 31, 32, 55, 41, 66, 38, 50, 63, 49], [66, 15, 52, 48, 21, 80, 51, 56, 54, 62, 32, 61]]
+Epoch 0: |                                                                          | 2619/? [2:31:39<00:00,  0.29it/s, v_num=seh9]train step 2620; scene = [['8481bc2172478625']]; loss = 0.019401
+Epoch 0: |                                                                          | 2620/? [2:31:43<00:00,  0.29it/s, v_num=seh9]context = [[151, 154, 158, 161, 167, 174, 180, 185, 189, 191, 193, 204, 209, 210, 211, 216, 224, 229, 237, 240, 242, 243, 244, 248]]target = [[179, 180, 165, 183, 223, 207, 176, 156, 194, 178, 243, 159, 214, 184, 210, 191, 168, 211, 237, 195, 170, 224, 166, 225]]
+Epoch 0: |                                                                          | 2629/? [2:32:12<00:00,  0.29it/s, v_num=seh9]train step 2630; scene = [['d8ad0faac7caff87'], ['673c44a2c8176eb6'], ['02a58849f00b13fc'], ['d5f11c320ee101a1']]; loss = 0.023816
+Epoch 0: |                                                                          | 2630/? [2:32:16<00:00,  0.29it/s, v_num=seh9]context = [[32, 77, 80], [67, 92, 114], [55, 108, 118], [153, 192, 205], [12, 21, 97], [214, 219, 273], [29, 81, 91], [145, 182, 222]]target = [[63, 68, 40], [102, 85, 77], [117, 89, 102], [169, 171, 167], [75, 68, 31], [242, 253, 252], [58, 59, 75], [198, 172, 197]]
+Epoch 0: |                                                                          | 2639/? [2:32:48<00:00,  0.29it/s, v_num=seh9]train step 2640; scene = [['08c09a60291a2f5b'], ['210ab3c1a9d655e5'], ['44f828162649e72a'], ['363ee2c7afc81e84'], ['78b83f9c9174e8cc'], ['d8a3b176b0529293'], ['f276b95e48af6e36'], ['5ae12f0ee3f7ad6f']]; loss = 0.015295
+Epoch 0: |                                                                          | 2640/? [2:32:51<00:00,  0.29it/s, v_num=seh9]context = [[178, 186, 187, 190, 197, 200, 203, 204, 221, 224, 227, 234], [0, 5, 15, 21, 28, 29, 33, 35, 37, 38, 49, 50]]target = [[209, 190, 203, 187, 191, 214, 205, 201, 202, 179, 212, 215], [25, 21, 46, 27, 44, 1, 40, 41, 16, 49, 35, 3]]
+Epoch 0: |                                                                          | 2649/? [2:33:23<00:00,  0.29it/s, v_num=seh9]train step 2650; scene = [['376c7ff59babc257']]; loss = 0.009468
+Epoch 0: |                                                                          | 2650/? [2:33:27<00:00,  0.29it/s, v_num=seh9]context = [[3, 18, 29, 58], [0, 17, 61, 84], [60, 99, 117, 131], [0, 7, 12, 45], [37, 60, 81, 101], [183, 190, 222, 244]]target = [[49, 31, 35, 25], [40, 50, 67, 10], [100, 115, 120, 107], [41, 34, 1, 29], [39, 43, 78, 67], [232, 242, 223, 191]]
+Epoch 0: |                                                                          | 2659/? [2:33:57<00:00,  0.29it/s, v_num=seh9]train step 2660; scene = [['5d50f21bfe94df8c'], ['a0f0af3e118b6ca6'], ['1ed3d0dda9649020'], ['0e613a7c35450462'], ['985af15298bdcea5'], ['fd859fb51b16cf6e'], ['c35de22a7238f9a5'], ['4e588d9d58d1ec06']]; loss = 0.022878
+Epoch 0: |                                                                          | 2660/? [2:34:01<00:00,  0.29it/s, v_num=seh9]context = [[12, 16, 20, 21, 23, 26, 27, 29, 32, 36, 46, 49, 55, 62, 71, 83, 84, 90, 95, 99, 100, 102, 103, 109]]target = [[34, 86, 61, 97, 67, 73, 17, 43, 72, 98, 92, 80, 62, 87, 70, 42, 85, 90, 44, 41, 77, 75, 21, 57]]
+Epoch 0: |                                                                          | 2669/? [2:34:33<00:00,  0.29it/s, v_num=seh9]train step 2670; scene = [['ac16e227bc0144d6'], ['6b1950140a598578']]; loss = 0.009029
+Epoch 0: |                                                                          | 2670/? [2:34:36<00:00,  0.29it/s, v_num=seh9]context = [[0, 6, 9, 19, 23, 31, 41, 48], [9, 21, 22, 23, 24, 33, 70, 72], [51, 58, 66, 72, 73, 75, 81, 97]]target = [[39, 40, 27, 21, 18, 10, 12, 5], [53, 64, 18, 47, 26, 11, 43, 30], [89, 59, 82, 80, 93, 88, 77, 79]]
+Epoch 0: |                                                                          | 2679/? [2:35:07<00:00,  0.29it/s, v_num=seh9]train step 2680; scene = [['b6dd69cee72df5a3'], ['1fb562b09fc361ea'], ['46ed182c11fd6b04']]; loss = 0.006167
+Epoch 0: |                                                                          | 2680/? [2:35:11<00:00,  0.29it/s, v_num=seh9]context = [[82, 83, 86, 89, 94, 96, 99, 106, 107, 109, 114, 120, 122, 130, 138, 140, 142, 145, 149, 156, 165, 170, 171, 179]]target = [[175, 93, 160, 127, 130, 125, 110, 117, 142, 153, 152, 103, 158, 85, 138, 155, 89, 119, 98, 118, 96, 154, 88, 116]]
+Epoch 0: |                                                                          | 2689/? [2:35:41<00:00,  0.29it/s, v_num=seh9]train step 2690; scene = [['8d7519b2e98e73b0'], ['9a89163b62f0a058']]; loss = 0.010032
+Epoch 0: |                                                                          | 2690/? [2:35:44<00:00,  0.29it/s, v_num=seh9]context = [[46, 47, 48, 93, 113, 133], [118, 122, 129, 135, 158, 166], [1, 6, 36, 44, 47, 69], [73, 102, 110, 130, 141, 142]]target = [[91, 96, 65, 88, 86, 66], [139, 120, 125, 135, 148, 153], [32, 65, 39, 55, 18, 36], [125, 86, 94, 91, 141, 83]]
+Epoch 0: |                                                                          | 2699/? [2:36:16<00:00,  0.29it/s, v_num=seh9]train step 2700; scene = [['1b24cf4a586a15e7'], ['58901334e2d813d9']]; loss = 0.008089
+Epoch 0: |                                                                          | 2700/? [2:36:19<00:00,  0.29it/s, v_num=seh9]context = [[3, 5, 8, 12, 15, 21, 27, 28, 29, 36, 37, 40, 41, 44, 46, 56, 58, 60, 71, 80, 82, 92, 99, 100]]target = [[9, 60, 19, 69, 72, 79, 63, 59, 55, 54, 75, 21, 29, 82, 46, 98, 40, 91, 86, 4, 71, 83, 48, 85]]
+Epoch 0: |                                                                          | 2709/? [2:36:51<00:00,  0.29it/s, v_num=seh9]train step 2710; scene = [['ef380960d4d8a206'], ['d351ba344c572b7c'], ['4beb0dd348a2c905'], ['2f6a5dd7b6a7e992']]; loss = 0.021361
+Epoch 0: |                                                                          | 2710/? [2:36:54<00:00,  0.29it/s, v_num=seh9]context = [[145, 155, 166, 167, 176, 194, 199, 201, 203, 204, 206, 209], [0, 5, 14, 19, 26, 29, 36, 37, 41, 43, 45, 50]]target = [[163, 201, 195, 199, 181, 185, 151, 194, 190, 154, 171, 198], [31, 29, 43, 23, 22, 34, 9, 11, 5, 44, 47, 7]]
+Epoch 0: |                                                                          | 2719/? [2:37:26<00:00,  0.29it/s, v_num=seh9]train step 2720; scene = [['76b44b96d0d32f80'], ['a9df659c4acffb49'], ['759e2542191f378d'], ['b7003ac834dc298b'], ['4d836c051f02de01'], ['107fabf0d1dce254']]; loss = 0.014851
+Epoch 0: |                                                                          | 2720/? [2:37:29<00:00,  0.29it/s, v_num=seh9]context = [[30, 33, 45, 50, 62, 67, 91, 97, 98, 106, 107, 115], [95, 98, 102, 104, 105, 109, 113, 116, 133, 136, 138, 147]]target = [[31, 88, 109, 102, 63, 68, 65, 93, 40, 35, 38, 39], [124, 140, 128, 139, 123, 136, 103, 134, 127, 135, 113, 121]]
+Epoch 0: |                                                                          | 2729/? [2:38:01<00:00,  0.29it/s, v_num=seh9]train step 2730; scene = [['5c0ddb9de8c16f05'], ['e82ae746fe86d59a'], ['ea02d0f42c603c21']]; loss = 0.027190
+Epoch 0: |                                                                          | 2730/? [2:38:05<00:00,  0.29it/s, v_num=seh9]context = [[42, 78, 108, 126], [5, 26, 68, 94], [194, 228, 242, 267], [107, 124, 139, 154], [10, 21, 31, 62], [3, 15, 31, 68]]target = [[116, 59, 85, 117], [16, 79, 78, 41], [248, 204, 196, 264], [113, 147, 111, 134], [30, 27, 31, 41], [17, 25, 7, 30]]
+Epoch 0: |                                                                          | 2739/? [2:38:36<00:00,  0.29it/s, v_num=seh9]train step 2740; scene = [['9eba6b410f166fe0'], ['3bcf3bcfc5d5c365'], ['7d8c26c8ac910aa8'], ['571879c55dd10963'], ['0e0e4a867359f360'], ['1e7d7ef1404597f0'], ['43205c12fd83c588'], ['f25aed34a7d73d42']]; loss = 0.012910
+Epoch 0: |                                                                          | 2740/? [2:38:40<00:00,  0.29it/s, v_num=seh9]context = [[159, 160, 163, 168, 170, 171, 179, 190, 193, 202, 214, 217, 219, 222, 229, 230, 234, 236, 238, 241, 245, 248, 250, 256]]target = [[218, 230, 175, 234, 202, 170, 210, 206, 200, 197, 231, 205, 193, 225, 166, 223, 228, 209, 211, 195, 188, 215, 199, 173]]
+Epoch 0: |                                                                          | 2749/? [2:39:10<00:00,  0.29it/s, v_num=seh9]train step 2750; scene = [['a9cd1a8fc1fa2269']]; loss = 0.018071
+Epoch 0: |                                                                          | 2750/? [2:39:13<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 2750; scene = ['3e07add8413f8157'];
+target intrinsic:  tensor(0.8521, device='cuda:0') tensor(0.8523, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8650, device='cuda:0') tensor(0.8630, device='cuda:0')
+[2026-02-25 04:20:24,888][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2750/? [2:39:14<00:00,  0.29it/s, v_num=seh9]context = [[2, 14, 28, 57], [18, 38, 82, 100], [12, 34, 76, 92], [0, 46, 61, 82], [78, 86, 110, 131], [4, 6, 63, 81]]target = [[17, 12, 31, 19], [48, 27, 88, 71], [58, 73, 82, 35], [32, 23, 40, 38], [119, 105, 130, 91], [34, 32, 40, 59]]
+Epoch 0: |                                                                          | 2759/? [2:39:45<00:00,  0.29it/s, v_num=seh9]train step 2760; scene = [['88c708a41a1fb05d']]; loss = 0.007987
+Epoch 0: |                                                                          | 2760/? [2:39:49<00:00,  0.29it/s, v_num=seh9]context = [[15, 19, 26, 33, 35, 37, 39, 42, 44, 48, 51, 64], [13, 21, 46, 47, 52, 62, 63, 68, 72, 76, 91, 103]]target = [[60, 51, 18, 36, 23, 31, 27, 19, 39, 55, 62, 52], [44, 37, 66, 31, 32, 42, 15, 51, 74, 79, 96, 73]]
+Epoch 0: |                                                                          | 2769/? [2:40:21<00:00,  0.29it/s, v_num=seh9]train step 2770; scene = [['0ca3f423c85f9681'], ['ecbf048504e09969'], ['edaf19c032358e0d']]; loss = 0.008459
+Epoch 0: |                                                                          | 2770/? [2:40:24<00:00,  0.29it/s, v_num=seh9]context = [[57, 62, 65, 71, 73, 76, 86, 87, 89, 91, 92, 113, 118, 119, 121, 123, 141, 142, 144, 146, 148, 151, 152, 154]]target = [[94, 88, 60, 148, 117, 113, 66, 120, 114, 103, 112, 131, 119, 108, 78, 137, 123, 90, 99, 130, 67, 118, 129, 59]]
+Epoch 0: |                                                                          | 2779/? [2:40:56<00:00,  0.29it/s, v_num=seh9]train step 2780; scene = [['c97fb236ee21af38'], ['4a0f95a3db913b56'], ['cbe7d9bfe38d2de8'], ['6f1508676b76c4f7'], ['3ad6f38502e90d67'], ['04c740af3de5ae4c']]; loss = 0.029495
+Epoch 0: |                                                                          | 2780/? [2:40:59<00:00,  0.29it/s, v_num=seh9]context = [[2, 6, 7, 8, 11, 13, 14, 15, 28, 59, 60, 63, 64, 67, 68, 77, 79, 81, 84, 85, 87, 91, 93, 99]]target = [[67, 33, 38, 74, 94, 54, 47, 35, 69, 55, 97, 84, 34, 61, 22, 60, 44, 36, 26, 86, 90, 50, 30, 75]]
+Epoch 0: |                                                                          | 2789/? [2:41:32<00:00,  0.29it/s, v_num=seh9]train step 2790; scene = [['0666c63c6c8d6a9b'], ['d0aafe6b7593a8c6'], ['5682eadfab7a6bcd'], ['08f41adb663ab4f4'], ['d8d10da8948e5676'], ['665db13a67f47d42']]; loss = 0.017370
+Epoch 0: |                                                                          | 2790/? [2:41:35<00:00,  0.29it/s, v_num=seh9]context = [[30, 55, 75, 84, 96, 101], [22, 32, 46, 65, 102, 103], [3, 27, 42, 50, 52, 58], [11, 47, 55, 89, 90, 91]]target = [[34, 39, 69, 58, 40, 91], [100, 47, 87, 95, 77, 28], [17, 51, 34, 28, 26, 29], [41, 50, 42, 27, 86, 61]]
+Epoch 0: |                                                                          | 2799/? [2:42:07<00:00,  0.29it/s, v_num=seh9]train step 2800; scene = [['be6ab5e68b93a77e'], ['bea0e295ee56d42c'], ['8b420b4f59fb2756'], ['c49e7882e04566c0']]; loss = 0.009797
+Epoch 0: |                                                                          | 2800/? [2:42:10<00:00,  0.29it/s, v_num=seh9]context = [[10, 12, 13, 14, 20, 24, 25, 34, 35, 36, 44, 46, 47, 50, 55, 63, 66, 70, 79, 93, 99, 103, 106, 107]]target = [[78, 50, 100, 34, 13, 21, 89, 20, 60, 40, 88, 86, 63, 77, 106, 95, 61, 94, 66, 91, 52, 49, 82, 83]]
+[2026-02-25 04:23:24,966][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2809/? [2:42:42<00:00,  0.29it/s, v_num=seh9]train step 2810; scene = [['b60f8abb905bb1e1']]; loss = 0.012039
+Epoch 0: |                                                                          | 2810/? [2:42:46<00:00,  0.29it/s, v_num=seh9]context = [[3, 19, 89], [188, 252, 275], [14, 25, 101], [3, 21, 83], [2, 56, 92], [0, 11, 58], [22, 42, 91], [4, 57, 81]]target = [[47, 18, 22], [197, 239, 260], [47, 63, 65], [73, 77, 49], [48, 9, 14], [43, 12, 10], [60, 32, 86], [20, 78, 75]]
+Epoch 0: |                                                                          | 2819/? [2:43:18<00:00,  0.29it/s, v_num=seh9]train step 2820; scene = [['d3dddf816450d1d0'], ['795b958ce87d556a']]; loss = 0.039688
+Epoch 0: |                                                                          | 2820/? [2:43:21<00:00,  0.29it/s, v_num=seh9]context = [[87, 93, 95, 97, 105, 110, 127, 129, 131, 132, 133, 135, 139, 140, 147, 149, 159, 161, 166, 168, 180, 181, 182, 184]]target = [[97, 127, 143, 95, 121, 96, 169, 157, 168, 142, 122, 131, 149, 136, 100, 125, 166, 116, 111, 172, 93, 163, 118, 153]]
+Epoch 0: |                                                                          | 2829/? [2:43:53<00:00,  0.29it/s, v_num=seh9]train step 2830; scene = [['bce06efba2b1d37c']]; loss = 0.009283
+Epoch 0: |                                                                          | 2830/? [2:43:56<00:00,  0.29it/s, v_num=seh9]context = [[3, 5, 15, 18, 38, 39, 42, 43, 53, 56, 57, 61], [11, 27, 36, 39, 45, 47, 56, 58, 69, 81, 83, 84]]target = [[5, 10, 56, 28, 23, 57, 8, 17, 20, 60, 53, 35], [39, 44, 40, 55, 12, 48, 19, 22, 58, 49, 60, 57]]
+Epoch 0: |                                                                          | 2839/? [2:44:27<00:00,  0.29it/s, v_num=seh9]train step 2840; scene = [['489e61a289f59044'], ['35b4f262621408f8'], ['a17672177fe8f694']]; loss = 0.013595
+Epoch 0: |                                                                          | 2840/? [2:44:31<00:00,  0.29it/s, v_num=seh9]context = [[49, 52, 55, 59, 61, 71, 83, 88, 92, 95, 99, 116], [14, 24, 39, 50, 57, 63, 66, 80, 83, 87, 100, 103]]target = [[70, 96, 60, 86, 113, 58, 85, 110, 81, 51, 89, 65], [34, 96, 39, 72, 65, 91, 24, 60, 28, 57, 19, 90]]
+Epoch 0: |                                                                          | 2849/? [2:45:02<00:00,  0.29it/s, v_num=seh9]train step 2850; scene = [['eea9a40d34d9e8f3'], ['da887eb3ac647f2c'], ['4e6ba49f61d7c52a'], ['f1d926a3302e8534']]; loss = 0.015195
+Epoch 0: |                                                                          | 2850/? [2:45:06<00:00,  0.29it/s, v_num=seh9]context = [[30, 31, 32, 34, 45, 51, 53, 59, 71, 72, 74, 79, 85, 88, 93, 104, 106, 112, 115, 118, 119, 123, 124, 127]]target = [[61, 117, 83, 66, 57, 48, 126, 54, 119, 112, 107, 79, 80, 74, 47, 114, 121, 78, 64, 71, 96, 116, 88, 62]]
+Epoch 0: |                                                                          | 2859/? [2:45:37<00:00,  0.29it/s, v_num=seh9]train step 2860; scene = [['f6a68923ac68bc4a'], ['c008476d54d1ee07'], ['8572c96093dc0b71']]; loss = 0.048482
+Epoch 0: |                                                                          | 2860/? [2:45:40<00:00,  0.29it/s, v_num=seh9]context = [[39, 41, 43, 47, 52, 54, 58, 63, 64, 66, 78, 89], [23, 29, 31, 38, 46, 48, 58, 70, 73, 79, 80, 86]]target = [[82, 66, 77, 72, 73, 69, 64, 52, 62, 68, 67, 51], [68, 33, 76, 45, 53, 55, 44, 64, 31, 37, 47, 79]]
+Epoch 0: |                                                                          | 2869/? [2:46:11<00:00,  0.29it/s, v_num=seh9]train step 2870; scene = [['0abbdaad60ccd753']]; loss = 0.015458
+Epoch 0: |                                                                          | 2870/? [2:46:14<00:00,  0.29it/s, v_num=seh9]context = [[44, 49, 51, 57, 92, 105], [2, 10, 30, 50, 58, 59], [68, 77, 104, 142, 145, 148], [4, 16, 32, 39, 85, 94]]target = [[65, 56, 101, 92, 49, 76], [34, 15, 13, 24, 54, 52], [143, 102, 129, 120, 130, 85], [29, 52, 8, 59, 28, 27]]
+Epoch 0: |                                                                          | 2879/? [2:46:46<00:00,  0.29it/s, v_num=seh9]train step 2880; scene = [['8f811c66bfac4e2b'], ['f388608cf7295e88'], ['3241037fc88ad609']]; loss = 0.030268
+Epoch 0: |                                                                          | 2880/? [2:46:50<00:00,  0.29it/s, v_num=seh9]context = [[26, 38, 39, 48, 50, 53, 54, 61, 77, 84, 94, 101], [67, 68, 71, 76, 79, 85, 97, 104, 111, 112, 118, 121]]target = [[82, 38, 79, 30, 47, 75, 78, 68, 88, 57, 98, 42], [87, 83, 104, 101, 112, 116, 105, 93, 76, 111, 113, 88]]
+Epoch 0: |                                                                          | 2889/? [2:47:20<00:00,  0.29it/s, v_num=seh9]train step 2890; scene = [['f8b0d1e280daed5d'], ['5cd3756227c3a8c9'], ['5746d03325bf70d3'], ['c146f23a3704fb63']]; loss = 0.017699
+Epoch 0: |                                                                          | 2890/? [2:47:23<00:00,  0.29it/s, v_num=seh9]context = [[175, 182, 188, 191, 197, 203, 205, 206, 212, 213, 220, 221, 227, 233, 234, 237, 241, 243, 248, 257, 258, 261, 271, 272]]target = [[227, 176, 252, 248, 190, 203, 228, 194, 255, 225, 205, 182, 207, 217, 238, 185, 257, 208, 268, 221, 242, 188, 231, 232]]
+Epoch 0: |                                                                          | 2899/? [2:47:55<00:00,  0.29it/s, v_num=seh9]train step 2900; scene = [['f1a0ce57e7071dac'], ['450f13d2be008e30'], ['60e0925e79a6c253'], ['a67ed4a351f4a8c2']]; loss = 0.020135
+Epoch 0: |                                                                          | 2900/? [2:47:58<00:00,  0.29it/s, v_num=seh9]context = [[107, 112, 118, 127, 129, 135, 143, 150, 151, 154, 163, 166], [91, 100, 105, 106, 107, 111, 118, 120, 123, 130, 132, 145]]target = [[144, 156, 129, 161, 130, 140, 121, 143, 165, 135, 146, 164], [137, 103, 130, 143, 109, 115, 140, 97, 114, 111, 112, 141]]
+Epoch 0: |                                                                          | 2909/? [2:48:30<00:00,  0.29it/s, v_num=seh9]train step 2910; scene = [['60ebe67fcdbb0767'], ['36f89cf6ea6d7736'], ['3d88d591a0323964'], ['fef48b769b17f0ed']]; loss = 0.012795
+Epoch 0: |                                                                          | 2910/? [2:48:34<00:00,  0.29it/s, v_num=seh9]context = [[0, 2, 56, 60, 64, 65, 76, 90], [23, 24, 26, 37, 52, 54, 67, 69], [26, 44, 56, 57, 60, 66, 80, 85]]target = [[45, 5, 35, 84, 86, 88, 12, 62], [24, 64, 47, 67, 52, 62, 44, 33], [33, 75, 62, 36, 28, 38, 43, 71]]
+Epoch 0: |                                                                          | 2919/? [2:49:06<00:00,  0.29it/s, v_num=seh9]train step 2920; scene = [['f0f815ffa7581003']]; loss = 0.010755
+Epoch 0: |                                                                          | 2920/? [2:49:09<00:00,  0.29it/s, v_num=seh9]context = [[17, 39, 70, 73], [53, 88, 112, 125], [8, 29, 36, 89], [77, 81, 82, 144], [15, 20, 39, 63], [122, 124, 156, 205]]target = [[43, 67, 71, 62], [101, 68, 60, 61], [75, 83, 38, 29], [79, 112, 114, 90], [30, 33, 39, 45], [152, 201, 162, 182]]
+Epoch 0: |                                                                          | 2929/? [2:49:40<00:00,  0.29it/s, v_num=seh9]train step 2930; scene = [['f78ae8f99b60b424'], ['73aee8654106974f']]; loss = 0.014482
+Epoch 0: |                                                                          | 2930/? [2:49:43<00:00,  0.29it/s, v_num=seh9]context = [[3, 18, 30, 37, 50, 53], [162, 163, 185, 195, 198, 216], [1, 15, 42, 45, 59, 74], [6, 27, 34, 45, 51, 59]]target = [[52, 51, 49, 50, 8, 20], [204, 164, 190, 174, 170, 166], [63, 9, 16, 72, 27, 54], [38, 20, 42, 28, 46, 19]]
+Epoch 0: |                                                                          | 2939/? [2:50:15<00:00,  0.29it/s, v_num=seh9]train step 2940; scene = [['6e7a6ed3e8593e75'], ['4099447674fb0515']]; loss = 0.008432
+Epoch 0: |                                                                          | 2940/? [2:50:18<00:00,  0.29it/s, v_num=seh9]context = [[13, 20, 28, 34, 50, 54, 84, 86], [15, 16, 36, 46, 52, 59, 95, 104], [58, 64, 83, 87, 103, 112, 113, 116]]target = [[41, 54, 44, 58, 42, 29, 34, 31], [79, 32, 46, 67, 78, 33, 99, 22], [68, 95, 70, 88, 104, 59, 80, 81]]
+Epoch 0: |                                                                          | 2949/? [2:50:49<00:00,  0.29it/s, v_num=seh9]train step 2950; scene = [['2cff2a771002750a'], ['6f4cc17690dcdd2e']]; loss = 0.011415
+Epoch 0: |                                                                          | 2950/? [2:50:52<00:00,  0.29it/s, v_num=seh9]context = [[0, 3, 7, 9, 16, 19, 20, 28, 29, 30, 35, 52, 55, 57, 58, 64, 71, 78, 79, 81, 84, 87, 93, 97]]target = [[70, 51, 67, 42, 87, 35, 3, 84, 96, 89, 46, 73, 32, 66, 12, 38, 13, 5, 14, 72, 80, 33, 23, 85]]
+Epoch 0: |                                                                          | 2959/? [2:51:23<00:00,  0.29it/s, v_num=seh9]train step 2960; scene = [['e7fad36853161638'], ['abc9cfcdc52bb1ea'], ['b054fb5ce2f8e1f2'], ['6a94bfa75e7988c8'], ['94d919d55f8b5d90'], ['34fa2f4f1daa9b1b']]; loss = 0.021121
+Epoch 0: |                                                                          | 2960/? [2:51:27<00:00,  0.29it/s, v_num=seh9]context = [[1, 19, 46, 59, 62, 83], [27, 46, 61, 65, 76, 86], [9, 13, 18, 52, 65, 71], [137, 139, 141, 142, 155, 182]]target = [[8, 46, 52, 39, 66, 32], [40, 59, 47, 60, 45, 50], [66, 25, 24, 36, 49, 53], [171, 152, 173, 150, 161, 144]]
+Epoch 0: |                                                                          | 2969/? [2:51:58<00:00,  0.29it/s, v_num=seh9]train step 2970; scene = [['f8b97071c3db77f7'], ['2f42d8f78b745b8f'], ['3a19113b55068671']]; loss = 0.012370
+Epoch 0: |                                                                          | 2970/? [2:52:01<00:00,  0.29it/s, v_num=seh9]context = [[1, 23, 50, 58], [77, 96, 101, 136], [0, 31, 43, 74], [87, 127, 137, 157], [34, 74, 83, 109], [7, 38, 39, 68]]target = [[43, 48, 2, 21], [87, 86, 106, 88], [29, 34, 39, 17], [95, 114, 115, 102], [66, 63, 81, 107], [40, 58, 21, 66]]
+Epoch 0: |                                                                          | 2979/? [2:52:31<00:00,  0.29it/s, v_num=seh9]train step 2980; scene = [['41b80c5fb43019e5'], ['64554b0854be0a81'], ['10c467dbab0f5134']]; loss = 0.015415
+Epoch 0: |                                                                          | 2980/? [2:52:35<00:00,  0.29it/s, v_num=seh9]context = [[58, 67, 72, 74, 75, 76, 79, 98, 99, 107, 132, 136], [11, 13, 22, 51, 53, 54, 64, 71, 72, 82, 94, 97]]target = [[87, 130, 123, 111, 120, 86, 73, 59, 95, 126, 66, 116], [16, 91, 76, 73, 35, 12, 57, 29, 56, 15, 44, 82]]
+Epoch 0: |                                                                          | 2989/? [2:53:06<00:00,  0.29it/s, v_num=seh9]train step 2990; scene = [['5f2d539c6dcc0a82'], ['78ca16a59dd41338']]; loss = 0.010495
+Epoch 0: |                                                                          | 2990/? [2:53:10<00:00,  0.29it/s, v_num=seh9]context = [[47, 53, 59, 63, 65, 67, 68, 69, 74, 76, 85, 86, 91, 93, 102, 103, 108, 117, 122, 131, 138, 139, 140, 144]]target = [[118, 141, 73, 69, 106, 52, 81, 102, 133, 49, 79, 88, 48, 87, 124, 132, 59, 51, 140, 100, 57, 50, 103, 56]]
+Epoch 0: |                                                                          | 2999/? [2:53:41<00:00,  0.29it/s, v_num=seh9]train step 3000; scene = [['86276412bbdb6b7a'], ['9a94e71cea472790'], ['aa2197dadd5f5f8d'], ['7d900c809e896e32']]; loss = 0.015684
+Epoch 0: |                                                                          | 3000/? [2:53:44<00:00,  0.29it/s, v_num=seh9]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 3000; scene = ['1072aae07584e091'];
+target intrinsic:  tensor(0.9886, device='cuda:0') tensor(0.9889, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9476, device='cuda:0') tensor(0.9471, device='cuda:0')
+[2026-02-25 04:35:15,500][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 3000/? [2:54:05<00:00,  0.29it/s, v_num=seh9]context = [[61, 62, 64, 68, 78, 82, 84, 87, 90, 91, 96, 107, 109, 116, 117, 126, 130, 135, 137, 143, 150, 153, 157, 158]]target = [[155, 85, 107, 142, 83, 78, 68, 126, 117, 133, 125, 99, 121, 141, 97, 81, 67, 100, 153, 146, 84, 118, 86, 150]]
+[2026-02-25 04:35:19,661][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 3001/? [2:54:10<00:00,  0.29it/s, v_num=seh9]
+`Trainer.fit` stopped: `max_steps=3001` reached.
+Peak VRAM: 78.034 GB (allocated), 137.805 GB (reserved)
+Total elapsed: 2.92 hours
+Saved memory info to: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain/peak_vram_memory.json
diff --git a/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/requirements.txt b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..7fbf9096f92b53f8bb2a7e5467c79ecbe64faca5
--- /dev/null
+++ b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/requirements.txt
@@ -0,0 +1,172 @@
+wheel==0.45.1
+pytz==2025.2
+easydict==1.13
+antlr4-python3-runtime==4.9.3
+wadler_lindig==0.1.7
+urllib3==2.5.0
+tzdata==2025.2
+typing-inspection==0.4.1
+tabulate==0.9.0
+smmap==5.0.2
+kornia_rs==0.1.9
+setuptools==78.1.1
+safetensors==0.5.3
+PyYAML==6.0.2
+PySocks==1.7.1
+pyparsing==3.2.5
+pydantic_core==2.33.2
+pycparser==2.23
+protobuf==6.32.1
+propcache==0.3.2
+proglog==0.1.12
+fsspec==2024.6.1
+platformdirs==4.4.0
+pip==25.2
+pillow==10.4.0
+frozenlist==1.7.0
+packaging==24.2
+opt_einsum==3.4.0
+numpy==1.26.4
+ninja==1.13.0
+fonttools==4.60.0
+networkx==3.4.2
+multidict==6.6.4
+mdurl==0.1.2
+MarkupSafe==3.0.2
+kiwisolver==1.4.9
+imageio-ffmpeg==0.6.0
+idna==3.7
+hf-xet==1.1.10
+gmpy2==2.2.1
+einops==0.8.1
+filelock==3.17.0
+decorator==4.4.2
+dacite==1.9.2
+cycler==0.12.1
+colorama==0.4.6
+click==8.3.0
+nvidia-nvtx-cu12==12.8.90
+charset-normalizer==3.3.2
+certifi==2025.8.3
+beartype==0.19.0
+attrs==25.3.0
+async-timeout==5.0.1
+annotated-types==0.7.0
+aiohappyeyeballs==2.6.1
+yarl==1.20.1
+tifffile==2025.5.10
+sentry-sdk==2.39.0
+scipy==1.15.3
+pydantic==2.11.9
+pandas==2.3.2
+opencv-python==4.11.0.86
+omegaconf==2.3.0
+markdown-it-py==4.0.0
+lightning-utilities==0.14.3
+lazy_loader==0.4
+jaxtyping==0.2.37
+imageio==2.37.0
+gitdb==4.0.12
+contourpy==1.3.2
+colorspacious==1.1.2
+cffi==1.17.1
+aiosignal==1.4.0
+scikit-video==1.1.11
+scikit-image==0.25.2
+rich==14.1.0
+moviepy==1.0.3
+matplotlib==3.10.6
+hydra-core==1.3.2
+nvidia-nccl-cu12==2.27.3
+huggingface-hub==0.35.1
+GitPython==3.1.45
+brotlicffi==1.0.9.2
+aiohttp==3.12.15
+torchmetrics==1.8.2
+opt-einsum-fx==0.1.4
+kornia==0.8.1
+pytorch-lightning==2.5.1
+lpips==0.1.4
+e3nn==0.6.0
+lightning==2.5.1
+nvidia-cusparselt-cu12==0.7.1
+triton==3.4.0
+nvidia-nvjitlink-cu12==12.8.93
+nvidia-curand-cu12==10.3.9.90
+nvidia-cufile-cu12==1.13.1.3
+nvidia-cuda-runtime-cu12==12.8.90
+nvidia-cuda-nvrtc-cu12==12.8.93
+nvidia-cuda-cupti-cu12==12.8.90
+nvidia-cublas-cu12==12.8.4.1
+nvidia-cusparse-cu12==12.5.8.93
+nvidia-cufft-cu12==11.3.3.83
+nvidia-cudnn-cu12==9.10.2.21
+nvidia-cusolver-cu12==11.7.3.90
+torch==2.8.0+cu128
+torchvision==0.23.0+cu128
+torchaudio==2.8.0+cu128
+torch_scatter==2.1.2+pt28cu128
+gsplat==1.5.3
+wandb==0.25.0
+cuda-bindings==13.0.3
+cuda-pathfinder==1.3.3
+Jinja2==3.1.6
+mpmath==1.3.0
+nvidia-cublas==13.1.0.3
+nvidia-cuda-cupti==13.0.85
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-runtime==13.0.96
+nvidia-cudnn-cu13==9.15.1.9
+nvidia-cufft==12.0.0.61
+nvidia-cufile==1.15.1.6
+nvidia-curand==10.4.0.35
+nvidia-cusolver==12.0.4.66
+nvidia-cusparse==12.6.3.3
+nvidia-cusparselt-cu13==0.8.0
+nvidia-nccl-cu13==2.28.9
+nvidia-nvjitlink==13.0.88
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvtx==13.0.85
+requests==2.32.5
+sentencepiece==0.2.1
+sympy==1.14.0
+torchcodec==0.10.0
+torchdata==0.10.0
+torchtext==0.6.0
+anyio==4.12.0
+asttokens==3.0.1
+comm==0.2.3
+debugpy==1.8.19
+executing==2.2.1
+h11==0.16.0
+httpcore==1.0.9
+httpx==0.28.1
+ipykernel==7.1.0
+ipython==9.8.0
+ipython_pygments_lexers==1.1.1
+ipywidgets==8.1.8
+jedi==0.19.2
+jupyter_client==8.7.0
+jupyter_core==5.9.1
+jupyterlab_widgets==3.0.16
+matplotlib-inline==0.2.1
+nest-asyncio==1.6.0
+parso==0.8.5
+pexpect==4.9.0
+prompt_toolkit==3.0.52
+psutil==7.2.1
+ptyprocess==0.7.0
+pure_eval==0.2.3
+Pygments==2.19.2
+python-dateutil==2.9.0.post0
+pyzmq==27.1.0
+shellingham==1.5.4
+six==1.17.0
+stack-data==0.6.3
+tornado==6.5.4
+tqdm==4.67.1
+traitlets==5.14.3
+typer-slim==0.21.0
+typing_extensions==4.15.0
+wcwidth==0.2.14
+widgetsnbextension==4.0.15
diff --git a/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/wandb-metadata.json b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/wandb-metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..c1245bb4b063caaaae067ed3acadf38cb13bd242
--- /dev/null
+++ b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/wandb-metadata.json
@@ -0,0 +1,94 @@
+{
+  "os":  "Linux-6.8.0-90-generic-x86_64-with-glibc2.39",
+  "python":  "CPython 3.12.12",
+  "startedAt":  "2026-02-25T01:40:59.849537Z",
+  "args":  [
+    "+experiment=re10k_ablation_24v",
+    "wandb.mode=online",
+    "wandb.name=ABLATION_0225_targetTrain",
+    "train.scene_scale_reg_loss=0.0",
+    "train.train_aux=false"
+  ],
+  "program":  "-m src.main",
+  "git":  {
+    "remote":  "git@github.com:K-nowing/CVPR2026.git",
+    "commit":  "2512754c6c27ca5150bf17fbcbdde3f192fd53cc"
+  },
+  "email":  "dna9041@korea.ac.kr",
+  "root":  "/workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain",
+  "host":  "27d18dedec6d",
+  "executable":  "/venv/main/bin/python",
+  "cpu_count":  128,
+  "cpu_count_logical":  256,
+  "gpu":  "NVIDIA H200",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "1170378588160",
+      "used":  "660778004480"
+    }
+  },
+  "memory":  {
+    "total":  "1622948257792"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-2649ab80-a3a6-5a1c-0fa5-12bc11bd75e9"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-e92921d9-c681-246f-af93-637e0dc938ca"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-ffe12ffc-9bb7-82de-5692-1ec0ee2e68d8"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-499e5acd-b6ab-2010-c51b-ee9b5aa65825"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-3b2522d9-1c72-e49b-2c30-96165680b74a"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-a9a280c5-b2f9-dc1e-a8a9-7326a74001ff"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-07d0167b-a6a1-1900-2d27-7c6c11598409"
+    },
+    {
+      "name":  "NVIDIA H200",
+      "memoryTotal":  "150754820096",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-8362a999-20d1-c27b-5d18-032d23f859ab"
+    }
+  ],
+  "cudaVersion":  "13.1",
+  "writerId":  "kk43v9bdhoc229qwa3b4aitwhvhz6zhc"
+}
\ No newline at end of file
diff --git a/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/wandb-summary.json b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/wandb-summary.json
new file mode 100644
index 0000000000000000000000000000000000000000..d28062a81729a2cb4e9f5371c0722cd4747324a0
--- /dev/null
+++ b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/files/wandb-summary.json
@@ -0,0 +1 @@
+{"epoch":0,"error_scores":{"captions":["1072aae07584e091"],"_type":"images/separated","width":800,"height":536,"format":"png","count":1,"filenames":["media/images/error_scores_184_3a82fda81a7ef629684d.png"]},"loss/final_3dgs/lpips":0.007702582515776157,"lr-AdamW/pg2":2e-05,"lr-AdamW/pg1":2.003594834351718e-05,"val/ssim":0.5377284288406372,"train/psnr_probabilistic":23.780838012695312,"train/scene_scale":0.7453143000602722,"comparison":{"filenames":["media/images/comparison_182_27183732573a913be14a.png"],"captions":["1072aae07584e091"],"_type":"images/separated","width":1064,"height":1098,"format":"png","count":1},"info/global_step":3000,"loss/final_3dgs/mse":0.004881734494119883,"_wandb":{"runtime":10466},"_timestamp":1.7719941209712121e+09,"loss/final_3dgs/error_score":0.21538545191287994,"val/psnr":17.19118881225586,"active_mask_imgs":{"count":1,"filenames":["media/images/active_mask_imgs_183_56f3be674a9a787ddf05.png"],"captions":["1072aae07584e091"],"_type":"images/separated","width":536,"height":800,"format":"png"},"val/gaussian_num_ratio":0.3937225341796875,"_step":186,"lr-AdamW/pg2-momentum":0.9,"val/lpips":0.3331141471862793,"_runtime":10466,"trainer/global_step":3001,"train/comparison":{"height":6378,"format":"png","count":1,"filenames":["media/images/train/comparison_186_63b8bd4de17df2033fed.png"],"captions":[["c270572a7f5ea828"]],"_type":"images/separated","width":536},"loss/camera":9.45475694607012e-05,"loss/total":0.015683647245168686,"lr-AdamW/pg1-momentum":0.9}
\ No newline at end of file
diff --git a/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug-core.log b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug-core.log
new file mode 100644
index 0000000000000000000000000000000000000000..ffe433147c5059cafa47269cb098276eeabc8a4d
--- /dev/null
+++ b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug-core.log
@@ -0,0 +1,15 @@
+{"time":"2026-02-25T01:40:59.917551837Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpgaq4tpjj/port-121656.txt","pid":121656,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-02-25T01:40:59.9185115Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":121656}
+{"time":"2026-02-25T01:40:59.918487499Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-121656-124088-1516356927/socket","Net":"unix"}}
+{"time":"2026-02-25T01:41:00.093328619Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-02-25T01:41:00.10359478Z","level":"INFO","msg":"handleInformInit: received","streamId":"qetzseh9","id":"1(@)"}
+{"time":"2026-02-25T01:41:00.515572336Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"qetzseh9","id":"1(@)"}
+{"time":"2026-02-25T01:41:07.147180846Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"n8r4dtceurnb"}
+{"time":"2026-02-25T04:35:28.261259207Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
+{"time":"2026-02-25T04:35:28.261349768Z","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-02-25T04:35:28.261324928Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
+{"time":"2026-02-25T04:35:28.261493701Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
+{"time":"2026-02-25T04:35:28.261493701Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-121656-124088-1516356927/socket","Net":"unix"}}
+{"time":"2026-02-25T04:35:29.936148035Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
+{"time":"2026-02-25T04:35:29.936192165Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
+{"time":"2026-02-25T04:35:29.936216326Z","level":"INFO","msg":"server is closed"}
diff --git a/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug-internal.log b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..c06eded7785be3e23e697052114b1fdf840b3371
--- /dev/null
+++ b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug-internal.log
@@ -0,0 +1,11 @@
+{"time":"2026-02-25T01:41:00.103713472Z","level":"INFO","msg":"stream: starting","core version":"0.25.0"}
+{"time":"2026-02-25T01:41:00.515197271Z","level":"INFO","msg":"stream: created new stream","id":"qetzseh9"}
+{"time":"2026-02-25T01:41:00.515336753Z","level":"INFO","msg":"handler: started","stream_id":"qetzseh9"}
+{"time":"2026-02-25T01:41:00.515557835Z","level":"INFO","msg":"stream: started","id":"qetzseh9"}
+{"time":"2026-02-25T01:41:00.515605436Z","level":"INFO","msg":"writer: started","stream_id":"qetzseh9"}
+{"time":"2026-02-25T01:41:00.515647397Z","level":"INFO","msg":"sender: started","stream_id":"qetzseh9"}
+{"time":"2026-02-25T04:35:28.261315178Z","level":"INFO","msg":"stream: closing","id":"qetzseh9"}
+{"time":"2026-02-25T04:35:29.641674674Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-02-25T04:35:29.934574479Z","level":"INFO","msg":"handler: closed","stream_id":"qetzseh9"}
+{"time":"2026-02-25T04:35:29.934901185Z","level":"INFO","msg":"sender: closed","stream_id":"qetzseh9"}
+{"time":"2026-02-25T04:35:29.934924995Z","level":"INFO","msg":"stream: closed","id":"qetzseh9"}
diff --git a/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug.log b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..e5fc69f9b164886e9cc9a2671e28569335bcb3fd
--- /dev/null
+++ b/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug.log
@@ -0,0 +1,21 @@
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_setup.py:_flush():81] Current SDK version is 0.25.0
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_setup.py:_flush():81] Configure stats pid to 121656
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug.log
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain/wandb/run-20260225_014059-qetzseh9/logs/debug-internal.log
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_init.py:init():844] calling init triggers
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
+config: {'model': {'encoder': {'name': 'dcsplat', 'input_image_shape': [518, 518], 'head_mode': 'pcd', 'num_level': 3, 'gs_param_dim': 256, 'align_corners': False, 'use_voxelize': True}, 'decoder': {'name': 'splatting_cuda', 'background_color': [0.0, 0.0, 0.0], 'make_scale_invariant': False}, 'density_control': {'name': 'density_control_module', 'mean_dim': 32, 'gs_param_dim': 256, 'refinement_layer_num': 1, 'num_level': 3, 'grad_mode': 'absgrad', 'use_mean_features': True, 'refinement_type': 'voxelize', 'refinement_hidden_dim': 32, 'aggregation_mode': 'mean', 'num_heads': 1, 'score_mode': 'absgrad', 'latent_dim': 128, 'num_latents': 64, 'num_self_attn_per_block': 2, 'voxel_size': 0.001, 'aux_refine': False, 'refine_error': False, 'use_refine_module': True, 'voxelize_activate': True, 'use_depth': False}}, 'render_loss': {'mse': {'weight': 1.0}, 'lpips': {'weight': 0.05, 'apply_after_step': 0}}, 'density_control_loss': {'error_score': {'weight': 0.01, 'log_scale': False, 'grad_scale': 10000.0, 'mode': 'original'}}, 'direct_loss': {'l1': {'weight': 0.8}, 'ssim': {'weight': 0.2}}, 'wandb': {'project': 'DCSplat', 'entity': 'scene-representation-group', 'name': 'ABLATION_0225_targetTrain', 'mode': 'online', 'tags': ['re10k', '256x256']}, 'mode': 'train', 'data_loader': {'train': {'num_workers': 16, 'persistent_workers': True, 'batch_size': 16, 'seed': 1234}, 'test': {'num_workers': 4, 'persistent_workers': False, 'batch_size': 1, 'seed': 2345}, 'val': {'num_workers': 1, 'persistent_workers': True, 'batch_size': 1, 'seed': 3456}}, 'optimizer': {'lr': 0.0002, 'warm_up_steps': 25, 'backbone_lr_multiplier': 0.1, 'backbone_trainable': 'T+H', 'accumulate': 1}, 'checkpointing': {'load': None, 'every_n_train_steps': 1500, 'save_top_k': 2, 'save_weights_only': False}, 'train': {'extended_visualization': False, 'print_log_every_n_steps': 10, 'camera_loss': 10.0, 'one_sample_validation': None, 'align_corners': False, 'intrinsic_scaling': False, 'verbose': False, 'beta_dist_param': [0.5, 4.0], 'use_refine_aux': False, 'train_target_set': True, 'train_gs_num': 1, 'ext_scale_detach': False, 'cam_scale_mode': 'sum', 'scene_scale_reg_loss': 0.0, 'train_aux': False, 'vggt_cam_loss': True, 'vggt_distil': False, 'context_view_train': False}, 'test': {'output_path': 'test/ablation/re10k', 'align_pose': False, 'pose_align_steps': 100, 'rot_opt_lr': 0.005, 'trans_opt_lr': 0.005, 'compute_scores': True, 'save_image': False, 'save_video': False, 'save_active_mask_image': False, 'save_error_score_image': False, 'save_compare': False, 'pred_intrinsic': False, 'error_threshold': 0.4, 'error_threshold_list': [0.2, 0.4, 0.6, 0.8, 1.0], 'threshold_mode': 'ratio', 'nvs_view_N_list': [3, 6, 16, 32, 64]}, 'seed': 111123, 'trainer': {'max_steps': 3001, 'val_check_interval': 250, 'gradient_clip_val': 0.5, 'num_nodes': 1}, 'dataset': {'re10k': {'make_baseline_1': True, 'relative_pose': True, 'augment': True, 'background_color': [0.0, 0.0, 0.0], 'overfit_to_scene': None, 'skip_bad_shape': True, 'view_sampler': {'name': 'bounded', 'num_target_views': 4, 'num_context_views': 2, 'min_distance_between_context_views': 45, 'max_distance_between_context_views': 90, 'min_distance_to_context_views': 0, 'warm_up_steps': 1000, 'initial_min_distance_between_context_views': 25, 'initial_max_distance_between_context_views': 25, 'same_target_gap': False, 'num_target_set': 3}, 'name': 're10k', 'roots': ['datasets/re10k'], 'input_image_shape': [256, 256], 'original_image_shape': [360, 640], 'cameras_are_circular': False, 'baseline_min': 0.001, 'baseline_max': 10000000000.0, 'max_fov': 100.0, 'dynamic_context_views': True, 'max_context_views_per_gpu': 24}}, '_wandb': {}}
+2026-02-25 01:40:59,851 INFO    MainThread:121656 [wandb_init.py:init():892] starting backend
+2026-02-25 01:41:00,093 INFO    MainThread:121656 [wandb_init.py:init():895] sending inform_init request
+2026-02-25 01:41:00,100 INFO    MainThread:121656 [wandb_init.py:init():903] backend started and connected
+2026-02-25 01:41:00,102 INFO    MainThread:121656 [wandb_init.py:init():973] updated telemetry
+2026-02-25 01:41:00,106 INFO    MainThread:121656 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
+2026-02-25 01:41:02,034 INFO    MainThread:121656 [wandb_init.py:init():1042] starting run threads in backend
+2026-02-25 01:41:02,143 INFO    MainThread:121656 [wandb_run.py:_console_start():2524] atexit reg
+2026-02-25 01:41:02,143 INFO    MainThread:121656 [wandb_run.py:_redirect():2373] redirect: wrap_raw
+2026-02-25 01:41:02,143 INFO    MainThread:121656 [wandb_run.py:_redirect():2442] Wrapping output streams.
+2026-02-25 01:41:02,143 INFO    MainThread:121656 [wandb_run.py:_redirect():2465] Redirects installed.
+2026-02-25 01:41:02,146 INFO    MainThread:121656 [wandb_init.py:init():1082] run started, returning control to user process
+2026-02-25 04:35:28,261 INFO    wandb-AsyncioManager-main:121656 [service_client.py:_forward_responses():134] Reached EOF.
+2026-02-25 04:35:28,261 INFO    wandb-AsyncioManager-main:121656 [mailbox.py:close():155] Closing mailbox, abandoning 1 handles.
diff --git a/ABLATION_0225_targetTrain_SSR/.hydra/config.yaml b/ABLATION_0225_targetTrain_SSR/.hydra/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..c342e11832d15291d463b4f3ec28e94a8a55722a
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/.hydra/config.yaml
@@ -0,0 +1,185 @@
+model:
+  encoder:
+    name: dcsplat
+    input_image_shape:
+    - 518
+    - 518
+    head_mode: pcd
+    num_level: 3
+    gs_param_dim: 256
+    align_corners: false
+    use_voxelize: true
+  decoder:
+    name: splatting_cuda
+    background_color:
+    - 0.0
+    - 0.0
+    - 0.0
+    make_scale_invariant: false
+  density_control:
+    name: density_control_module
+    mean_dim: 32
+    gs_param_dim: 256
+    refinement_layer_num: 1
+    num_level: 3
+    grad_mode: absgrad
+    use_mean_features: true
+    refinement_type: voxelize
+    refinement_hidden_dim: 32
+    aggregation_mode: mean
+    num_heads: 1
+    score_mode: absgrad
+    latent_dim: 128
+    num_latents: 64
+    num_self_attn_per_block: 2
+    voxel_size: 0.001
+    aux_refine: false
+    refine_error: false
+    use_refine_module: true
+    voxelize_activate: true
+    use_depth: false
+render_loss:
+  mse:
+    weight: 1.0
+  lpips:
+    weight: 0.05
+    apply_after_step: 0
+density_control_loss:
+  error_score:
+    weight: 0.01
+    log_scale: false
+    grad_scale: 10000.0
+    mode: original
+direct_loss:
+  l1:
+    weight: 0.8
+  ssim:
+    weight: 0.2
+wandb:
+  project: DCSplat
+  entity: scene-representation-group
+  name: ABLATION_0225_targetTrain_SSR
+  mode: online
+  tags:
+  - re10k
+  - 256x256
+mode: train
+data_loader:
+  train:
+    num_workers: 16
+    persistent_workers: true
+    batch_size: 16
+    seed: 1234
+  test:
+    num_workers: 4
+    persistent_workers: false
+    batch_size: 1
+    seed: 2345
+  val:
+    num_workers: 1
+    persistent_workers: true
+    batch_size: 1
+    seed: 3456
+optimizer:
+  lr: 0.0002
+  warm_up_steps: 25
+  backbone_lr_multiplier: 0.1
+  backbone_trainable: T+H
+  accumulate: 1
+checkpointing:
+  load: null
+  every_n_train_steps: 1500
+  save_top_k: 2
+  save_weights_only: false
+train:
+  extended_visualization: false
+  print_log_every_n_steps: 10
+  camera_loss: 10.0
+  one_sample_validation: null
+  align_corners: false
+  intrinsic_scaling: false
+  verbose: false
+  beta_dist_param:
+  - 0.5
+  - 4.0
+  use_refine_aux: false
+  train_target_set: true
+  train_gs_num: 1
+  ext_scale_detach: false
+  cam_scale_mode: sum
+  scene_scale_reg_loss: 0.01
+  train_aux: false
+  vggt_cam_loss: true
+  vggt_distil: false
+  context_view_train: false
+test:
+  output_path: test/ablation/re10k
+  align_pose: false
+  pose_align_steps: 100
+  rot_opt_lr: 0.005
+  trans_opt_lr: 0.005
+  compute_scores: true
+  save_image: false
+  save_video: false
+  save_active_mask_image: false
+  save_error_score_image: false
+  save_compare: false
+  pred_intrinsic: false
+  error_threshold: 0.4
+  error_threshold_list:
+  - 0.2
+  - 0.4
+  - 0.6
+  - 0.8
+  - 1.0
+  threshold_mode: ratio
+  nvs_view_N_list:
+  - 3
+  - 6
+  - 16
+  - 32
+  - 64
+seed: 111123
+trainer:
+  max_steps: 3001
+  val_check_interval: 250
+  gradient_clip_val: 0.5
+  num_nodes: 1
+dataset:
+  re10k:
+    make_baseline_1: true
+    relative_pose: true
+    augment: true
+    background_color:
+    - 0.0
+    - 0.0
+    - 0.0
+    overfit_to_scene: null
+    skip_bad_shape: true
+    view_sampler:
+      name: bounded
+      num_target_views: 4
+      num_context_views: 2
+      min_distance_between_context_views: 45
+      max_distance_between_context_views: 90
+      min_distance_to_context_views: 0
+      warm_up_steps: 1000
+      initial_min_distance_between_context_views: 25
+      initial_max_distance_between_context_views: 25
+      same_target_gap: false
+      num_target_set: 3
+    name: re10k
+    roots:
+    - datasets/re10k
+    input_image_shape:
+    - 256
+    - 256
+    original_image_shape:
+    - 360
+    - 640
+    cameras_are_circular: false
+    baseline_min: 0.001
+    baseline_max: 10000000000.0
+    max_fov: 100.0
+    dynamic_context_views: true
+    max_context_views_per_gpu: 24
diff --git a/ABLATION_0225_targetTrain_SSR/.hydra/overrides.yaml b/ABLATION_0225_targetTrain_SSR/.hydra/overrides.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..11e96a08d5e7a257baad4deec4c88a752b193338
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/.hydra/overrides.yaml
@@ -0,0 +1,4 @@
+- +experiment=re10k_ablation_24v
+- wandb.mode=online
+- wandb.name=ABLATION_0225_targetTrain_SSR
+- train.train_aux=false
diff --git a/ABLATION_0225_targetTrain_SSR/train_ddp_process_2.log b/ABLATION_0225_targetTrain_SSR/train_ddp_process_2.log
new file mode 100644
index 0000000000000000000000000000000000000000..226307fc3472615705c169b1ac4bd0d24e2c3b99
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/train_ddp_process_2.log
@@ -0,0 +1,66 @@
+[2026-02-25 04:35:57,812][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 04:36:27,679][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 04:36:27,679][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 04:36:41,642][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 04:36:54,726][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [57, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [57, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 04:36:54,833][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 04:38:19,474][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 04:48:25,990][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:00:01,620][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:11:31,315][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:23:05,955][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:34:39,775][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:46:15,971][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:57:53,153][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:09:43,025][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:21:18,230][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:32:53,993][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:44:28,348][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:56:07,074][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:07:47,162][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:19:25,492][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:31:15,203][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_targetTrain_SSR/train_ddp_process_4.log b/ABLATION_0225_targetTrain_SSR/train_ddp_process_4.log
new file mode 100644
index 0000000000000000000000000000000000000000..21036055a6d60dc570de4839cf39349b36a5fa2e
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/train_ddp_process_4.log
@@ -0,0 +1,66 @@
+[2026-02-25 04:35:57,758][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 04:36:17,388][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 04:36:17,389][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 04:36:41,642][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 04:36:54,720][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [57, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [57, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 04:36:54,831][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 04:38:19,450][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 04:48:25,988][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:00:01,622][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:11:31,315][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:23:05,955][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:34:39,775][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:46:15,973][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:57:53,153][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:09:43,025][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:21:18,231][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:32:53,995][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:44:28,348][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:56:07,074][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:07:47,163][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:19:25,492][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:31:15,203][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_targetTrain_SSR/train_ddp_process_7.log b/ABLATION_0225_targetTrain_SSR/train_ddp_process_7.log
new file mode 100644
index 0000000000000000000000000000000000000000..7f731315065b35c179c4ac6dd554251ba5bff44f
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/train_ddp_process_7.log
@@ -0,0 +1,66 @@
+[2026-02-25 04:35:57,716][dinov2][INFO] - using MLP layer as FFN
+[2026-02-25 04:36:27,252][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 04:36:27,252][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+[2026-02-25 04:36:41,642][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. 
+  warnings.warn(  # warn only once
+
+[2026-02-25 04:36:54,221][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [57, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [57, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 04:36:54,830][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 04:38:19,480][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+[2026-02-25 04:48:25,989][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:00:01,622][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:11:31,316][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:23:05,954][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:34:39,775][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:46:15,971][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 05:57:53,153][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:09:43,026][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:21:18,231][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:32:53,993][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:44:28,348][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 06:56:07,074][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:07:47,164][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:19:25,494][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 07:31:15,203][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
diff --git a/ABLATION_0225_targetTrain_SSR/wandb/debug-internal.log b/ABLATION_0225_targetTrain_SSR/wandb/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..23b5edfa711dc7262f610c0b8f4f02b4d40c1dff
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/wandb/debug-internal.log
@@ -0,0 +1,11 @@
+{"time":"2026-02-25T04:36:37.632656448Z","level":"INFO","msg":"stream: starting","core version":"0.25.0"}
+{"time":"2026-02-25T04:36:38.083360107Z","level":"INFO","msg":"stream: created new stream","id":"hluaxp6d"}
+{"time":"2026-02-25T04:36:38.083494289Z","level":"INFO","msg":"handler: started","stream_id":"hluaxp6d"}
+{"time":"2026-02-25T04:36:38.083807484Z","level":"INFO","msg":"stream: started","id":"hluaxp6d"}
+{"time":"2026-02-25T04:36:38.083822184Z","level":"INFO","msg":"writer: started","stream_id":"hluaxp6d"}
+{"time":"2026-02-25T04:36:38.083845724Z","level":"INFO","msg":"sender: started","stream_id":"hluaxp6d"}
+{"time":"2026-02-25T07:31:23.586161027Z","level":"INFO","msg":"stream: closing","id":"hluaxp6d"}
+{"time":"2026-02-25T07:31:24.4560367Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-02-25T07:31:24.656085012Z","level":"INFO","msg":"handler: closed","stream_id":"hluaxp6d"}
+{"time":"2026-02-25T07:31:24.656271365Z","level":"INFO","msg":"sender: closed","stream_id":"hluaxp6d"}
+{"time":"2026-02-25T07:31:24.656292415Z","level":"INFO","msg":"stream: closed","id":"hluaxp6d"}
diff --git a/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/files/output.log b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/files/output.log
new file mode 100644
index 0000000000000000000000000000000000000000..e2a0b54e3ee6c8ff0ede0f0703cfab5285e75094
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/files/output.log
@@ -0,0 +1,800 @@
+LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0,1,2,3,4,5,6,7]
+
+  | Name                   | Type                 | Params | Mode
+------------------------------------------------------------------------
+0 | encoder                | OurSplat             | 888 M  | train
+1 | density_control_module | DensityControlModule | 2.6 M  | train
+2 | decoder                | DecoderSplattingCUDA | 0      | train
+3 | render_losses          | ModuleList           | 0      | train
+4 | density_control_losses | ModuleList           | 0      | train
+5 | direct_losses          | ModuleList           | 0      | train
+------------------------------------------------------------------------
+891 M     Trainable params
+0         Non-trainable params
+891 M     Total params
+3,564.328 Total estimated model params size (MB)
+1231      Modules in train mode
+522       Modules in eval mode
+Sanity Checking: |                                                                                           | 0/? [00:00<?, ?it/s][2026-02-25 04:36:41,641][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/data_connector.py:425: The 'val_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=31` in the `DataLoader` to improve performance.
+
+[2026-02-25 04:36:41,642][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/distributed/distributed_c10d.py:4807: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user.
+  warnings.warn(  # warn only once
+
+Validation epoch start on rank 0
+Sanity Checking DataLoader 0:   0%|                                                                          | 0/1 [00:00<?, ?it/s]validation step 0; scene = ['306e2b7785657539'];
+target intrinsic:  tensor(0.8595, device='cuda:0') tensor(0.8597, device='cuda:0')
+pred intrinsic:  tensor(0.8779, device='cuda:0') tensor(0.8773, device='cuda:0')
+[rank0]:W0225 04:36:44.049000 129608 site-packages/torch/utils/cpp_extension.py:2425] TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation.
+[rank0]:W0225 04:36:44.049000 129608 site-packages/torch/utils/cpp_extension.py:2425] If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'] to specific architectures.
+[2026-02-25 04:36:44,110][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+[2026-02-25 04:36:44,120][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/utilities/data.py:79: Trying to infer the `batch_size` from an ambiguous collection. The batch size we found is 1. To avoid any miscalculations, use `self.log(..., batch_size=batch_size)`.
+
+Setting up [LPIPS] perceptual loss: trunk [vgg], v[0.1], spatial [off]
+[2026-02-25 04:36:44,120][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
+  warnings.warn(
+
+[2026-02-25 04:36:44,121][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
+  warnings.warn(msg)
+
+Loading model from: /venv/main/lib/python3.12/site-packages/lpips/weights/v0.1/vgg.pth
+[2026-02-25 04:36:45,824][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/functional.py:554: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /pytorch/aten/src/ATen/native/TensorShape.cpp:4322.)
+  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
+
+Sanity Checking DataLoader 0: 100%|██████████████████████████████████████████████████████████████████| 1/1 [00:04<00:00,  0.25it/s][2026-02-25 04:36:46,108][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/psnr', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 04:36:46,109][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/lpips', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 04:36:46,110][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/ssim', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 04:36:46,110][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('val/gaussian_num_ratio', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+[2026-02-25 04:36:46,110][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/lightning/pytorch/trainer/connectors/logger_connector/result.py:434: It is recommended to use `self.log('info/global_step', ..., sync_dist=True)` when logging on epoch level in distributed setting to accumulate the metric across devices.
+
+Epoch 0: |                                                                                                   | 0/? [00:00<?, ?it/s]context = [[34, 36, 50, 53, 54, 60, 63, 70, 76, 78, 79, 80, 81, 88, 92, 94, 102, 110, 112, 114, 122, 125, 126, 131]]target = [[126, 96, 109, 55, 99, 116, 43, 60, 113, 85, 103, 90, 130, 62, 76, 123, 35, 102, 125, 128, 98, 67, 129, 79]]
+[2026-02-25 04:36:54,723][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/autograd/graph.py:829: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed.  This is not an error, but may impair performance.
+grad.sizes() = [57, 256, 1, 1], strides() = [256, 1, 256, 256]
+bucket_view.sizes() = [57, 256, 1, 1], strides() = [256, 1, 1, 1] (Triggered internally at /pytorch/torch/csrc/distributed/c10d/reducer.cpp:334.)
+  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+
+[2026-02-25 04:36:54,811][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                               | 9/? [00:35<00:00,  0.25it/s, v_num=xp6d]train step 10; scene = [['36047ec1694f9d49'], ['eac6cedeba1f720b']]; loss = 0.105256
+Epoch 0: |                                                                              | 10/? [00:39<00:00,  0.25it/s, v_num=xp6d]context = [[241, 243, 246, 266], [36, 42, 60, 61], [91, 100, 108, 116], [19, 27, 28, 44], [9, 11, 19, 34], [19, 24, 42, 44]]target = [[244, 251, 250, 260], [52, 41, 43, 49], [113, 114, 115, 93], [30, 40, 20, 26], [20, 13, 24, 28], [37, 24, 22, 39]]
+Epoch 0: |                                                                              | 19/? [01:11<00:00,  0.27it/s, v_num=xp6d]train step 20; scene = [['ce3f0dbbab9e2619'], ['35f1002ddc4fbebf'], ['37b468e1f381e86f'], ['bcd63850409a4c42'], ['91ced81093fc5294'], ['7544731be485052b'], ['e4d41618c44cfb3a'], ['f6144e0803cf99db']]; loss = 0.098995
+Epoch 0: |                                                                              | 20/? [01:14<00:00,  0.27it/s, v_num=xp6d]context = [[33, 35, 49, 53, 55, 56, 62, 66], [55, 60, 70, 71, 80, 83, 87, 88], [1, 4, 8, 9, 10, 15, 24, 34]]target = [[49, 36, 48, 65, 57, 60, 50, 41], [60, 82, 78, 73, 67, 77, 87, 61], [3, 20, 13, 9, 14, 17, 12, 7]]
+Epoch 0: |                                                                              | 24/? [01:27<00:00,  0.27it/s, v_num=xp6d][2026-02-25 04:38:19,468][py.warnings][WARNING] - /venv/main/lib/python3.12/site-packages/torch/optim/lr_scheduler.py:209: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
+  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
+
+Epoch 0: |                                                                              | 29/? [01:45<00:00,  0.28it/s, v_num=xp6d]train step 30; scene = [['1e7c432d2207b6f2'], ['c495f01f294333ee'], ['0636f8d5854771ca'], ['8ef87b7c44d0d34f']]; loss = 0.045924
+Epoch 0: |                                                                              | 30/? [01:48<00:00,  0.28it/s, v_num=xp6d]context = [[10, 14, 16, 19, 20, 24, 33, 34, 42, 48, 51, 59], [54, 60, 62, 65, 71, 76, 77, 80, 91, 92, 98, 103]]target = [[14, 34, 26, 32, 11, 44, 29, 20, 15, 27, 52, 35], [97, 77, 72, 62, 61, 102, 67, 101, 89, 58, 81, 88]]
+Epoch 0: |                                                                              | 39/? [02:20<00:00,  0.28it/s, v_num=xp6d]train step 40; scene = [['3931c6a7e69ffff1']]; loss = 0.035875
+Epoch 0: |                                                                              | 40/? [02:24<00:00,  0.28it/s, v_num=xp6d]context = [[75, 90, 91, 94, 96, 100], [89, 97, 103, 109, 113, 114], [36, 39, 42, 57, 59, 62], [144, 147, 150, 151, 159, 170]]target = [[90, 79, 86, 92, 99, 85], [106, 90, 112, 96, 104, 110], [52, 44, 56, 49, 55, 41], [153, 149, 167, 166, 150, 147]]
+Epoch 0: |                                                                              | 49/? [02:55<00:00,  0.28it/s, v_num=xp6d]train step 50; scene = [['a559ed659c3d87e2']]; loss = 0.042900
+Epoch 0: |                                                                              | 50/? [02:59<00:00,  0.28it/s, v_num=xp6d]context = [[171, 177, 196], [2, 12, 29], [53, 67, 80], [160, 161, 185], [234, 237, 260], [5, 27, 32], [139, 140, 164], [4, 30, 31]]target = [[192, 188, 187], [4, 21, 24], [67, 79, 58], [161, 177, 163], [239, 258, 244], [13, 31, 25], [140, 163, 142], [14, 9, 20]]
+Epoch 0: |                                                                              | 59/? [03:30<00:00,  0.28it/s, v_num=xp6d]train step 60; scene = [['db139805c0d52235']]; loss = 0.031063
+Epoch 0: |                                                                              | 60/? [03:33<00:00,  0.28it/s, v_num=xp6d]context = [[21, 31, 32, 42, 43, 48, 49, 52, 56, 58, 59, 60, 62, 66, 72, 78, 88, 91, 93, 99, 100, 102, 105, 118]]target = [[106, 62, 43, 74, 92, 26, 38, 65, 85, 100, 68, 70, 81, 98, 115, 32, 107, 59, 110, 96, 77, 116, 114, 35]]
+Epoch 0: |                                                                              | 69/? [04:04<00:00,  0.28it/s, v_num=xp6d]train step 70; scene = [['cd6c21656a85e9b9'], ['3fe4e74e7531c300'], ['229581725830ce75'], ['47396d5a5299873e']]; loss = 0.030921
+Epoch 0: |                                                                              | 70/? [04:08<00:00,  0.28it/s, v_num=xp6d]context = [[3, 7, 11, 14, 15, 17, 18, 21, 33, 37, 45, 52], [68, 69, 76, 82, 86, 87, 93, 98, 100, 107, 108, 117]]target = [[27, 45, 17, 44, 9, 5, 34, 30, 43, 22, 11, 31], [84, 76, 90, 97, 87, 110, 105, 111, 95, 102, 113, 116]]
+Epoch 0: |                                                                              | 79/? [04:38<00:00,  0.28it/s, v_num=xp6d]train step 80; scene = [['1683c797f3feac29'], ['1916ec563ca4d9b6']]; loss = 0.046912
+Epoch 0: |                                                                              | 80/? [04:41<00:00,  0.28it/s, v_num=xp6d]context = [[9, 17, 27, 28, 35, 37, 40, 42], [36, 37, 43, 47, 52, 61, 63, 69], [67, 71, 74, 90, 95, 96, 99, 100]]target = [[23, 36, 14, 16, 34, 28, 17, 13], [63, 37, 50, 54, 39, 42, 57, 64], [75, 77, 84, 91, 98, 88, 83, 93]]
+Epoch 0: |                                                                              | 89/? [05:12<00:00,  0.28it/s, v_num=xp6d]train step 90; scene = [['ffae0c358d55ccd6']]; loss = 0.060047
+Epoch 0: |                                                                              | 90/? [05:16<00:00,  0.28it/s, v_num=xp6d]context = [[37, 39, 40, 43, 44, 45, 49, 51, 52, 54, 55, 59, 62, 74, 80, 83, 89, 92, 95, 96, 102, 118, 133, 134]]target = [[46, 52, 108, 96, 76, 118, 59, 73, 105, 109, 83, 125, 48, 39, 111, 98, 124, 79, 130, 132, 100, 104, 131, 97]]
+Epoch 0: |                                                                              | 99/? [05:47<00:00,  0.29it/s, v_num=xp6d]train step 100; scene = [['baab02d64ec03a56'], ['d56e9c764f637e1c'], ['a6990abb3591ce5a'], ['8ad7f684b4dddb11']]; loss = 0.032833
+Epoch 0: |                                                                             | 100/? [05:49<00:00,  0.29it/s, v_num=xp6d]context = [[41, 43, 50, 59, 66, 67, 69, 74], [5, 8, 12, 14, 24, 28, 31, 38], [1, 12, 13, 15, 25, 30, 32, 34]]target = [[66, 55, 53, 56, 44, 54, 59, 57], [30, 12, 29, 19, 18, 20, 17, 32], [11, 29, 21, 8, 16, 31, 22, 23]]
+Epoch 0: |                                                                             | 109/? [06:21<00:00,  0.29it/s, v_num=xp6d]train step 110; scene = [['124a8e1a8219f20f'], ['3a40230e7d49bd72'], ['47a1772c9348c0be'], ['893b447be20a73fc']]; loss = 0.023461
+Epoch 0: |                                                                             | 110/? [06:24<00:00,  0.29it/s, v_num=xp6d]context = [[77, 81, 84, 87, 88, 95, 101, 104, 109, 111, 122, 126], [52, 66, 67, 69, 74, 76, 82, 89, 92, 98, 100, 101]]target = [[92, 107, 90, 124, 120, 84, 117, 110, 105, 97, 106, 78], [87, 65, 84, 75, 92, 60, 74, 79, 98, 71, 76, 57]]
+Epoch 0: |                                                                             | 119/? [06:56<00:00,  0.29it/s, v_num=xp6d]train step 120; scene = [['2054cdda3bb0e2fa']]; loss = 0.024278
+Epoch 0: |                                                                             | 120/? [06:59<00:00,  0.29it/s, v_num=xp6d]context = [[136, 140, 142, 144, 152, 158, 159, 164, 165, 171, 172, 173, 179, 185, 187, 190, 193, 196, 197, 204, 206, 218, 232, 233]]target = [[190, 137, 170, 210, 220, 168, 155, 172, 146, 149, 202, 183, 166, 232, 203, 177, 225, 208, 139, 180, 199, 154, 216, 161]]
+Epoch 0: |                                                                             | 129/? [07:30<00:00,  0.29it/s, v_num=xp6d]train step 130; scene = [['40c0d605c4f8c69b'], ['a8b72199cf4cf5e2'], ['b75f3820760d835c']]; loss = 0.025531
+Epoch 0: |                                                                             | 130/? [07:33<00:00,  0.29it/s, v_num=xp6d]context = [[173, 179, 180, 182, 183, 184, 187, 188, 190, 193, 194, 195, 202, 204, 207, 213, 218, 224, 225, 249, 256, 260, 267, 270]]target = [[215, 236, 199, 243, 185, 249, 266, 259, 204, 189, 228, 188, 245, 203, 182, 198, 251, 195, 261, 238, 237, 257, 240, 212]]
+Epoch 0: |                                                                             | 139/? [08:04<00:00,  0.29it/s, v_num=xp6d]train step 140; scene = [['35fcc43842bf847e'], ['633e76d7ccbeb3ed'], ['0207b0ec0cc851f6'], ['63341a860ea3a43a'], ['b4e9a9bf77f35701'], ['482ceff1527f21e1']]; loss = 0.027928
+Epoch 0: |                                                                             | 140/? [08:07<00:00,  0.29it/s, v_num=xp6d]context = [[143, 164, 173], [112, 114, 142], [144, 152, 171], [50, 77, 78], [139, 148, 170], [75, 86, 107], [7, 18, 38], [99, 125, 132]]target = [[154, 148, 172], [129, 115, 128], [151, 168, 166], [65, 61, 57], [163, 141, 160], [82, 97, 106], [33, 11, 22], [112, 111, 103]]
+Epoch 0: |                                                                             | 149/? [08:39<00:00,  0.29it/s, v_num=xp6d]train step 150; scene = [['bb2afd35d6f8a765'], ['e1da66dcfb584564'], ['fc434865599d2fe6']]; loss = 0.022252
+Epoch 0: |                                                                             | 150/? [08:42<00:00,  0.29it/s, v_num=xp6d]context = [[76, 77, 96, 98, 112, 114, 116, 117, 121, 123, 124, 125], [3, 8, 17, 22, 23, 24, 26, 41, 42, 44, 51, 52]]target = [[121, 115, 81, 79, 105, 86, 114, 122, 109, 102, 101, 116], [9, 38, 27, 24, 46, 49, 18, 44, 31, 43, 15, 12]]
+Epoch 0: |                                                                             | 159/? [09:13<00:00,  0.29it/s, v_num=xp6d]train step 160; scene = [['e726107dc0960b84']]; loss = 0.025946
+Epoch 0: |                                                                             | 160/? [09:16<00:00,  0.29it/s, v_num=xp6d]context = [[91, 98, 100, 109, 113, 118, 119, 124, 128, 131, 138, 140], [59, 61, 71, 73, 75, 81, 86, 96, 97, 100, 106, 108]]target = [[137, 105, 134, 128, 136, 101, 125, 133, 114, 107, 98, 123], [107, 95, 90, 68, 72, 98, 62, 71, 67, 76, 69, 99]]
+Epoch 0: |                                                                             | 169/? [09:48<00:00,  0.29it/s, v_num=xp6d]train step 170; scene = [['6c383c3e7ece2df7']]; loss = 0.014947
+Epoch 0: |                                                                             | 170/? [09:51<00:00,  0.29it/s, v_num=xp6d]context = [[4, 5, 7, 12, 16, 19, 36, 39, 41, 47, 50, 51, 54, 58, 60, 63, 69, 71, 74, 75, 76, 86, 91, 101]]target = [[34, 93, 44, 16, 49, 25, 12, 28, 94, 48, 11, 58, 8, 76, 80, 57, 40, 13, 88, 39, 92, 41, 30, 61]]
+Epoch 0: |                                                                             | 179/? [10:23<00:00,  0.29it/s, v_num=xp6d]train step 180; scene = [['a52931a53e49657f'], ['5a2447dbfc1bdfac']]; loss = 0.018196
+Epoch 0: |                                                                             | 180/? [10:26<00:00,  0.29it/s, v_num=xp6d]context = [[86, 97, 99, 102, 109, 110, 118, 121], [152, 154, 164, 177, 181, 184, 186, 187], [29, 32, 35, 37, 42, 55, 57, 64]]target = [[100, 118, 98, 109, 94, 120, 116, 103], [162, 153, 181, 171, 159, 154, 183, 160], [33, 58, 50, 40, 35, 56, 34, 52]]
+Epoch 0: |                                                                             | 189/? [10:58<00:00,  0.29it/s, v_num=xp6d]train step 190; scene = [['fa5df7455c4df198'], ['d3136cedcdc0fc89']]; loss = 0.029244
+Epoch 0: |                                                                             | 190/? [11:01<00:00,  0.29it/s, v_num=xp6d]context = [[75, 84, 89, 102, 105, 108], [135, 136, 145, 150, 159, 164], [25, 27, 29, 56, 59, 61], [44, 51, 57, 70, 72, 73]]target = [[93, 89, 85, 105, 101, 96], [147, 137, 139, 158, 142, 159], [34, 40, 28, 38, 39, 49], [66, 49, 56, 47, 62, 70]]
+Epoch 0: |                                                                             | 199/? [11:30<00:00,  0.29it/s, v_num=xp6d]train step 200; scene = [['4d0d26a83b1768dd']]; loss = 0.024278
+Epoch 0: |                                                                             | 200/? [11:34<00:00,  0.29it/s, v_num=xp6d]context = [[13, 28, 48], [39, 63, 69], [58, 80, 92], [14, 17, 47], [1, 4, 36], [20, 44, 50], [31, 39, 65], [14, 33, 51]]target = [[18, 32, 38], [62, 60, 58], [77, 89, 82], [46, 27, 29], [4, 15, 18], [36, 23, 45], [42, 40, 45], [40, 19, 27]]
+[2026-02-25 04:48:25,988][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 209/? [12:06<00:00,  0.29it/s, v_num=xp6d]train step 210; scene = [['3ffaf1e704427fc8'], ['87194474b33de348']]; loss = 0.023494
+Epoch 0: |                                                                             | 210/? [12:09<00:00,  0.29it/s, v_num=xp6d]context = [[78, 86, 87, 94, 96, 108, 109, 119, 122, 127, 134, 139, 143, 145, 149, 150, 151, 156, 157, 159, 161, 164, 166, 175]]target = [[132, 83, 156, 168, 127, 89, 174, 159, 123, 105, 171, 141, 98, 112, 113, 137, 150, 131, 124, 109, 97, 120, 152, 133]]
+Epoch 0: |                                                                             | 219/? [12:39<00:00,  0.29it/s, v_num=xp6d]train step 220; scene = [['87491783936d3687']]; loss = 0.030345
+Epoch 0: |                                                                             | 220/? [12:42<00:00,  0.29it/s, v_num=xp6d]context = [[2, 8, 9, 34], [5, 16, 31, 37], [21, 31, 50, 52], [107, 120, 134, 142], [77, 86, 93, 114], [34, 50, 66, 67]]target = [[13, 21, 6, 16], [14, 11, 8, 28], [27, 50, 39, 43], [116, 140, 127, 141], [96, 104, 80, 91], [36, 52, 39, 66]]
+Epoch 0: |                                                                             | 229/? [13:13<00:00,  0.29it/s, v_num=xp6d]train step 230; scene = [['34c8c62d878eca66'], ['4c691165a406de40'], ['87c9128b943f1b3f'], ['8b052fd18b3b7c20'], ['4d27fb96530fe02b'], ['ee19678152518002'], ['c49bd62c183dd925'], ['da388971863189eb']]; loss = 0.025467
+Epoch 0: |                                                                             | 230/? [13:16<00:00,  0.29it/s, v_num=xp6d]context = [[0, 1, 5, 6, 10, 13, 18, 24, 32, 34, 35, 37, 48, 58, 59, 68, 70, 73, 74, 81, 89, 90, 92, 97]]target = [[53, 56, 14, 5, 44, 96, 33, 65, 40, 24, 20, 17, 36, 19, 9, 21, 2, 7, 80, 28, 73, 27, 79, 87]]
+Epoch 0: |                                                                             | 239/? [13:47<00:00,  0.29it/s, v_num=xp6d]train step 240; scene = [['86dfadf971ff9ff5'], ['68e586537c71a833'], ['9ee2ca77349564bd'], ['9794641b7e015578'], ['d8b22b4eb5e28e71'], ['be0d02ca6abeb470']]; loss = 0.038116
+Epoch 0: |                                                                             | 240/? [13:51<00:00,  0.29it/s, v_num=xp6d]context = [[46, 52, 55, 56, 64, 67, 72, 73, 74, 77, 78, 80, 84, 90, 93, 94, 95, 101, 104, 109, 112, 113, 124, 143]]target = [[47, 57, 142, 76, 92, 60, 72, 66, 71, 115, 86, 105, 119, 70, 65, 100, 95, 101, 133, 111, 116, 62, 130, 126]]
+Epoch 0: |                                                                             | 249/? [14:21<00:00,  0.29it/s, v_num=xp6d]train step 250; scene = [['0ccfac8b4f535035']]; loss = 0.015247
+Epoch 0: |                                                                             | 250/? [14:25<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 250; scene = ['49b8f80c849dc341'];
+target intrinsic:  tensor(0.8891, device='cuda:0') tensor(0.8894, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8834, device='cuda:0') tensor(0.8844, device='cuda:0')
+[2026-02-25 04:51:14,127][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 250/? [14:26<00:00,  0.29it/s, v_num=xp6d]context = [[27, 30, 31, 33, 34, 39, 40, 44, 50, 51, 72, 76], [19, 25, 30, 32, 34, 37, 39, 43, 51, 58, 63, 68]]target = [[34, 31, 43, 37, 52, 59, 54, 48, 55, 65, 44, 71], [65, 45, 51, 58, 53, 67, 57, 22, 52, 26, 46, 61]]
+Epoch 0: |                                                                             | 259/? [14:57<00:00,  0.29it/s, v_num=xp6d]train step 260; scene = [['c0e3203eaee46f03'], ['47cd9a752d3fb63f'], ['3155b05b96e2111e']]; loss = 0.013979
+Epoch 0: |                                                                             | 260/? [15:01<00:00,  0.29it/s, v_num=xp6d]context = [[41, 43, 45, 49, 59, 60, 68, 73, 77, 83, 89, 90], [9, 12, 15, 16, 23, 25, 27, 34, 42, 51, 53, 58]]target = [[84, 73, 81, 76, 55, 42, 62, 80, 68, 83, 47, 49], [45, 34, 54, 32, 49, 37, 17, 12, 18, 11, 16, 27]]
+Epoch 0: |                                                                             | 269/? [15:32<00:00,  0.29it/s, v_num=xp6d]train step 270; scene = [['e7a85a4315af4cc1']]; loss = 0.020629
+Epoch 0: |                                                                             | 270/? [15:36<00:00,  0.29it/s, v_num=xp6d]context = [[5, 17, 20, 21, 27, 30, 31, 35, 36, 41, 47, 50, 53, 57, 71, 75, 77, 87, 89, 90, 95, 99, 100, 102]]target = [[92, 95, 40, 64, 67, 37, 31, 47, 42, 55, 51, 33, 53, 59, 68, 9, 50, 93, 46, 71, 14, 44, 45, 28]]
+Epoch 0: |                                                                             | 279/? [16:08<00:00,  0.29it/s, v_num=xp6d]train step 280; scene = [['1b778f72bbee1f27']]; loss = 0.019173
+Epoch 0: |                                                                             | 280/? [16:11<00:00,  0.29it/s, v_num=xp6d]context = [[70, 72, 76, 90, 100, 112], [34, 40, 47, 48, 60, 64], [80, 89, 97, 105, 112, 115], [0, 2, 13, 24, 28, 36]]target = [[96, 110, 76, 95, 71, 75], [58, 40, 55, 51, 49, 35], [104, 96, 81, 110, 85, 97], [11, 35, 26, 21, 10, 28]]
+Epoch 0: |                                                                             | 289/? [16:41<00:00,  0.29it/s, v_num=xp6d]train step 290; scene = [['e7a04da784bddf9b'], ['5e9846591aa4776c'], ['b1d7d8d0b587ffc4']]; loss = 0.021174
+Epoch 0: |                                                                             | 290/? [16:45<00:00,  0.29it/s, v_num=xp6d]context = [[34, 37, 60, 61, 63, 73], [18, 22, 35, 37, 39, 52], [22, 23, 25, 35, 37, 61], [13, 22, 33, 38, 42, 44]]target = [[53, 55, 41, 47, 42, 66], [38, 43, 51, 41, 40, 39], [53, 33, 46, 57, 44, 36], [27, 24, 35, 31, 23, 26]]
+Epoch 0: |                                                                             | 299/? [17:17<00:00,  0.29it/s, v_num=xp6d]train step 300; scene = [['58ae98561b032336'], ['a942015dc7100bd5'], ['cc68efcc2b123690'], ['120391d54ad880a1']]; loss = 0.016904
+Epoch 0: |                                                                             | 300/? [17:20<00:00,  0.29it/s, v_num=xp6d]context = [[43, 55, 58, 61, 63, 65, 68, 70, 79, 83, 91, 92], [96, 98, 101, 105, 108, 117, 118, 123, 132, 141, 143, 145]]target = [[82, 45, 49, 79, 50, 62, 58, 61, 59, 72, 51, 44], [131, 124, 140, 108, 137, 123, 120, 122, 141, 102, 134, 143]]
+Epoch 0: |                                                                             | 309/? [17:52<00:00,  0.29it/s, v_num=xp6d]train step 310; scene = [['7e58b6857e275547']]; loss = 0.017676
+Epoch 0: |                                                                             | 310/? [17:55<00:00,  0.29it/s, v_num=xp6d]context = [[7, 13, 26, 32, 42, 49], [106, 110, 121, 124, 145, 146], [88, 99, 102, 103, 104, 127], [73, 85, 97, 99, 102, 103]]target = [[45, 46, 15, 38, 17, 41], [122, 137, 124, 126, 132, 142], [102, 117, 123, 119, 109, 125], [94, 96, 102, 77, 84, 79]]
+Epoch 0: |                                                                             | 319/? [18:27<00:00,  0.29it/s, v_num=xp6d]train step 320; scene = [['82376678602466d7'], ['3d4b9b728d676007'], ['62dbaf64a3a79445'], ['ec36aa235dc5a597']]; loss = 0.015307
+Epoch 0: |                                                                             | 320/? [18:30<00:00,  0.29it/s, v_num=xp6d]context = [[1, 3, 10, 11, 31, 37], [50, 51, 52, 71, 80, 93], [60, 64, 82, 87, 101, 103], [40, 47, 51, 67, 68, 71]]target = [[8, 20, 11, 3, 5, 14], [54, 63, 66, 91, 79, 76], [77, 96, 81, 94, 75, 85], [66, 54, 60, 64, 43, 53]]
+Epoch 0: |                                                                             | 329/? [19:02<00:00,  0.29it/s, v_num=xp6d]train step 330; scene = [['05c6dc05aaf40f45']]; loss = 0.020776
+Epoch 0: |                                                                             | 330/? [19:05<00:00,  0.29it/s, v_num=xp6d]context = [[58, 59, 61, 68, 72, 84, 89, 90, 95, 100, 103, 107], [35, 37, 41, 49, 52, 56, 58, 59, 66, 69, 80, 84]]target = [[89, 75, 61, 103, 60, 79, 62, 93, 99, 106, 81, 78], [53, 79, 64, 76, 38, 45, 49, 83, 56, 61, 69, 43]]
+Epoch 0: |                                                                             | 339/? [19:37<00:00,  0.29it/s, v_num=xp6d]train step 340; scene = [['6fceeb6dbfb1d42d'], ['89a30f540e35e364'], ['aa2c714ead9d4071']]; loss = 0.018094
+Epoch 0: |                                                                             | 340/? [19:41<00:00,  0.29it/s, v_num=xp6d]context = [[15, 54], [43, 84], [0, 31], [7, 52], [7, 51], [21, 57], [45, 89], [17, 51], [33, 69], [61, 100], [2, 35], [27, 67]]target = [[34, 28], [44, 83], [14, 7], [26, 15], [17, 21], [37, 23], [49, 71], [30, 36], [58, 41], [77, 73], [24, 23], [34, 46]]
+Epoch 0: |                                                                             | 349/? [20:12<00:00,  0.29it/s, v_num=xp6d]train step 350; scene = [['d17d6e951b1fb862'], ['dad4c7a3ecb07b58'], ['e2f3157c10aa655b'], ['7af404ed5bc4ae26']]; loss = 0.017567
+Epoch 0: |                                                                             | 350/? [20:16<00:00,  0.29it/s, v_num=xp6d]context = [[67, 78, 104, 105], [24, 28, 36, 59], [77, 81, 104, 109], [67, 71, 103, 110], [64, 82, 90, 95], [10, 33, 46, 52]]target = [[75, 83, 70, 102], [52, 37, 25, 45], [106, 100, 80, 95], [91, 106, 72, 81], [82, 70, 75, 74], [26, 30, 28, 38]]
+Epoch 0: |                                                                             | 359/? [20:47<00:00,  0.29it/s, v_num=xp6d]train step 360; scene = [['c7d35bb824ce8724'], ['fa06053c189cf9bd']]; loss = 0.025049
+Epoch 0: |                                                                             | 360/? [20:51<00:00,  0.29it/s, v_num=xp6d]context = [[8, 12, 13, 16, 19, 24, 31, 38, 46, 49, 52, 57], [34, 36, 42, 48, 50, 54, 60, 61, 62, 67, 81, 83]]target = [[46, 45, 12, 37, 35, 17, 40, 30, 49, 53, 28, 11], [68, 36, 78, 56, 82, 55, 51, 42, 54, 40, 60, 46]]
+Epoch 0: |                                                                             | 369/? [21:22<00:00,  0.29it/s, v_num=xp6d]train step 370; scene = [['a162b27e42a276da'], ['4b1ace7056c3ef7c'], ['f46fdb94ef36d8db'], ['6545c42d1fa5df33'], ['d85bf23db2a97502'], ['cc68c11bf97c66c3'], ['560c5720aa4920fd'], ['599f80ec8db49f3f']]; loss = 0.026724
+Epoch 0: |                                                                             | 370/? [21:26<00:00,  0.29it/s, v_num=xp6d]context = [[3, 4, 10, 13, 27, 28, 32, 40], [3, 6, 11, 19, 23, 32, 35, 44], [36, 42, 43, 45, 46, 54, 59, 69]]target = [[20, 18, 33, 29, 23, 35, 15, 13], [39, 15, 34, 30, 16, 20, 4, 11], [48, 46, 42, 39, 49, 57, 51, 67]]
+Epoch 0: |                                                                             | 379/? [21:57<00:00,  0.29it/s, v_num=xp6d]train step 380; scene = [['89e29a4ab957f358'], ['8dcd060a4bdf34b9'], ['23c1f863919d9a5a']]; loss = 0.016466
+Epoch 0: |                                                                             | 380/? [22:01<00:00,  0.29it/s, v_num=xp6d]context = [[141, 142, 146, 151, 159, 162, 163, 176, 177, 185, 189, 190], [23, 26, 30, 31, 33, 34, 43, 60, 65, 69, 70, 72]]target = [[182, 158, 178, 155, 176, 156, 186, 165, 188, 159, 181, 157], [40, 29, 39, 52, 51, 65, 28, 68, 37, 70, 53, 34]]
+Epoch 0: |                                                                             | 389/? [22:33<00:00,  0.29it/s, v_num=xp6d]train step 390; scene = [['298c3bccb02eb533'], ['452625cd6b071b87'], ['14900b71ac66b7bd'], ['9b8b7ae8b1327717'], ['012ae7e6ae8eb7b1'], ['ac8497e8d2e2d395']]; loss = 0.018271
+Epoch 0: |                                                                             | 390/? [22:35<00:00,  0.29it/s, v_num=xp6d]context = [[176, 183, 192, 194, 199, 214, 215, 216, 218, 220, 224, 225], [86, 88, 90, 92, 97, 103, 109, 110, 112, 118, 126, 135]]target = [[218, 220, 197, 217, 180, 216, 214, 219, 186, 210, 212, 213], [103, 101, 94, 108, 96, 105, 118, 88, 89, 123, 127, 109]]
+Epoch 0: |                                                                             | 399/? [23:07<00:00,  0.29it/s, v_num=xp6d]train step 400; scene = [['b290b6a0afa1dac7'], ['b25963fb28a6fd6a'], ['c0f67af5cd34e8d8'], ['a7c776da58a96494'], ['3b22bca6f62b1f3e'], ['99947f86c2cc4108']]; loss = 0.022146
+Epoch 0: |                                                                             | 400/? [23:10<00:00,  0.29it/s, v_num=xp6d]context = [[5, 7, 9, 19, 26, 53], [9, 15, 18, 37, 39, 47], [41, 56, 59, 63, 64, 76], [26, 29, 49, 57, 62, 65]]target = [[43, 52, 51, 29, 40, 39], [42, 31, 45, 32, 14, 37], [45, 51, 67, 60, 65, 70], [42, 41, 63, 35, 43, 54]]
+[2026-02-25 05:00:01,620][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 409/? [23:40<00:00,  0.29it/s, v_num=xp6d]train step 410; scene = [['8d8bd167d9a3ab72'], ['8c40e162a6c8015d'], ['6803b3d46e5a2c7f'], ['3eb44f0a98eaa502']]; loss = 0.017140
+Epoch 0: |                                                                             | 410/? [23:44<00:00,  0.29it/s, v_num=xp6d]context = [[59, 60, 62, 63, 68, 70, 71, 72, 81, 82, 89, 98, 106, 107, 112, 120, 121, 126, 129, 130, 142, 146, 152, 156]]target = [[70, 97, 139, 124, 103, 150, 132, 63, 127, 62, 142, 136, 90, 107, 152, 131, 108, 69, 86, 91, 128, 148, 147, 117]]
+Epoch 0: |                                                                             | 419/? [24:14<00:00,  0.29it/s, v_num=xp6d]train step 420; scene = [['f10456015e79b2b7']]; loss = 0.011576
+Epoch 0: |                                                                             | 420/? [24:18<00:00,  0.29it/s, v_num=xp6d]context = [[66, 69, 71, 73, 78, 79, 83, 86, 90, 97, 103, 110, 112, 117, 118, 120, 139, 140, 143, 144, 151, 152, 154, 163]]target = [[73, 132, 67, 122, 110, 151, 120, 89, 148, 84, 150, 126, 88, 116, 157, 138, 134, 145, 76, 105, 74, 119, 158, 121]]
+Epoch 0: |                                                                             | 429/? [24:47<00:00,  0.29it/s, v_num=xp6d]train step 430; scene = [['423716b02f72a648']]; loss = 0.013662
+Epoch 0: |                                                                             | 430/? [24:51<00:00,  0.29it/s, v_num=xp6d]context = [[110, 111, 122, 124, 125, 127, 138, 140, 142, 145, 150, 151, 152, 155, 156, 160, 170, 181, 182, 189, 194, 200, 203, 207]]target = [[165, 170, 159, 162, 180, 164, 143, 138, 203, 161, 129, 160, 168, 118, 128, 139, 111, 202, 148, 200, 204, 141, 188, 132]]
+Epoch 0: |                                                                             | 439/? [25:22<00:00,  0.29it/s, v_num=xp6d]train step 440; scene = [['a56f0d62418193a5'], ['297b64e30f27b9d2']]; loss = 0.015875
+Epoch 0: |                                                                             | 440/? [25:26<00:00,  0.29it/s, v_num=xp6d]context = [[92, 99, 106, 109, 115, 117, 120, 127, 130, 131, 139, 140, 141, 144, 148, 153, 154, 155, 169, 175, 177, 178, 179, 189]]target = [[140, 93, 165, 158, 179, 104, 95, 150, 125, 177, 122, 143, 108, 134, 139, 101, 148, 185, 117, 119, 169, 118, 159, 161]]
+Epoch 0: |                                                                             | 449/? [25:57<00:00,  0.29it/s, v_num=xp6d]train step 450; scene = [['ac2fc17a4de90b7b'], ['d8a3b176b0529293'], ['60a93ae5bb06a238'], ['53dc437d11e76ef6']]; loss = 0.020717
+Epoch 0: |                                                                             | 450/? [26:01<00:00,  0.29it/s, v_num=xp6d]context = [[3, 19, 25, 26, 28, 31, 32, 33, 37, 44, 48, 56], [225, 232, 239, 241, 255, 257, 262, 263, 264, 269, 271, 275]]target = [[20, 7, 24, 36, 23, 21, 29, 39, 50, 54, 6, 42], [268, 254, 231, 245, 248, 241, 256, 267, 265, 258, 250, 227]]
+Epoch 0: |                                                                             | 459/? [26:31<00:00,  0.29it/s, v_num=xp6d]train step 460; scene = [['02f8df567f0c3db6']]; loss = 0.126944
+Epoch 0: |                                                                             | 460/? [26:35<00:00,  0.29it/s, v_num=xp6d]context = [[18, 22, 25, 28, 35, 36, 42, 43, 48, 58, 59, 71], [65, 70, 78, 83, 85, 89, 90, 91, 93, 97, 99, 114]]target = [[37, 36, 30, 24, 32, 69, 19, 25, 52, 62, 26, 51], [86, 112, 101, 106, 71, 98, 77, 70, 87, 111, 99, 74]]
+Epoch 0: |                                                                             | 469/? [27:05<00:00,  0.29it/s, v_num=xp6d]train step 470; scene = [['a52d26a78b04aebd']]; loss = 0.012159
+Epoch 0: |                                                                             | 470/? [27:09<00:00,  0.29it/s, v_num=xp6d]context = [[27, 34, 38, 43, 48, 49, 56, 61, 74, 78, 79, 81], [0, 2, 11, 13, 16, 21, 29, 33, 40, 42, 48, 50]]target = [[58, 28, 41, 60, 49, 65, 44, 52, 36, 55, 61, 78], [20, 26, 29, 31, 36, 5, 47, 7, 3, 42, 22, 17]]
+Epoch 0: |                                                                             | 479/? [27:39<00:00,  0.29it/s, v_num=xp6d]train step 480; scene = [['3d9f44d7bdd0796d']]; loss = 0.011741
+Epoch 0: |                                                                             | 480/? [27:42<00:00,  0.29it/s, v_num=xp6d]context = [[60, 62, 75, 76, 85, 89, 104, 107], [4, 18, 26, 30, 35, 36, 39, 45], [80, 88, 100, 102, 109, 112, 115, 127]]target = [[97, 77, 98, 94, 96, 91, 105, 95], [39, 19, 29, 28, 26, 43, 41, 36], [98, 109, 94, 88, 100, 101, 87, 95]]
+Epoch 0: |                                                                             | 489/? [28:13<00:00,  0.29it/s, v_num=xp6d]train step 490; scene = [['8dd25b0a12c6a8f8'], ['6882e4590bb8ff85'], ['24f3efef10906531'], ['3bff367484a62a13'], ['62216d162b71b5b4'], ['6cb7f85c92dc6f1a'], ['dfb8f1b208a8949f'], ['0d7c1a3319b74e43']]; loss = 0.019287
+Epoch 0: |                                                                             | 490/? [28:17<00:00,  0.29it/s, v_num=xp6d]context = [[0, 1, 31, 35, 48, 55], [18, 20, 43, 47, 55, 60], [35, 58, 61, 73, 77, 80], [5, 15, 26, 29, 59, 60]]target = [[47, 35, 1, 37, 36, 28], [38, 42, 41, 36, 31, 52], [79, 42, 49, 71, 54, 72], [22, 49, 8, 41, 47, 18]]
+Epoch 0: |                                                                             | 499/? [28:49<00:00,  0.29it/s, v_num=xp6d]train step 500; scene = [['63e22bbf20853cd9'], ['bcef3076b93012b1'], ['65c3f29c43dd1e63'], ['e8d6100917c31f4c'], ['4e4314a00227ed83'], ['964a8c79fc153038'], ['3062ac95a34b9b4f'], ['3d7dfcfce85588a6']]; loss = 0.026151
+Epoch 0: |                                                                             | 500/? [28:52<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 500; scene = ['73d6f935f31b3fd4'];
+target intrinsic:  tensor(0.8576, device='cuda:0') tensor(0.8579, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8740, device='cuda:0') tensor(0.8758, device='cuda:0')
+[2026-02-25 05:05:41,307][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 500/? [28:53<00:00,  0.29it/s, v_num=xp6d]context = [[1, 3, 6, 8, 12, 32, 33, 41], [1, 19, 24, 25, 27, 29, 33, 44], [77, 79, 83, 92, 94, 103, 112, 119]]target = [[31, 8, 16, 22, 14, 27, 7, 32], [9, 33, 22, 23, 7, 19, 32, 17], [80, 101, 103, 106, 96, 97, 86, 114]]
+Epoch 0: |                                                                             | 509/? [29:24<00:00,  0.29it/s, v_num=xp6d]train step 510; scene = [['2b1f47da224557a3'], ['407f9d609bda4fb1'], ['23b99c0bdf36a8ab']]; loss = 0.016039
+Epoch 0: |                                                                             | 510/? [29:27<00:00,  0.29it/s, v_num=xp6d]context = [[184, 205, 217, 221, 222, 228], [5, 16, 19, 28, 42, 47], [0, 1, 2, 9, 11, 36], [8, 28, 31, 41, 51, 58]]target = [[214, 190, 193, 202, 199, 196], [14, 41, 8, 40, 29, 19], [28, 16, 32, 3, 33, 13], [44, 35, 34, 23, 51, 17]]
+Epoch 0: |                                                                             | 519/? [29:59<00:00,  0.29it/s, v_num=xp6d]train step 520; scene = [['9741eb65c3d6d1b0'], ['61049a4bf21ec488'], ['176ba1bba0744fe1']]; loss = 0.019990
+Epoch 0: |                                                                             | 520/? [30:02<00:00,  0.29it/s, v_num=xp6d]context = [[79, 80, 96, 113, 116, 118, 123, 124, 125, 130, 133, 134], [9, 17, 18, 23, 29, 35, 36, 39, 47, 56, 60, 65]]target = [[126, 101, 85, 92, 98, 115, 120, 119, 130, 117, 91, 105], [50, 35, 20, 21, 58, 46, 40, 11, 60, 62, 37, 15]]
+Epoch 0: |                                                                             | 529/? [30:32<00:00,  0.29it/s, v_num=xp6d]train step 530; scene = [['5a040cbc8d351ff7']]; loss = 0.029114
+Epoch 0: |                                                                             | 530/? [30:35<00:00,  0.29it/s, v_num=xp6d]context = [[20, 54, 60], [22, 26, 77], [75, 102, 125], [45, 73, 97], [197, 236, 247], [12, 26, 47], [62, 112, 113], [145, 174, 180]]target = [[37, 33, 30], [66, 56, 24], [91, 103, 88], [47, 80, 62], [209, 222, 239], [18, 45, 38], [105, 92, 110], [156, 178, 160]]
+Epoch 0: |                                                                             | 539/? [31:07<00:00,  0.29it/s, v_num=xp6d]train step 540; scene = [['5a43331e136e1666'], ['e20fa4c9c8fc8f42'], ['9d8ddcdbe1f7ac42'], ['ab2680bf91942e23']]; loss = 0.018895
+Epoch 0: |                                                                             | 540/? [31:10<00:00,  0.29it/s, v_num=xp6d]context = [[18, 21, 28, 29, 30, 31, 33, 39, 40, 43, 61, 68, 70, 72, 74, 77, 80, 83, 91, 104, 105, 108, 113, 115]]target = [[38, 62, 51, 34, 61, 58, 30, 89, 113, 56, 82, 112, 64, 60, 73, 40, 71, 41, 109, 96, 100, 97, 47, 24]]
+Epoch 0: |                                                                             | 549/? [31:41<00:00,  0.29it/s, v_num=xp6d]train step 550; scene = [['46eccfae1e09e577']]; loss = 0.013489
+Epoch 0: |                                                                             | 550/? [31:45<00:00,  0.29it/s, v_num=xp6d]context = [[23, 24, 51, 59], [63, 79, 117, 121], [31, 40, 43, 72], [210, 211, 229, 262], [108, 119, 135, 160], [33, 53, 74, 86]]target = [[34, 55, 37, 39], [92, 76, 107, 77], [36, 61, 57, 70], [233, 244, 230, 253], [150, 120, 147, 110], [85, 46, 80, 54]]
+Epoch 0: |                                                                             | 559/? [32:16<00:00,  0.29it/s, v_num=xp6d]train step 560; scene = [['afc970648a89e04a'], ['100270e080fe5b87'], ['90d87190e41314e7'], ['267277440899ef99']]; loss = 0.018804
+Epoch 0: |                                                                             | 560/? [32:20<00:00,  0.29it/s, v_num=xp6d]context = [[125, 126, 130, 136, 139, 142, 144, 149, 151, 154, 171, 176, 184, 188, 192, 193, 197, 200, 206, 207, 210, 217, 220, 222]]target = [[195, 179, 184, 167, 187, 208, 126, 166, 127, 146, 180, 134, 207, 206, 145, 181, 143, 165, 209, 218, 156, 171, 141, 190]]
+Epoch 0: |                                                                             | 569/? [32:52<00:00,  0.29it/s, v_num=xp6d]train step 570; scene = [['2bf30153b26e2060'], ['9c41c4a921687df3'], ['7f1f3578729394b9'], ['8eb8e4b1dfd6afbd']]; loss = 0.015460
+Epoch 0: |                                                                             | 570/? [32:55<00:00,  0.29it/s, v_num=xp6d]context = [[54, 56, 58, 59, 63, 74, 81, 85, 87, 90, 93, 101, 102, 103, 117, 118, 119, 127, 130, 133, 136, 140, 150, 151]]target = [[85, 91, 142, 119, 147, 80, 121, 104, 65, 105, 113, 134, 79, 90, 149, 71, 88, 102, 60, 73, 123, 64, 130, 66]]
+Epoch 0: |                                                                             | 579/? [33:27<00:00,  0.29it/s, v_num=xp6d]train step 580; scene = [['e9e24b6ec4ba54e2'], ['7ac2f82f912c0ed2'], ['d4775a73d902fab3']]; loss = 0.022987
+Epoch 0: |                                                                             | 580/? [33:31<00:00,  0.29it/s, v_num=xp6d]context = [[102, 106, 110, 111, 115, 116, 122, 131, 135, 137, 138, 146, 151, 158, 163, 165, 167, 170, 174, 176, 183, 186, 191, 199]]target = [[197, 181, 194, 189, 140, 142, 125, 162, 135, 127, 121, 129, 111, 168, 176, 179, 183, 184, 177, 108, 193, 148, 107, 134]]
+Epoch 0: |                                                                             | 589/? [34:02<00:00,  0.29it/s, v_num=xp6d]train step 590; scene = [['82b22a784b4af245'], ['f0a89e470598390d'], ['7ea719e779680555'], ['59f9c9b55158d47d']]; loss = 0.019261
+Epoch 0: |                                                                             | 590/? [34:05<00:00,  0.29it/s, v_num=xp6d]context = [[4, 14, 15, 18, 28, 32, 36, 39, 41, 42, 45, 46, 51, 59, 66, 71, 73, 75, 77, 85, 86, 90, 97, 101]]target = [[53, 55, 72, 86, 33, 97, 31, 88, 18, 25, 75, 99, 22, 81, 59, 7, 38, 90, 85, 54, 60, 32, 57, 92]]
+Epoch 0: |                                                                             | 599/? [34:36<00:00,  0.29it/s, v_num=xp6d]train step 600; scene = [['e734e244799d0bfd'], ['c62f05d391aa6ac6'], ['bc500f01a9368065'], ['8f8405a3407168ff'], ['30124191dafb3383'], ['43947a751e7c8059']]; loss = 0.027477
+Epoch 0: |                                                                             | 600/? [34:39<00:00,  0.29it/s, v_num=xp6d]context = [[33, 69], [3, 42], [37, 76], [17, 66], [4, 44], [22, 78], [41, 79], [36, 93], [97, 133], [50, 89], [19, 80], [7, 58]]target = [[63, 36], [19, 20], [41, 52], [57, 37], [29, 22], [58, 74], [67, 65], [72, 56], [122, 117], [84, 88], [22, 69], [22, 56]]
+[2026-02-25 05:11:31,315][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 609/? [35:10<00:00,  0.29it/s, v_num=xp6d]train step 610; scene = [['3bca29d56784093e']]; loss = 0.018810
+Epoch 0: |                                                                             | 610/? [35:13<00:00,  0.29it/s, v_num=xp6d]context = [[29, 34, 35, 37, 42, 44, 53, 54, 63, 72, 73, 80, 82, 84, 95, 96, 101, 102, 103, 106, 109, 112, 114, 126]]target = [[65, 35, 40, 57, 117, 30, 102, 46, 41, 112, 58, 39, 52, 114, 87, 89, 77, 31, 72, 91, 53, 59, 74, 120]]
+Epoch 0: |                                                                             | 619/? [35:45<00:00,  0.29it/s, v_num=xp6d]train step 620; scene = [['873339b3cca41d9d'], ['cbb3ec1745f4d00a'], ['a8f7946996aa6d67'], ['6093fe2362714108'], ['bcec4cead9a8da81'], ['ef685fbdf3e576a7'], ['6515de6c81e7c80c'], ['547583d5f8ae1238'], ['adde543e98ec0342'], ['bc5ebbad305b5d0f'], ['5dd26c5a45c5e40c'], ['9fde0d7ec4b06f6c']]; loss = 0.030415
+Epoch 0: |                                                                             | 620/? [35:48<00:00,  0.29it/s, v_num=xp6d]context = [[165, 181, 184, 186, 191, 195, 199, 204, 205, 212, 213, 214, 216, 217, 222, 223, 232, 237, 238, 246, 248, 256, 259, 262]]target = [[219, 247, 207, 235, 190, 229, 176, 186, 218, 237, 166, 256, 252, 233, 178, 192, 234, 226, 182, 191, 203, 212, 198, 217]]
+Epoch 0: |                                                                             | 629/? [36:20<00:00,  0.29it/s, v_num=xp6d]train step 630; scene = [['b29d69f0e6f21178']]; loss = 0.019952
+Epoch 0: |                                                                             | 630/? [36:24<00:00,  0.29it/s, v_num=xp6d]context = [[129, 130, 131, 132, 137, 147, 155, 160, 162, 181, 192, 193, 194, 195, 198, 200, 203, 210, 211, 215, 216, 219, 222, 226]]target = [[172, 169, 157, 221, 143, 186, 151, 205, 159, 219, 178, 146, 191, 194, 145, 162, 190, 222, 209, 193, 216, 225, 170, 217]]
+Epoch 0: |                                                                             | 639/? [36:55<00:00,  0.29it/s, v_num=xp6d]train step 640; scene = [['2726f059d69c25f4'], ['f95349aeab4d3dc0'], ['da963e97170fd9e9'], ['43c4c89b3414a0fc'], ['38e8301a4405eca6'], ['513aa200a2b07a1e'], ['992f739dba95f93a'], ['f8532bb5eca0e29d'], ['6e47f42627792c8a'], ['27c7362a27f418ef'], ['5874223a67b1e443'], ['00a9f110ad222aa4']]; loss = 0.041816
+Epoch 0: |                                                                             | 640/? [36:58<00:00,  0.29it/s, v_num=xp6d]context = [[11, 12, 14, 19, 20, 21, 26, 34, 35, 40, 41, 43, 45, 56, 59, 66, 72, 74, 76, 81, 85, 90, 96, 108]]target = [[61, 58, 36, 99, 89, 93, 56, 34, 38, 88, 39, 81, 32, 51, 84, 62, 48, 69, 64, 82, 29, 53, 15, 72]]
+Epoch 0: |                                                                             | 649/? [37:30<00:00,  0.29it/s, v_num=xp6d]train step 650; scene = [['9254db8c00084848'], ['467e6cb4bcb7fc53']]; loss = 0.011713
+Epoch 0: |                                                                             | 650/? [37:33<00:00,  0.29it/s, v_num=xp6d]context = [[90, 96, 97, 100, 112, 114, 117, 122, 131, 136, 140, 144, 146, 150, 161, 164, 173, 176, 177, 178, 179, 180, 186, 187]]target = [[164, 120, 180, 177, 118, 142, 122, 154, 105, 172, 128, 168, 116, 147, 119, 160, 166, 101, 169, 152, 121, 157, 130, 125]]
+Epoch 0: |                                                                             | 659/? [38:05<00:00,  0.29it/s, v_num=xp6d]train step 660; scene = [['a6d8d6545b09da76'], ['eb453d1c11da72d4'], ['46d291b62f5032b2'], ['eeb55a861130a21d']]; loss = 0.017422
+Epoch 0: |                                                                             | 660/? [38:07<00:00,  0.29it/s, v_num=xp6d]context = [[22, 33, 43, 47, 51, 53, 55, 56, 61, 71, 74, 78], [18, 31, 35, 41, 42, 43, 44, 47, 48, 67, 69, 73]]target = [[36, 65, 30, 66, 31, 27, 25, 43, 69, 76, 34, 45], [22, 37, 28, 23, 20, 19, 69, 24, 44, 72, 36, 30]]
+Epoch 0: |                                                                             | 669/? [38:39<00:00,  0.29it/s, v_num=xp6d]train step 670; scene = [['823a32785927191a'], ['ca2389f5e4fcfe61'], ['174ebd189316bd92'], ['b5260341870c7aa0']]; loss = 0.012756
+Epoch 0: |                                                                             | 670/? [38:42<00:00,  0.29it/s, v_num=xp6d]context = [[1, 5, 11, 22, 27, 29, 36, 40], [47, 67, 72, 89, 91, 94, 99, 101], [7, 16, 22, 23, 32, 38, 42, 47]]target = [[34, 2, 31, 30, 39, 35, 25, 6], [93, 58, 87, 51, 66, 92, 79, 98], [20, 29, 19, 33, 24, 21, 10, 43]]
+Epoch 0: |                                                                             | 679/? [39:13<00:00,  0.29it/s, v_num=xp6d]train step 680; scene = [['b703bdb45d172fe7'], ['d5d88e90c900def8'], ['7a20ba81fb778529'], ['bf2b8a204cdeb69f'], ['6e47af576b38cfa9'], ['c86cfcc517103af2'], ['17c29c5294714185'], ['97e7e5b8b6e493dd']]; loss = 0.028768
+Epoch 0: |                                                                             | 680/? [39:16<00:00,  0.29it/s, v_num=xp6d]context = [[2, 7, 32, 33, 35, 46, 49, 57], [45, 55, 69, 70, 74, 88, 89, 99], [60, 63, 74, 78, 81, 88, 91, 100]]target = [[39, 31, 7, 24, 45, 51, 26, 46], [78, 64, 90, 96, 73, 57, 67, 77], [81, 93, 66, 69, 84, 80, 85, 88]]
+Epoch 0: |                                                                             | 689/? [39:48<00:00,  0.29it/s, v_num=xp6d]train step 690; scene = [['43b4e4c3f5ae7c81'], ['80fdc99ef1b23262'], ['3ec9b76cab1555e4']]; loss = 0.018041
+Epoch 0: |                                                                             | 690/? [39:52<00:00,  0.29it/s, v_num=xp6d]context = [[125, 130, 131, 133, 139, 142, 153, 156, 160, 161, 166, 179], [0, 6, 18, 31, 32, 36, 50, 58, 59, 64, 65, 68]]target = [[133, 164, 151, 155, 128, 139, 142, 159, 160, 129, 174, 147], [29, 47, 32, 64, 33, 56, 63, 66, 24, 65, 8, 55]]
+Epoch 0: |                                                                             | 699/? [40:24<00:00,  0.29it/s, v_num=xp6d]train step 700; scene = [['3d30b440244efcd5'], ['8301348e86826b01']]; loss = 0.033009
+Epoch 0: |                                                                             | 700/? [40:27<00:00,  0.29it/s, v_num=xp6d]context = [[112, 118, 122, 123, 125, 131, 136, 139, 152, 154, 157, 165, 167, 168, 176, 177, 178, 186, 188, 194, 199, 203, 207, 209]]target = [[185, 195, 177, 197, 145, 155, 138, 151, 192, 130, 162, 184, 173, 113, 139, 202, 178, 198, 206, 182, 114, 169, 180, 115]]
+Epoch 0: |                                                                             | 709/? [40:58<00:00,  0.29it/s, v_num=xp6d]train step 710; scene = [['43c939b11c5fed4a']]; loss = 0.024066
+Epoch 0: |                                                                             | 710/? [41:02<00:00,  0.29it/s, v_num=xp6d]context = [[0, 3, 6, 11, 14, 27, 31, 36, 43, 46, 51, 54], [6, 8, 9, 23, 26, 28, 42, 44, 47, 60, 62, 63]]target = [[50, 44, 31, 45, 12, 7, 14, 43, 33, 11, 42, 20], [7, 27, 25, 55, 31, 51, 44, 21, 56, 60, 46, 57]]
+Epoch 0: |                                                                             | 719/? [41:32<00:00,  0.29it/s, v_num=xp6d]train step 720; scene = [['cf8fc3268c3034d6'], ['0715871c63748812']]; loss = 0.015014
+Epoch 0: |                                                                             | 720/? [41:36<00:00,  0.29it/s, v_num=xp6d]context = [[96, 103, 105, 116, 126, 128, 137, 145], [22, 31, 44, 45, 51, 52, 62, 79], [2, 19, 33, 55, 57, 58, 60, 66]]target = [[137, 129, 116, 118, 142, 119, 135, 108], [43, 55, 54, 69, 45, 38, 65, 27], [30, 28, 57, 15, 43, 16, 12, 3]]
+Epoch 0: |                                                                             | 729/? [42:06<00:00,  0.29it/s, v_num=xp6d]train step 730; scene = [['076a6c5199542c91']]; loss = 0.016440
+Epoch 0: |                                                                             | 730/? [42:10<00:00,  0.29it/s, v_num=xp6d]context = [[44, 67, 78, 90, 92, 93], [95, 98, 99, 119, 138, 140], [6, 16, 44, 64, 67, 69], [61, 89, 91, 97, 113, 131]]target = [[50, 84, 81, 90, 56, 67], [116, 108, 99, 107, 115, 103], [51, 39, 45, 53, 60, 38], [70, 67, 119, 62, 104, 75]]
+Epoch 0: |                                                                             | 739/? [42:40<00:00,  0.29it/s, v_num=xp6d]train step 740; scene = [['106fa5a7ff2a8839'], ['df378b8aa82b58d8'], ['cd6fdd63dcab9099']]; loss = 0.015074
+Epoch 0: |                                                                             | 740/? [42:44<00:00,  0.29it/s, v_num=xp6d]context = [[19, 23, 25, 28, 35, 39, 40, 44, 45, 47, 55, 62, 63, 69, 73, 76, 78, 85, 87, 92, 98, 99, 113, 116]]target = [[76, 92, 111, 107, 100, 47, 40, 87, 73, 57, 58, 66, 55, 109, 46, 88, 29, 51, 89, 69, 32, 53, 83, 93]]
+Epoch 0: |                                                                             | 749/? [43:14<00:00,  0.29it/s, v_num=xp6d]train step 750; scene = [['400098a7ab312bdc']]; loss = 0.069957
+Epoch 0: |                                                                             | 750/? [43:18<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 750; scene = ['91fda69e1cda4602'];
+target intrinsic:  tensor(0.8937, device='cuda:0') tensor(0.8939, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9100, device='cuda:0') tensor(0.9092, device='cuda:0')
+[2026-02-25 05:20:07,059][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 750/? [43:19<00:00,  0.29it/s, v_num=xp6d]context = [[105, 125, 144, 147, 160, 173], [31, 58, 61, 63, 68, 90], [11, 20, 27, 40, 42, 50], [127, 138, 144, 154, 173, 183]]target = [[167, 135, 112, 157, 107, 116], [68, 76, 58, 64, 52, 88], [18, 34, 30, 32, 27, 12], [180, 143, 138, 141, 154, 149]]
+Epoch 0: |                                                                             | 759/? [43:50<00:00,  0.29it/s, v_num=xp6d]train step 760; scene = [['882c2c3ccb2f06c3'], ['e5f6c40c0454e1aa'], ['d10eef0641b38981'], ['dd549a909b62ec24'], ['8d45251c07bebb7b'], ['8eb9e437797482f6']]; loss = 0.020692
+Epoch 0: |                                                                             | 760/? [43:53<00:00,  0.29it/s, v_num=xp6d]context = [[3, 13, 14, 21, 22, 23, 36, 44, 47, 53, 55, 56], [1, 9, 14, 15, 22, 23, 40, 41, 43, 44, 52, 63]]target = [[25, 46, 19, 38, 22, 31, 36, 8, 39, 17, 48, 28], [12, 29, 62, 39, 37, 55, 28, 56, 41, 60, 34, 48]]
+Epoch 0: |                                                                             | 769/? [44:25<00:00,  0.29it/s, v_num=xp6d]train step 770; scene = [['944e92ff3fea78eb'], ['f946d94d8adc2178']]; loss = 0.024745
+Epoch 0: |                                                                             | 770/? [44:28<00:00,  0.29it/s, v_num=xp6d]context = [[10, 17, 28, 40, 62, 64], [175, 183, 185, 199, 201, 239], [18, 21, 30, 44, 70, 84], [72, 78, 84, 94, 113, 117]]target = [[60, 55, 35, 49, 20, 17], [219, 207, 191, 196, 200, 212], [81, 66, 62, 36, 69, 70], [115, 83, 88, 116, 101, 113]]
+Epoch 0: |                                                                             | 779/? [45:00<00:00,  0.29it/s, v_num=xp6d]train step 780; scene = [['ea037bd5bd5dd020']]; loss = 0.013167
+Epoch 0: |                                                                             | 780/? [45:04<00:00,  0.29it/s, v_num=xp6d]context = [[27, 38, 49, 70], [6, 22, 76, 79], [5, 10, 24, 58], [56, 60, 78, 107], [63, 78, 89, 109], [22, 43, 73, 96]]target = [[66, 56, 58, 40], [34, 61, 42, 55], [57, 51, 26, 38], [76, 59, 103, 90], [105, 89, 100, 84], [93, 61, 63, 47]]
+Epoch 0: |                                                                             | 789/? [45:36<00:00,  0.29it/s, v_num=xp6d]train step 790; scene = [['a611f23daa8ebe85'], ['e6f06611137751ed'], ['f95e0bbb315250f4'], ['f9e4a4a27b9f0530'], ['65b40c140ed34d82'], ['643a5274a1acd8fe'], ['b1c505350a76d200'], ['db602b5c46a6037e'], ['5b7c70d9fd79a963'], ['a9cd1a8fc1fa2269'], ['111a2975a86f6e89'], ['42ac9272440cdbce']]; loss = 0.032267
+Epoch 0: |                                                                             | 790/? [45:39<00:00,  0.29it/s, v_num=xp6d]context = [[2, 8, 10, 13, 16, 18, 22, 24, 29, 32, 39, 44, 50, 54, 60, 70, 76, 79, 90, 93, 94, 95, 97, 99]]target = [[82, 57, 4, 35, 18, 6, 81, 47, 43, 97, 63, 20, 13, 56, 75, 39, 92, 51, 66, 71, 34, 86, 96, 64]]
+Epoch 0: |                                                                             | 799/? [46:11<00:00,  0.29it/s, v_num=xp6d]train step 800; scene = [['2698ac35434969d2'], ['75c89a32bc3a65d0'], ['f67633dab609f48a'], ['0c23c5d9c2010333']]; loss = 0.022041
+Epoch 0: |                                                                             | 800/? [46:15<00:00,  0.29it/s, v_num=xp6d]context = [[23, 33, 34, 41, 47, 50, 63, 67], [28, 33, 46, 64, 66, 76, 86, 90], [106, 119, 120, 134, 151, 156, 157, 158]]target = [[52, 58, 55, 61, 62, 36, 29, 31], [77, 40, 63, 79, 32, 57, 70, 42], [129, 154, 156, 119, 120, 124, 110, 157]]
+[2026-02-25 05:23:05,955][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                             | 809/? [46:45<00:00,  0.29it/s, v_num=xp6d]train step 810; scene = [['a12ec328b1ce8ed3']]; loss = 0.009147
+Epoch 0: |                                                                             | 810/? [46:49<00:00,  0.29it/s, v_num=xp6d]context = [[6, 7, 8, 12, 16, 17, 22, 33, 36, 42, 44, 47, 56, 64, 66, 74, 75, 84, 86, 90, 92, 98, 100, 103]]target = [[31, 53, 66, 12, 16, 52, 25, 8, 35, 62, 42, 43, 81, 34, 49, 79, 82, 101, 84, 59, 83, 14, 29, 69]]
+Epoch 0: |                                                                             | 819/? [47:20<00:00,  0.29it/s, v_num=xp6d]train step 820; scene = [['13e915d66f8c9bf3'], ['f1dab83bdccd9a64'], ['afba4de2c78eb96f'], ['f97a862ad842d7b4']]; loss = 0.016911
+Epoch 0: |                                                                             | 820/? [47:23<00:00,  0.29it/s, v_num=xp6d]context = [[1, 56], [10, 74], [41, 84], [7, 49], [28, 87], [21, 65], [1, 57], [0, 49], [25, 95], [16, 90], [58, 104], [114, 157]]target = [[7, 24], [63, 72], [74, 43], [24, 28], [37, 45], [52, 35], [5, 27], [9, 41], [35, 58], [73, 82], [96, 89], [152, 148]]
+Epoch 0: |                                                                             | 829/? [47:55<00:00,  0.29it/s, v_num=xp6d]train step 830; scene = [['38d4b64a2c62a10c']]; loss = 0.011600
+Epoch 0: |                                                                             | 830/? [47:58<00:00,  0.29it/s, v_num=xp6d]context = [[1, 8, 34, 37, 41, 65, 66, 70, 71, 72, 74, 75], [39, 40, 42, 43, 56, 62, 66, 92, 98, 99, 101, 102]]target = [[34, 71, 31, 37, 55, 42, 72, 70, 49, 28, 18, 58], [93, 96, 61, 60, 73, 58, 76, 80, 47, 97, 98, 49]]
+Epoch 0: |                                                                             | 839/? [48:30<00:00,  0.29it/s, v_num=xp6d]train step 840; scene = [['6e46379a7ad5ea92'], ['2abf4fde25e3836a'], ['a86ad11e2813e855']]; loss = 0.014510
+Epoch 0: |                                                                             | 840/? [48:33<00:00,  0.29it/s, v_num=xp6d]context = [[1, 8, 22, 30, 31, 42, 53, 60, 62, 66, 69, 72], [76, 79, 85, 89, 90, 92, 101, 102, 104, 113, 127, 129]]target = [[50, 65, 54, 30, 11, 7, 51, 26, 33, 28, 19, 9], [113, 116, 84, 92, 115, 89, 104, 90, 83, 78, 98, 122]]
+Epoch 0: |                                                                             | 849/? [49:03<00:00,  0.29it/s, v_num=xp6d]train step 850; scene = [['95522bc4e658c4bb'], ['a54a8ec49d0c4c52'], ['6c292d2668c5df52']]; loss = 0.021119
+Epoch 0: |                                                                             | 850/? [49:07<00:00,  0.29it/s, v_num=xp6d]context = [[210, 276], [10, 52], [100, 172], [26, 85], [15, 79], [0, 43], [9, 65], [79, 122], [7, 50], [1, 76], [10, 86], [7, 58]]target = [[271, 260], [39, 32], [117, 155], [46, 71], [24, 72], [20, 4], [49, 30], [90, 110], [19, 16], [36, 35], [81, 85], [57, 17]]
+Epoch 0: |                                                                             | 859/? [49:37<00:00,  0.29it/s, v_num=xp6d]train step 860; scene = [['056b5030399a131f'], ['066424b7f19bf6e5']]; loss = 0.011051
+Epoch 0: |                                                                             | 860/? [49:39<00:00,  0.29it/s, v_num=xp6d]context = [[18, 30, 31, 45, 47, 59, 64, 66], [0, 13, 19, 25, 31, 54, 59, 69], [5, 7, 8, 24, 28, 43, 54, 57]]target = [[19, 22, 45, 35, 25, 39, 36, 33], [2, 62, 51, 18, 68, 19, 3, 41], [52, 19, 23, 34, 28, 55, 33, 26]]
+Epoch 0: |                                                                             | 869/? [50:11<00:00,  0.29it/s, v_num=xp6d]train step 870; scene = [['2f2587735ce5a386'], ['c0f128f783625ed0']]; loss = 0.010556
+Epoch 0: |                                                                             | 870/? [50:15<00:00,  0.29it/s, v_num=xp6d]context = [[0, 14, 25, 44], [4, 17, 71, 83], [3, 8, 47, 52], [75, 84, 126, 153], [7, 9, 43, 66], [74, 102, 104, 134]]target = [[11, 38, 22, 37], [54, 23, 45, 28], [12, 50, 5, 17], [90, 135, 84, 122], [35, 31, 11, 54], [76, 104, 116, 94]]
+Epoch 0: |                                                                             | 879/? [50:46<00:00,  0.29it/s, v_num=xp6d]train step 880; scene = [['fdd01ef3e4a926df'], ['430d79082d999336'], ['3140f7bea5597af4'], ['4b4fbc022aa36b37'], ['b903f35fbb538e4d'], ['39d9f692bfb58d80'], ['e9905e5bf1c49ce7'], ['f44cc142d9796ff7']]; loss = 0.031137
+Epoch 0: |                                                                             | 880/? [50:50<00:00,  0.29it/s, v_num=xp6d]context = [[67, 119, 129, 145], [46, 82, 98, 100], [69, 88, 89, 136], [45, 74, 106, 124], [2, 30, 43, 55], [29, 57, 72, 108]]target = [[84, 122, 110, 68], [83, 61, 84, 52], [100, 83, 74, 103], [90, 62, 77, 118], [43, 36, 50, 48], [100, 77, 92, 58]]
+Epoch 0: |                                                                             | 889/? [51:22<00:00,  0.29it/s, v_num=xp6d]train step 890; scene = [['58b43b9b210b78b8']]; loss = 0.020084
+Epoch 0: |                                                                             | 890/? [51:25<00:00,  0.29it/s, v_num=xp6d]context = [[1, 7, 40, 62], [7, 14, 23, 64], [112, 129, 170, 191], [10, 11, 35, 83], [11, 12, 63, 84], [4, 25, 59, 67]]target = [[20, 43, 34, 55], [40, 29, 43, 18], [190, 173, 175, 134], [55, 73, 69, 12], [76, 45, 65, 55], [54, 15, 25, 22]]
+Epoch 0: |                                                                             | 899/? [51:56<00:00,  0.29it/s, v_num=xp6d]train step 900; scene = [['db02cd4ba6a027da'], ['7b4630c7ece2e8ab'], ['7ff18d239739a030']]; loss = 0.013643
+Epoch 0: |                                                                             | 900/? [51:59<00:00,  0.29it/s, v_num=xp6d]context = [[31, 35, 86], [20, 35, 73], [66, 69, 139], [2, 5, 58], [26, 84, 88], [94, 153, 158], [1, 4, 51], [75, 128, 131]]target = [[44, 67, 39], [23, 55, 38], [135, 89, 103], [33, 24, 6], [72, 58, 70], [119, 112, 101], [47, 13, 37], [102, 111, 88]]
+Epoch 0: |                                                                             | 909/? [52:31<00:00,  0.29it/s, v_num=xp6d]train step 910; scene = [['e0222b577fdfde97'], ['245ccba767bcd121'], ['51587e352ab35ba1'], ['9b4d466924c40d8b']]; loss = 0.025624
+Epoch 0: |                                                                             | 910/? [52:34<00:00,  0.29it/s, v_num=xp6d]context = [[4, 7, 12, 13, 16, 22, 25, 30, 34, 46, 52, 57], [9, 19, 26, 31, 40, 41, 44, 46, 51, 57, 61, 63]]target = [[36, 53, 17, 24, 22, 20, 12, 50, 51, 39, 41, 33], [27, 41, 11, 20, 28, 46, 48, 32, 44, 12, 49, 31]]
+Epoch 0: |                                                                             | 919/? [53:06<00:00,  0.29it/s, v_num=xp6d]train step 920; scene = [['42512ca51e222e21'], ['d2739b0ce00e63cf'], ['7a92426dbce9e920']]; loss = 0.014625
+Epoch 0: |                                                                             | 920/? [53:09<00:00,  0.29it/s, v_num=xp6d]context = [[4, 25, 48, 58, 65, 78], [0, 8, 11, 34, 41, 45], [13, 16, 17, 22, 33, 57], [125, 135, 140, 149, 171, 202]]target = [[62, 56, 77, 28, 24, 42], [9, 11, 16, 40, 22, 23], [31, 47, 42, 54, 32, 33], [143, 195, 128, 146, 127, 151]]
+Epoch 0: |                                                                             | 929/? [53:40<00:00,  0.29it/s, v_num=xp6d]train step 930; scene = [['c8c034439fbe43f2'], ['4b1409aed425c619'], ['bac018643bed4fe8'], ['6920a4f8f496ef7c'], ['bfc897cfc8733b10'], ['56b98fa7036d9f70'], ['cb2506c57773a6bd'], ['c7f66b7e91a9e97e']]; loss = 0.021024
+Epoch 0: |                                                                             | 930/? [53:44<00:00,  0.29it/s, v_num=xp6d]context = [[15, 23, 27, 58, 64, 80], [38, 53, 54, 66, 79, 100], [0, 19, 24, 34, 55, 65], [5, 6, 16, 28, 30, 58]]target = [[66, 61, 41, 52, 33, 59], [90, 72, 39, 42, 82, 89], [36, 23, 10, 24, 11, 51], [27, 12, 40, 21, 42, 29]]
+Epoch 0: |                                                                             | 939/? [54:14<00:00,  0.29it/s, v_num=xp6d]train step 940; scene = [['4ae003ef13ad4ffd'], ['ef8f0fd78fcb44e0']]; loss = 0.010745
+Epoch 0: |                                                                             | 940/? [54:17<00:00,  0.29it/s, v_num=xp6d]context = [[33, 35, 45, 50, 69, 70, 87, 93], [40, 53, 63, 70, 78, 96, 97, 99], [31, 37, 39, 48, 57, 61, 71, 80]]target = [[42, 58, 77, 64, 63, 39, 82, 79], [71, 59, 76, 92, 84, 95, 61, 79], [74, 32, 44, 61, 72, 64, 73, 77]]
+Epoch 0: |                                                                             | 949/? [54:48<00:00,  0.29it/s, v_num=xp6d]train step 950; scene = [['cfe7d2764a367f81']]; loss = 0.023603
+Epoch 0: |                                                                             | 950/? [54:51<00:00,  0.29it/s, v_num=xp6d]context = [[160, 166, 180, 184, 202, 212, 241, 243], [212, 213, 216, 229, 230, 247, 271, 275], [13, 14, 27, 33, 47, 62, 64, 71]]target = [[172, 241, 230, 237, 235, 164, 207, 179], [242, 224, 219, 231, 244, 260, 227, 230], [55, 69, 22, 60, 66, 37, 38, 41]]
+Epoch 0: |                                                                             | 959/? [55:22<00:00,  0.29it/s, v_num=xp6d]train step 960; scene = [['ef6401c117a2701a'], ['7f9480301fa3e38b'], ['a2ae856e2faf3097'], ['7b4006c02dddd695'], ['5b33a7567dabea55'], ['be1b98bfac2059b2']]; loss = 0.016406
+Epoch 0: |                                                                             | 960/? [55:26<00:00,  0.29it/s, v_num=xp6d]context = [[139, 146, 152, 173, 214, 224], [16, 23, 33, 48, 66, 87], [20, 26, 54, 60, 63, 103], [42, 56, 61, 71, 74, 127]]target = [[174, 145, 187, 203, 152, 159], [49, 20, 43, 26, 55, 31], [80, 30, 56, 41, 21, 81], [51, 73, 78, 70, 97, 48]]
+Epoch 0: |                                                                             | 969/? [55:58<00:00,  0.29it/s, v_num=xp6d]train step 970; scene = [['77f1b53620811084'], ['3c077d5c8af0bd36']]; loss = 0.020387
+Epoch 0: |                                                                             | 970/? [56:00<00:00,  0.29it/s, v_num=xp6d]context = [[3, 15, 17, 31, 41, 45, 63, 70, 71, 76, 89, 90], [15, 16, 29, 31, 35, 52, 69, 73, 83, 86, 94, 98]]target = [[79, 61, 68, 22, 81, 12, 29, 74, 8, 4, 69, 35], [59, 86, 45, 77, 36, 25, 51, 16, 23, 91, 82, 24]]
+Epoch 0: |                                                                             | 979/? [56:32<00:00,  0.29it/s, v_num=xp6d]train step 980; scene = [['2ff32d3c912d6d4b'], ['b7ffe15e4ef99466'], ['477b227f30b26d5f']]; loss = 0.010502
+Epoch 0: |                                                                             | 980/? [56:36<00:00,  0.29it/s, v_num=xp6d]context = [[2, 4, 10, 15, 19, 31, 32, 35, 36, 48, 54, 59, 71, 73, 75, 78, 81, 84, 89, 90, 91, 92, 96, 99]]target = [[78, 75, 60, 26, 20, 87, 76, 85, 16, 58, 46, 42, 44, 6, 50, 15, 43, 62, 5, 96, 89, 66, 83, 92]]
+Epoch 0: |                                                                             | 989/? [57:07<00:00,  0.29it/s, v_num=xp6d]train step 990; scene = [['d78079ff1a0f045d'], ['3b42fa1245f6b00b'], ['b66059464cc436c0']]; loss = 0.016426
+Epoch 0: |                                                                             | 990/? [57:11<00:00,  0.29it/s, v_num=xp6d]context = [[168, 175, 176, 185, 186, 188, 197, 208, 211, 223, 233, 245], [21, 34, 37, 41, 42, 45, 48, 51, 58, 62, 65, 70]]target = [[202, 190, 206, 174, 221, 205, 196, 211, 232, 219, 238, 178], [50, 45, 51, 29, 63, 54, 23, 52, 26, 41, 43, 61]]
+Epoch 0: |                                                                             | 999/? [57:43<00:00,  0.29it/s, v_num=xp6d]train step 1000; scene = [['c2a108b0ff836b87'], ['395b6ac27ddcc9d1'], ['e80ac219ae75ce5e'], ['2ac1cf1adda42447'], ['2de40cefd70bc708'], ['25d17315f4472e27'], ['658d3f56767f0396'], ['b953af75bafccce8']]; loss = 0.018113
+Epoch 0: |                                                                            | 1000/? [57:46<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 1000; scene = ['647f2049bf4cb3f3'];
+target intrinsic:  tensor(0.8998, device='cuda:0') tensor(0.9001, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8981, device='cuda:0') tensor(0.9012, device='cuda:0')
+[2026-02-25 05:34:35,366][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                            | 1000/? [57:47<00:00,  0.29it/s, v_num=xp6d]context = [[6, 9, 21, 32, 39, 40, 49, 54, 61, 62, 65, 68, 71, 75, 76, 78, 85, 87, 88, 91, 96, 99, 102, 103]]target = [[56, 19, 49, 28, 80, 26, 35, 23, 82, 39, 13, 15, 17, 44, 7, 58, 21, 36, 45, 40, 91, 53, 74, 70]]
+[2026-02-25 05:34:39,776][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                            | 1009/? [58:20<00:00,  0.29it/s, v_num=xp6d]train step 1010; scene = [['2beac37f687333f5'], ['8030e59b90ab02d6'], ['6e58885ab481a69e'], ['2f3f9d1ca227e778'], ['8ebbf8173feadc66'], ['6755d1366bce2ef0']]; loss = 0.028158
+Epoch 0: |                                                                            | 1010/? [58:24<00:00,  0.29it/s, v_num=xp6d]context = [[88, 90, 94, 95, 98, 99, 103, 122, 125, 128, 143, 144, 147, 151, 152, 155, 161, 164, 165, 166, 173, 181, 182, 185]]target = [[163, 135, 183, 143, 105, 90, 169, 97, 151, 127, 140, 138, 91, 104, 157, 102, 146, 152, 145, 129, 100, 162, 99, 130]]
+Epoch 0: |                                                                            | 1019/? [58:55<00:00,  0.29it/s, v_num=xp6d]train step 1020; scene = [['dd657030a3c637d2']]; loss = 0.028413
+Epoch 0: |                                                                            | 1020/? [58:59<00:00,  0.29it/s, v_num=xp6d]context = [[77, 80, 91, 96, 98, 105, 108, 112, 130, 131, 134, 136], [10, 18, 20, 26, 34, 36, 65, 75, 78, 80, 81, 82]]target = [[102, 105, 104, 88, 119, 127, 110, 121, 101, 80, 84, 122], [65, 17, 15, 23, 21, 52, 48, 50, 38, 68, 42, 47]]
+Epoch 0: |                                                                            | 1029/? [59:30<00:00,  0.29it/s, v_num=xp6d]train step 1030; scene = [['b6a629f2e3cbf2c3'], ['04d96fd21593f793'], ['3dc32c403d742301']]; loss = 0.019320
+Epoch 0: |                                                                            | 1030/? [59:33<00:00,  0.29it/s, v_num=xp6d]context = [[58, 61, 64, 72, 73, 77, 79, 85, 95, 101, 103, 112], [13, 14, 21, 27, 36, 38, 49, 53, 54, 69, 71, 79]]target = [[78, 77, 97, 71, 100, 101, 99, 68, 73, 105, 87, 88], [53, 36, 27, 57, 47, 64, 28, 39, 49, 58, 33, 35]]
+Epoch 0: |                                                                          | 1039/? [1:00:05<00:00,  0.29it/s, v_num=xp6d]train step 1040; scene = [['94aa5fd022f4400b'], ['21be2154f3c073c1'], ['103ff72c7bde3074'], ['db811a2460c4f9b5'], ['27159dbf7b134cd9'], ['db61533e680f5d74']]; loss = 0.022184
+Epoch 0: |                                                                          | 1040/? [1:00:09<00:00,  0.29it/s, v_num=xp6d]context = [[1, 8, 13, 22, 24, 25, 31, 39, 43, 44, 47, 58, 60, 64, 65, 70, 76, 77, 85, 87, 90, 91, 94, 98]]target = [[30, 38, 34, 60, 10, 39, 94, 80, 35, 59, 20, 52, 6, 97, 61, 72, 40, 63, 93, 46, 41, 7, 55, 23]]
+Epoch 0: |                                                                          | 1049/? [1:00:40<00:00,  0.29it/s, v_num=xp6d]train step 1050; scene = [['94ddacdba86c4757'], ['2650dfe1a0b7976a'], ['7273382ad589d7ac'], ['dcfb0ddcac6008e6'], ['b382af5f342061fa'], ['d325bd35d81548c3']]; loss = 0.015390
+Epoch 0: |                                                                          | 1050/? [1:00:44<00:00,  0.29it/s, v_num=xp6d]context = [[68, 69, 80, 82, 84, 86, 90, 98, 104, 108, 117, 120, 122, 126, 136, 138, 144, 149, 150, 154, 157, 159, 160, 165]]target = [[138, 111, 147, 162, 96, 129, 95, 97, 103, 117, 90, 164, 78, 130, 155, 119, 123, 75, 107, 80, 82, 143, 72, 150]]
+Epoch 0: |                                                                          | 1059/? [1:01:15<00:00,  0.29it/s, v_num=xp6d]train step 1060; scene = [['1fdf19d7c4402d40'], ['57520a34789bd5d8'], ['5431b4af9712cad6']]; loss = 0.011825
+Epoch 0: |                                                                          | 1060/? [1:01:18<00:00,  0.29it/s, v_num=xp6d]context = [[23, 28, 29, 30, 35, 38, 41, 47, 56, 60, 61, 66, 83, 84, 86, 91, 95, 96, 102, 107, 109, 110, 111, 120]]target = [[49, 44, 112, 36, 32, 68, 78, 38, 75, 113, 34, 50, 91, 54, 116, 82, 33, 87, 24, 117, 48, 98, 43, 96]]
+Epoch 0: |                                                                          | 1069/? [1:01:49<00:00,  0.29it/s, v_num=xp6d]train step 1070; scene = [['16fd029febc69afa'], ['901e47c506f0d2c1']]; loss = 0.014449
+Epoch 0: |                                                                          | 1070/? [1:01:53<00:00,  0.29it/s, v_num=xp6d]context = [[1, 10, 55, 65], [52, 104, 105, 106], [124, 138, 171, 197], [1, 25, 73, 76], [9, 48, 62, 99], [129, 166, 172, 195]]target = [[32, 14, 59, 35], [87, 80, 69, 98], [186, 172, 133, 129], [53, 21, 42, 7], [57, 81, 40, 91], [161, 159, 155, 175]]
+Epoch 0: |                                                                          | 1079/? [1:02:24<00:00,  0.29it/s, v_num=xp6d]train step 1080; scene = [['99e7a4ff1897a94d'], ['d0e7e477ff1174d4'], ['b21723140f5d4a70'], ['5102042496230ec1'], ['3c5825296fdc2a5f'], ['f8532bb5eca0e29d'], ['c58abf74527c5ce1'], ['d7decc9863ca94ef']]; loss = 0.029849
+Epoch 0: |                                                                          | 1080/? [1:02:27<00:00,  0.29it/s, v_num=xp6d]context = [[6, 7, 10, 11, 13, 17, 28, 31, 32, 41, 46, 50, 51, 52, 57, 64, 65, 74, 80, 83, 86, 93, 101, 103]]target = [[44, 96, 81, 43, 67, 92, 30, 36, 83, 74, 79, 78, 59, 41, 85, 10, 16, 64, 52, 12, 24, 58, 35, 54]]
+Epoch 0: |                                                                          | 1089/? [1:02:59<00:00,  0.29it/s, v_num=xp6d]train step 1090; scene = [['9724d02dd7954ece'], ['0cf733dbbb0e017f'], ['436ccd6836a9c74e'], ['d4cc5c4887f72d3e'], ['d3711afbbda1b025'], ['18bc67e83f83fd74'], ['2517e14bc498b4f8'], ['8edf724cd98c75c0']]; loss = 0.025792
+Epoch 0: |                                                                          | 1090/? [1:03:02<00:00,  0.29it/s, v_num=xp6d]context = [[5, 9, 16, 17, 18, 19, 24, 27, 32, 42, 64, 71], [19, 22, 24, 36, 40, 51, 52, 57, 60, 62, 65, 70]]target = [[9, 13, 11, 40, 43, 34, 6, 52, 18, 61, 56, 49], [51, 48, 39, 43, 58, 68, 25, 44, 21, 22, 29, 62]]
+Epoch 0: |                                                                          | 1099/? [1:03:33<00:00,  0.29it/s, v_num=xp6d]train step 1100; scene = [['6cee8362ee57a4c7']]; loss = 0.019661
+Epoch 0: |                                                                          | 1100/? [1:03:37<00:00,  0.29it/s, v_num=xp6d]context = [[140, 161, 164, 170, 177, 205, 207, 217], [11, 12, 17, 21, 27, 44, 63, 72], [0, 7, 10, 29, 45, 54, 55, 62]]target = [[167, 185, 195, 197, 177, 196, 168, 192], [17, 42, 22, 47, 65, 21, 50, 56], [3, 58, 22, 9, 25, 24, 37, 52]]
+Epoch 0: |                                                                          | 1109/? [1:04:08<00:00,  0.29it/s, v_num=xp6d]train step 1110; scene = [['7b2c118f021e6902'], ['05ed7f84c7d34ed9'], ['ced7fe64b7867119'], ['e1d9628095f76a9f']]; loss = 0.019010
+Epoch 0: |                                                                          | 1110/? [1:04:11<00:00,  0.29it/s, v_num=xp6d]context = [[4, 11, 12, 16, 20, 24, 25, 29, 33, 35, 44, 46, 47, 53, 57, 58, 62, 63, 83, 84, 90, 94, 98, 101]]target = [[98, 46, 60, 18, 67, 28, 23, 16, 37, 34, 53, 71, 80, 57, 88, 64, 54, 36, 42, 78, 17, 89, 27, 45]]
+Epoch 0: |                                                                          | 1119/? [1:04:42<00:00,  0.29it/s, v_num=xp6d]train step 1120; scene = [['1b449c2b2ba10206']]; loss = 0.045272
+Epoch 0: |                                                                          | 1120/? [1:04:46<00:00,  0.29it/s, v_num=xp6d]context = [[161, 199, 216, 230], [200, 229, 244, 248], [86, 110, 132, 147], [21, 31, 50, 70], [7, 56, 60, 71], [0, 56, 73, 84]]target = [[218, 207, 225, 224], [236, 220, 241, 212], [100, 131, 108, 89], [31, 35, 58, 32], [13, 59, 40, 22], [37, 35, 29, 70]]
+Epoch 0: |                                                                          | 1129/? [1:05:16<00:00,  0.29it/s, v_num=xp6d]train step 1130; scene = [['373f042a22f7d6ff'], ['4fd852226711ffa9'], ['013e509c64d311ba'], ['2dc9c96992a50777']]; loss = 0.014604
+Epoch 0: |                                                                          | 1130/? [1:05:20<00:00,  0.29it/s, v_num=xp6d]context = [[105, 136, 160], [2, 89, 90], [42, 113, 117], [16, 53, 66], [82, 110, 133], [4, 17, 62], [111, 147, 181], [25, 27, 79]]target = [[121, 151, 140], [28, 11, 50], [97, 75, 99], [64, 24, 29], [116, 105, 120], [37, 38, 59], [147, 136, 154], [43, 68, 33]]
+Epoch 0: |                                                                          | 1139/? [1:05:51<00:00,  0.29it/s, v_num=xp6d]train step 1140; scene = [['04719d552f4d9d25'], ['066f14037c220103'], ['2d6dc5466ac4ed93'], ['4f9716bb3dc7feec']]; loss = 0.014498
+Epoch 0: |                                                                          | 1140/? [1:05:55<00:00,  0.29it/s, v_num=xp6d]context = [[23, 33, 34, 35, 51, 58, 70, 71, 89, 91, 95, 106], [10, 16, 24, 26, 30, 40, 45, 46, 48, 53, 55, 64]]target = [[53, 100, 54, 79, 60, 98, 36, 78, 105, 28, 52, 42], [59, 31, 46, 41, 63, 58, 22, 43, 17, 24, 13, 15]]
+Epoch 0: |                                                                          | 1149/? [1:06:26<00:00,  0.29it/s, v_num=xp6d]train step 1150; scene = [['308ec88c1ad9b768'], ['b3c43704ee2efd4f'], ['7239b2a7554c75fe'], ['5645a008715acf0a']]; loss = 0.014149
+Epoch 0: |                                                                          | 1150/? [1:06:30<00:00,  0.29it/s, v_num=xp6d]context = [[10, 20, 28, 29, 34, 35, 40, 42, 50, 51, 54, 59], [18, 19, 26, 31, 34, 36, 37, 39, 46, 52, 69, 76]]target = [[29, 38, 24, 15, 11, 54, 21, 26, 28, 36, 19, 37], [71, 38, 24, 72, 27, 29, 30, 19, 60, 58, 53, 40]]
+Epoch 0: |                                                                          | 1159/? [1:07:01<00:00,  0.29it/s, v_num=xp6d]train step 1160; scene = [['016349b94babf6d4'], ['d72595b1e4d300b8'], ['b072d5e75e95e9cd'], ['3dc75de99ef09a0c'], ['8cfd52fbbac85c10'], ['742eaf7317415c35']]; loss = 0.031924
+Epoch 0: |                                                                          | 1160/? [1:07:04<00:00,  0.29it/s, v_num=xp6d]context = [[87, 88, 95, 99, 102, 112, 125, 156, 159, 160, 162, 167], [3, 4, 6, 8, 22, 25, 27, 30, 33, 37, 48, 58]]target = [[104, 129, 155, 116, 114, 147, 98, 143, 108, 162, 166, 131], [20, 40, 11, 42, 7, 16, 21, 56, 27, 50, 28, 6]]
+Epoch 0: |                                                                          | 1169/? [1:07:36<00:00,  0.29it/s, v_num=xp6d]train step 1170; scene = [['f75b1a55b42e01fc'], ['fc33d4814d21b0ff']]; loss = 0.007344
+Epoch 0: |                                                                          | 1170/? [1:07:39<00:00,  0.29it/s, v_num=xp6d]context = [[10, 31, 58], [30, 58, 84], [178, 228, 258], [17, 48, 85], [56, 96, 104], [99, 112, 153], [28, 44, 91], [4, 16, 60]]target = [[37, 30, 54], [35, 45, 78], [184, 229, 232], [38, 66, 72], [69, 81, 101], [143, 127, 121], [33, 89, 50], [9, 18, 29]]
+Epoch 0: |                                                                          | 1179/? [1:08:10<00:00,  0.29it/s, v_num=xp6d]train step 1180; scene = [['93b36a54151e085e']]; loss = 0.012459
+Epoch 0: |                                                                          | 1180/? [1:08:14<00:00,  0.29it/s, v_num=xp6d]context = [[35, 69, 77, 82], [22, 66, 77, 92], [3, 23, 51, 59], [0, 11, 35, 67], [179, 193, 225, 245], [60, 64, 90, 110]]target = [[59, 78, 50, 41], [29, 66, 30, 42], [35, 30, 12, 5], [29, 46, 1, 16], [222, 211, 185, 205], [61, 78, 84, 71]]
+Epoch 0: |                                                                          | 1189/? [1:08:46<00:00,  0.29it/s, v_num=xp6d]train step 1190; scene = [['cc20d092a599c493']]; loss = 0.008231
+Epoch 0: |                                                                          | 1190/? [1:08:49<00:00,  0.29it/s, v_num=xp6d]context = [[151, 152, 163, 164, 166, 176, 186, 189, 190, 191, 196, 197, 201, 203, 206, 211, 218, 220, 226, 233, 235, 242, 244, 248]]target = [[183, 188, 157, 243, 213, 164, 235, 182, 234, 225, 247, 221, 194, 212, 218, 224, 153, 198, 214, 197, 184, 166, 210, 191]]
+Epoch 0: |                                                                          | 1199/? [1:09:20<00:00,  0.29it/s, v_num=xp6d]train step 1200; scene = [['ad3c210c5cdc595f'], ['fca33c5a37b6c7bf']]; loss = 0.017915
+Epoch 0: |                                                                          | 1200/? [1:09:24<00:00,  0.29it/s, v_num=xp6d]context = [[0, 8, 13, 14, 21, 61, 83, 88], [4, 25, 31, 32, 42, 47, 55, 56], [60, 65, 69, 74, 92, 105, 114, 133]]target = [[39, 31, 85, 21, 17, 38, 22, 29], [18, 41, 37, 6, 22, 52, 39, 31], [77, 123, 120, 62, 65, 68, 78, 67]]
+[2026-02-25 05:46:15,972][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1209/? [1:09:55<00:00,  0.29it/s, v_num=xp6d]train step 1210; scene = [['41986a305a19ca1d'], ['0b707735ebc3eeed']]; loss = 0.011759
+Epoch 0: |                                                                          | 1210/? [1:09:58<00:00,  0.29it/s, v_num=xp6d]context = [[135, 145, 149, 150, 152, 153, 157, 158, 160, 163, 167, 174, 178, 183, 185, 187, 192, 193, 195, 198, 204, 212, 230, 232]]target = [[157, 224, 203, 149, 225, 192, 197, 217, 139, 142, 211, 218, 215, 156, 173, 155, 188, 136, 160, 180, 222, 178, 137, 204]]
+Epoch 0: |                                                                          | 1219/? [1:10:30<00:00,  0.29it/s, v_num=xp6d]train step 1220; scene = [['5795c143b1adc665'], ['d51d569c27d0b2b5'], ['f51a6709d0525ad3']]; loss = 0.013269
+Epoch 0: |                                                                          | 1220/? [1:10:33<00:00,  0.29it/s, v_num=xp6d]context = [[0, 11, 45, 56], [3, 17, 48, 54], [59, 78, 108, 111], [15, 30, 46, 71], [3, 39, 51, 53], [21, 46, 71, 99]]target = [[9, 30, 19, 54], [18, 27, 48, 49], [110, 101, 89, 90], [33, 58, 57, 70], [27, 41, 36, 43], [75, 35, 48, 49]]
+Epoch 0: |                                                                          | 1229/? [1:11:05<00:00,  0.29it/s, v_num=xp6d]train step 1230; scene = [['3d2496a06b64ab52'], ['95a572d344c5c54f']]; loss = 0.010825
+Epoch 0: |                                                                          | 1230/? [1:11:08<00:00,  0.29it/s, v_num=xp6d]context = [[12, 21, 27, 73, 83, 95, 98, 99], [15, 27, 37, 59, 68, 78, 80, 98], [181, 182, 200, 210, 221, 236, 250, 257]]target = [[62, 45, 80, 83, 92, 37, 72, 56], [74, 95, 23, 71, 43, 89, 19, 56], [193, 224, 203, 192, 249, 255, 202, 207]]
+Epoch 0: |                                                                          | 1239/? [1:11:39<00:00,  0.29it/s, v_num=xp6d]train step 1240; scene = [['b07961433f13f70e'], ['89a3b0eed03cf522']]; loss = 0.011066
+Epoch 0: |                                                                          | 1240/? [1:11:42<00:00,  0.29it/s, v_num=xp6d]context = [[200, 206, 207, 208, 209, 216, 218, 219, 226, 233, 237, 261], [16, 21, 25, 30, 32, 61, 67, 73, 76, 94, 97, 99]]target = [[252, 211, 215, 220, 225, 235, 239, 203, 229, 217, 231, 230], [64, 42, 62, 98, 88, 28, 63, 66, 45, 90, 36, 89]]
+Epoch 0: |                                                                          | 1249/? [1:12:14<00:00,  0.29it/s, v_num=xp6d]train step 1250; scene = [['8bd8e3f6166b3bf2']]; loss = 0.035816
+Epoch 0: |                                                                          | 1250/? [1:12:18<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 1250; scene = ['70b0a33083333dc9'];
+target intrinsic:  tensor(0.8872, device='cuda:0') tensor(0.8874, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8750, device='cuda:0') tensor(0.8746, device='cuda:0')
+[2026-02-25 05:49:06,959][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1250/? [1:12:19<00:00,  0.29it/s, v_num=xp6d]context = [[49, 53, 64, 77, 88, 90, 103, 106], [33, 57, 60, 68, 70, 77, 78, 79], [67, 85, 87, 108, 109, 112, 131, 136]]target = [[98, 95, 87, 69, 100, 50, 82, 61], [36, 74, 51, 57, 62, 35, 77, 39], [119, 68, 85, 82, 77, 113, 96, 73]]
+Epoch 0: |                                                                          | 1259/? [1:12:49<00:00,  0.29it/s, v_num=xp6d]train step 1260; scene = [['8875b49387b58221']]; loss = 0.010710
+Epoch 0: |                                                                          | 1260/? [1:12:53<00:00,  0.29it/s, v_num=xp6d]context = [[62, 63, 66, 67, 71, 78, 82, 83, 86, 92, 95, 99, 102, 106, 108, 110, 119, 125, 126, 140, 146, 147, 152, 159]]target = [[88, 73, 95, 75, 125, 147, 90, 157, 91, 79, 104, 120, 132, 98, 107, 111, 118, 129, 154, 112, 114, 156, 134, 78]]
+Epoch 0: |                                                                          | 1269/? [1:13:24<00:00,  0.29it/s, v_num=xp6d]train step 1270; scene = [['336b3140d9d8bebd']]; loss = 0.012364
+Epoch 0: |                                                                          | 1270/? [1:13:27<00:00,  0.29it/s, v_num=xp6d]context = [[7, 10, 11, 24, 28, 33, 35, 47, 51, 57, 63, 64], [151, 162, 169, 174, 176, 179, 180, 181, 187, 197, 213, 216]]target = [[25, 26, 24, 54, 8, 18, 55, 38, 15, 58, 41, 61], [179, 193, 206, 204, 202, 191, 187, 166, 189, 182, 154, 196]]
+Epoch 0: |                                                                          | 1279/? [1:13:58<00:00,  0.29it/s, v_num=xp6d]train step 1280; scene = [['06ed4433d2640109'], ['6ca39802dcef328e']]; loss = 0.015476
+Epoch 0: |                                                                          | 1280/? [1:14:01<00:00,  0.29it/s, v_num=xp6d]context = [[215, 218, 240, 241, 254, 272], [47, 63, 108, 113, 114, 116], [77, 91, 108, 114, 118, 151], [19, 22, 24, 36, 72, 74]]target = [[242, 267, 217, 259, 226, 220], [48, 63, 79, 78, 100, 108], [81, 147, 143, 129, 82, 112], [53, 71, 20, 36, 21, 41]]
+Epoch 0: |                                                                          | 1289/? [1:14:33<00:00,  0.29it/s, v_num=xp6d]train step 1290; scene = [['9138f683b523d708'], ['c1289a667237cd36'], ['9c1dfe99d620e067'], ['77c03290d901ec15'], ['37f2b83c868084df'], ['06ea674963ad018e'], ['4797e813ea4a5ed2'], ['aa5175d18a929cb8']]; loss = 0.018690
+Epoch 0: |                                                                          | 1290/? [1:14:36<00:00,  0.29it/s, v_num=xp6d]context = [[198, 205, 214, 215, 235, 236, 245, 246], [21, 24, 28, 52, 75, 78, 83, 108], [8, 18, 20, 49, 60, 80, 83, 86]]target = [[222, 235, 234, 215, 219, 240, 218, 200], [24, 35, 28, 23, 46, 32, 90, 84], [33, 38, 57, 55, 20, 52, 58, 26]]
+Epoch 0: |                                                                          | 1299/? [1:15:06<00:00,  0.29it/s, v_num=xp6d]train step 1300; scene = [['6b2e49b1e748eb08'], ['20a321699aab25aa'], ['d15f97346b6c753c']]; loss = 0.027593
+Epoch 0: |                                                                          | 1300/? [1:15:10<00:00,  0.29it/s, v_num=xp6d]context = [[104, 106, 107, 112, 113, 116, 119, 120, 128, 137, 138, 146, 148, 150, 152, 153, 158, 164, 168, 173, 177, 193, 198, 201]]target = [[117, 169, 142, 191, 108, 125, 146, 177, 126, 116, 155, 161, 135, 132, 110, 122, 160, 199, 162, 190, 150, 168, 127, 140]]
+Epoch 0: |                                                                          | 1309/? [1:15:41<00:00,  0.29it/s, v_num=xp6d]train step 1310; scene = [['a4db4f86a3f6a3c9']]; loss = 0.020897
+Epoch 0: |                                                                          | 1310/? [1:15:45<00:00,  0.29it/s, v_num=xp6d]context = [[95, 103, 109, 123, 131, 140, 149, 153], [2, 3, 10, 15, 31, 53, 71, 75], [97, 98, 114, 118, 121, 126, 137, 143]]target = [[142, 152, 122, 137, 115, 143, 109, 107], [64, 11, 71, 42, 74, 41, 39, 24], [117, 99, 135, 133, 98, 116, 127, 129]]
+Epoch 0: |                                                                          | 1319/? [1:16:17<00:00,  0.29it/s, v_num=xp6d]train step 1320; scene = [['651489c4eabb5ef9']]; loss = 0.016437
+Epoch 0: |                                                                          | 1320/? [1:16:20<00:00,  0.29it/s, v_num=xp6d]context = [[5, 7, 14, 25, 27, 30, 34, 37, 39, 43, 47, 51, 55, 59, 63, 69, 78, 81, 87, 90, 93, 96, 99, 102]]target = [[40, 74, 11, 26, 92, 6, 61, 96, 12, 78, 19, 93, 54, 9, 33, 21, 66, 62, 32, 52, 60, 99, 41, 83]]
+Epoch 0: |                                                                          | 1329/? [1:16:52<00:00,  0.29it/s, v_num=xp6d]train step 1330; scene = [['36047ec1694f9d49'], ['1781ab25cd748560'], ['23ac87fd90e9a51a']]; loss = 0.032128
+Epoch 0: |                                                                          | 1330/? [1:16:56<00:00,  0.29it/s, v_num=xp6d]context = [[105, 138, 140, 169, 183, 184], [36, 62, 69, 77, 87, 91], [24, 43, 49, 54, 95, 97], [69, 84, 88, 104, 106, 123]]target = [[167, 150, 143, 112, 123, 113], [81, 61, 85, 38, 57, 84], [52, 41, 36, 89, 91, 63], [76, 78, 83, 74, 96, 86]]
+Epoch 0: |                                                                          | 1339/? [1:17:26<00:00,  0.29it/s, v_num=xp6d]train step 1340; scene = [['1eb41a7aa81df97b']]; loss = 0.010616
+Epoch 0: |                                                                          | 1340/? [1:17:30<00:00,  0.29it/s, v_num=xp6d]context = [[32, 46, 61, 65, 77, 96, 101, 102], [18, 23, 25, 40, 57, 58, 71, 80], [1, 6, 15, 35, 42, 52, 67, 75]]target = [[50, 85, 100, 73, 48, 89, 44, 42], [19, 36, 69, 52, 72, 70, 22, 67], [14, 13, 26, 23, 47, 53, 17, 34]]
+Epoch 0: |                                                                          | 1349/? [1:18:02<00:00,  0.29it/s, v_num=xp6d]train step 1350; scene = [['aa293d98738dc39f'], ['40dc627bd4ddc389'], ['fb45a86f6154e126'], ['baa5664b5ea5348e'], ['bd814603b3e38fb1'], ['252a4d522d463417'], ['68df2105110fbafe'], ['19dbf45ccabba3ff']]; loss = 0.016348
+Epoch 0: |                                                                          | 1350/? [1:18:06<00:00,  0.29it/s, v_num=xp6d]context = [[10, 13, 14, 22, 24, 27, 29, 31, 34, 38, 50, 59, 60, 61, 62, 65, 68, 71, 81, 87, 93, 103, 104, 107]]target = [[59, 64, 73, 49, 15, 84, 12, 70, 47, 33, 65, 97, 101, 75, 25, 27, 28, 42, 76, 17, 98, 86, 88, 68]]
+Epoch 0: |                                                                          | 1359/? [1:18:37<00:00,  0.29it/s, v_num=xp6d]train step 1360; scene = [['2d78bdcafda0e615'], ['fb526fa7f45a3ae4']]; loss = 0.013357
+Epoch 0: |                                                                          | 1360/? [1:18:40<00:00,  0.29it/s, v_num=xp6d]context = [[32, 53, 57, 60, 98, 103, 104, 119], [10, 16, 17, 25, 30, 34, 58, 65], [6, 40, 41, 43, 46, 52, 58, 63]]target = [[112, 55, 85, 93, 52, 106, 102, 92], [57, 58, 33, 47, 63, 64, 50, 59], [56, 30, 40, 10, 29, 50, 37, 59]]
+Epoch 0: |                                                                          | 1369/? [1:19:12<00:00,  0.29it/s, v_num=xp6d]train step 1370; scene = [['b7aeefdcf8c006ec']]; loss = 0.009820
+Epoch 0: |                                                                          | 1370/? [1:19:15<00:00,  0.29it/s, v_num=xp6d]context = [[42, 53, 56, 57, 63, 70, 77, 87], [21, 30, 36, 54, 60, 73, 84, 91], [121, 130, 134, 141, 151, 160, 161, 166]]target = [[46, 85, 73, 76, 49, 52, 75, 60], [59, 83, 46, 71, 74, 23, 82, 49], [156, 130, 139, 123, 161, 158, 145, 135]]
+Epoch 0: |                                                                          | 1379/? [1:19:47<00:00,  0.29it/s, v_num=xp6d]train step 1380; scene = [['db25ce184054217a']]; loss = 0.010593
+Epoch 0: |                                                                          | 1380/? [1:19:51<00:00,  0.29it/s, v_num=xp6d]context = [[20, 21, 23, 34, 36, 43, 46, 47, 49, 60, 61, 65, 68, 83, 84, 86, 87, 91, 93, 94, 106, 112, 113, 117]]target = [[22, 102, 111, 79, 52, 59, 108, 32, 88, 90, 68, 39, 113, 24, 71, 85, 89, 67, 62, 30, 95, 92, 44, 86]]
+Epoch 0: |                                                                          | 1389/? [1:20:22<00:00,  0.29it/s, v_num=xp6d]train step 1390; scene = [['73341b79eaf41f6a'], ['7f4353922e24e719'], ['572f15c53545baaf']]; loss = 0.009988
+Epoch 0: |                                                                          | 1390/? [1:20:26<00:00,  0.29it/s, v_num=xp6d]context = [[64, 65, 69, 73, 74, 77, 81, 89, 96, 99, 101, 103, 104, 108, 114, 116, 118, 129, 132, 137, 138, 146, 150, 161]]target = [[68, 159, 160, 90, 82, 79, 78, 110, 155, 108, 115, 65, 121, 153, 67, 156, 92, 139, 114, 73, 80, 107, 109, 135]]
+Epoch 0: |                                                                          | 1399/? [1:20:57<00:00,  0.29it/s, v_num=xp6d]train step 1400; scene = [['7d634ffc2785a9f4'], ['acd293363db06c1c']]; loss = 0.011710
+Epoch 0: |                                                                          | 1400/? [1:21:01<00:00,  0.29it/s, v_num=xp6d]context = [[48, 50, 53, 54, 75, 87, 89, 115], [5, 20, 24, 31, 38, 72, 78, 85], [11, 18, 24, 53, 66, 76, 88, 91]]target = [[77, 79, 96, 75, 102, 87, 62, 54], [14, 61, 24, 79, 43, 11, 13, 77], [38, 25, 29, 20, 69, 43, 60, 28]]
+[2026-02-25 05:57:53,153][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1409/? [1:21:33<00:00,  0.29it/s, v_num=xp6d]train step 1410; scene = [['3ec980a78e9467e0'], ['6744c1cf4e4dee91']]; loss = 0.011946
+Epoch 0: |                                                                          | 1410/? [1:21:36<00:00,  0.29it/s, v_num=xp6d]context = [[1, 2, 6, 9, 17, 27, 43, 80], [40, 43, 54, 57, 67, 73, 97, 103], [89, 91, 110, 112, 125, 131, 140, 147]]target = [[39, 53, 4, 2, 75, 68, 31, 51], [102, 56, 62, 45, 92, 90, 60, 75], [127, 128, 144, 94, 130, 124, 113, 112]]
+Epoch 0: |                                                                          | 1419/? [1:22:08<00:00,  0.29it/s, v_num=xp6d]train step 1420; scene = [['9ac8241a27b0ec6d']]; loss = 0.014101
+Epoch 0: |                                                                          | 1420/? [1:22:11<00:00,  0.29it/s, v_num=xp6d]context = [[29, 36, 38, 41, 42, 43, 48, 51, 56, 59, 62, 69, 78, 87, 90, 92, 94, 97, 99, 101, 105, 106, 124, 126]]target = [[41, 97, 113, 39, 116, 48, 107, 102, 36, 121, 54, 50, 115, 61, 94, 123, 57, 82, 86, 30, 101, 93, 117, 32]]
+Epoch 0: |                                                                          | 1429/? [1:22:43<00:00,  0.29it/s, v_num=xp6d]train step 1430; scene = [['54d3668b6c7ed4b8']]; loss = 0.008274
+Epoch 0: |                                                                          | 1430/? [1:22:46<00:00,  0.29it/s, v_num=xp6d]context = [[12, 18, 32, 39, 56, 72, 76, 92], [70, 74, 86, 100, 115, 118, 125, 128], [22, 54, 56, 65, 68, 70, 72, 77]]target = [[29, 33, 27, 22, 90, 81, 54, 63], [123, 120, 72, 92, 74, 101, 104, 125], [49, 56, 58, 69, 31, 37, 32, 73]]
+Epoch 0: |                                                                          | 1439/? [1:23:18<00:00,  0.29it/s, v_num=xp6d]train step 1440; scene = [['070e6ca07b5f898c'], ['6c383c3e7ece2df7'], ['fbc5d8715d6debec']]; loss = 0.012010
+Epoch 0: |                                                                          | 1440/? [1:23:22<00:00,  0.29it/s, v_num=xp6d]context = [[1, 7, 12, 17, 24, 26, 27, 29, 32, 34, 38, 42, 45, 55, 63, 67, 69, 71, 72, 75, 88, 91, 95, 98]]target = [[62, 12, 35, 63, 55, 46, 47, 38, 82, 22, 4, 89, 75, 25, 9, 54, 70, 60, 76, 90, 28, 2, 67, 18]]
+Epoch 0: |                                                                          | 1449/? [1:23:51<00:00,  0.29it/s, v_num=xp6d]train step 1450; scene = [['61aef773e015ee1b'], ['895fe07be01436f1']]; loss = 0.013833
+Epoch 0: |                                                                          | 1450/? [1:23:55<00:00,  0.29it/s, v_num=xp6d]context = [[52, 60, 116], [92, 107, 146], [81, 134, 149], [11, 15, 85], [60, 67, 136], [32, 78, 121], [9, 24, 78], [25, 52, 77]]target = [[63, 103, 95], [107, 94, 138], [131, 129, 114], [73, 36, 42], [73, 65, 123], [61, 83, 107], [50, 38, 34], [42, 62, 74]]
+Epoch 0: |                                                                          | 1459/? [1:24:26<00:00,  0.29it/s, v_num=xp6d]train step 1460; scene = [['913bc26ba7d3afc5']]; loss = 0.022054
+Epoch 0: |                                                                          | 1460/? [1:24:29<00:00,  0.29it/s, v_num=xp6d]context = [[53, 54, 55, 58, 67, 68, 74, 76, 81, 82, 85, 86, 97, 104, 109, 112, 114, 115, 119, 127, 141, 142, 144, 150]]target = [[71, 143, 95, 81, 66, 54, 140, 125, 141, 94, 131, 146, 86, 124, 129, 144, 92, 111, 91, 110, 138, 93, 109, 128]]
+Epoch 0: |                                                                          | 1469/? [1:25:00<00:00,  0.29it/s, v_num=xp6d]train step 1470; scene = [['f73db02fdfe72073'], ['1bb8fe74783c5161'], ['7fa967710229854e'], ['4ec72717091e92c7']]; loss = 0.009566
+Epoch 0: |                                                                          | 1470/? [1:25:04<00:00,  0.29it/s, v_num=xp6d]context = [[115, 121, 127, 134, 143, 144, 149, 150, 151, 161, 165, 197], [1, 3, 4, 11, 29, 31, 37, 49, 58, 72, 80, 87]]target = [[143, 184, 154, 183, 132, 151, 157, 144, 168, 130, 169, 153], [77, 4, 82, 21, 18, 76, 29, 10, 71, 15, 56, 50]]
+Epoch 0: |                                                                          | 1479/? [1:25:35<00:00,  0.29it/s, v_num=xp6d]train step 1480; scene = [['8b91f281b1c3feec'], ['801309986851fc16']]; loss = 0.012915
+Epoch 0: |                                                                          | 1480/? [1:25:38<00:00,  0.29it/s, v_num=xp6d]context = [[27, 30, 33, 36, 57, 58, 59, 61, 71, 75, 82, 93], [78, 79, 82, 87, 98, 99, 108, 110, 123, 125, 132, 136]]target = [[47, 64, 90, 74, 32, 44, 79, 86, 81, 88, 30, 82], [81, 104, 82, 119, 130, 113, 97, 87, 105, 125, 93, 102]]
+Epoch 0: |                                                                          | 1489/? [1:26:10<00:00,  0.29it/s, v_num=xp6d]train step 1490; scene = [['bbc6cab96f65f530'], ['62755beb1ebd7813']]; loss = 0.014407
+Epoch 0: |                                                                          | 1490/? [1:26:14<00:00,  0.29it/s, v_num=xp6d]context = [[17, 37, 94], [50, 65, 101], [89, 93, 143], [1, 27, 61], [17, 66, 81], [1, 50, 52], [23, 70, 78], [1, 17, 72]]target = [[35, 92, 33], [68, 63, 97], [111, 133, 120], [50, 42, 29], [55, 59, 60], [22, 10, 15], [48, 46, 64], [19, 14, 34]]
+Epoch 0: |                                                                          | 1499/? [1:26:45<00:00,  0.29it/s, v_num=xp6d]train step 1500; scene = [['35a4da1b96cb6e60'], ['a5bcdef3441b37cd']]; loss = 0.013761
+Epoch 0: |                                                                          | 1500/? [1:26:49<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 1500; scene = ['45592a7f307bccd0'];
+target intrinsic:  tensor(0.8508, device='cuda:0') tensor(0.8510, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8670, device='cuda:0') tensor(0.8656, device='cuda:0')
+[2026-02-25 06:03:52,411][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1500/? [1:27:04<00:00,  0.29it/s, v_num=xp6d]context = [[121, 160, 183], [50, 62, 95], [5, 48, 57], [27, 29, 90], [23, 107, 111], [10, 25, 55], [1, 50, 89], [17, 60, 67]]target = [[128, 139, 182], [60, 93, 89], [15, 25, 17], [79, 84, 54], [80, 51, 60], [12, 13, 28], [80, 51, 57], [47, 36, 52]]
+Epoch 0: |                                                                          | 1509/? [1:27:34<00:00,  0.29it/s, v_num=xp6d]train step 1510; scene = [['7415ffa5c457e330'], ['abdfc86976ee11de'], ['3e30d1a14cad58d5'], ['a7b928c84f551702'], ['9de970f9a14770a9'], ['4558408811e71246']]; loss = 0.022813
+Epoch 0: |                                                                          | 1510/? [1:27:38<00:00,  0.29it/s, v_num=xp6d]context = [[3, 6, 11, 16, 24, 30, 37, 40, 41, 51, 55, 58, 61, 66, 71, 73, 77, 79, 86, 87, 95, 97, 98, 100]]target = [[31, 62, 24, 19, 4, 88, 98, 60, 6, 55, 52, 82, 36, 74, 69, 46, 37, 22, 23, 58, 87, 8, 38, 89]]
+Epoch 0: |                                                                          | 1519/? [1:28:10<00:00,  0.29it/s, v_num=xp6d]train step 1520; scene = [['17bfe61015a33214'], ['aecaf8906a0a9bb1']]; loss = 0.016376
+Epoch 0: |                                                                          | 1520/? [1:28:13<00:00,  0.29it/s, v_num=xp6d]context = [[3, 8, 9, 10, 13, 21, 28, 30, 35, 38, 42, 44, 46, 52, 57, 62, 63, 75, 84, 85, 90, 93, 94, 100]]target = [[48, 12, 4, 84, 73, 17, 71, 22, 96, 43, 94, 40, 27, 39, 66, 63, 59, 50, 26, 56, 37, 76, 68, 64]]
+Epoch 0: |                                                                          | 1529/? [1:28:45<00:00,  0.29it/s, v_num=xp6d]train step 1530; scene = [['aa790649db11f096']]; loss = 0.020501
+Epoch 0: |                                                                          | 1530/? [1:28:48<00:00,  0.29it/s, v_num=xp6d]context = [[166, 172, 173, 176, 179, 184, 196, 211], [30, 31, 35, 40, 51, 67, 89, 117], [11, 18, 21, 35, 41, 44, 70, 74]]target = [[180, 174, 203, 206, 179, 177, 192, 197], [55, 85, 98, 79, 62, 53, 108, 70], [12, 22, 56, 32, 65, 17, 21, 49]]
+Epoch 0: |                                                                          | 1539/? [1:29:19<00:00,  0.29it/s, v_num=xp6d]train step 1540; scene = [['1661797137c6af53']]; loss = 0.020419
+Epoch 0: |                                                                          | 1540/? [1:29:22<00:00,  0.29it/s, v_num=xp6d]context = [[144, 209, 210], [50, 103, 119], [99, 111, 146], [163, 198, 217], [51, 65, 124], [8, 58, 95], [44, 97, 113], [2, 44, 88]]target = [[187, 173, 188], [83, 59, 103], [145, 117, 108], [201, 207, 194], [121, 53, 112], [24, 23, 53], [78, 69, 83], [38, 29, 82]]
+Epoch 0: |                                                                          | 1549/? [1:29:54<00:00,  0.29it/s, v_num=xp6d]train step 1550; scene = [['db24baad7f78f763']]; loss = 0.010073
+Epoch 0: |                                                                          | 1550/? [1:29:57<00:00,  0.29it/s, v_num=xp6d]context = [[109, 112, 117, 133, 138, 147, 154, 156], [9, 12, 23, 58, 60, 62, 88, 94], [0, 3, 5, 13, 16, 22, 48, 59]]target = [[145, 149, 124, 121, 152, 130, 122, 153], [74, 49, 44, 28, 36, 68, 19, 26], [26, 53, 14, 17, 13, 31, 55, 29]]
+Epoch 0: |                                                                          | 1559/? [1:30:28<00:00,  0.29it/s, v_num=xp6d]train step 1560; scene = [['8d89803ec0446d0e'], ['e86d5510c0fb2a93'], ['6549544819976e02'], ['a7bc2227bd3a3006'], ['a6e57377594a7a40'], ['4c76d28a44171812']]; loss = 0.031100
+Epoch 0: |                                                                          | 1560/? [1:30:31<00:00,  0.29it/s, v_num=xp6d]context = [[14, 24, 28, 30, 34, 38, 46, 48, 51, 52, 55, 57, 64, 71, 73, 75, 81, 82, 83, 89, 98, 107, 108, 111]]target = [[107, 43, 55, 51, 29, 22, 64, 58, 105, 85, 56, 28, 65, 32, 36, 53, 102, 21, 78, 80, 39, 76, 74, 87]]
+Epoch 0: |                                                                          | 1569/? [1:31:03<00:00,  0.29it/s, v_num=xp6d]train step 1570; scene = [['df29eae05ee5e69f'], ['af8f48c2701a86d7'], ['46f7affec5786351']]; loss = 0.013935
+Epoch 0: |                                                                          | 1570/? [1:31:06<00:00,  0.29it/s, v_num=xp6d]context = [[2, 8, 10, 19, 23, 28, 29, 32, 34, 42, 45, 46, 48, 53, 54, 56, 58, 64, 66, 69, 74, 80, 90, 99]]target = [[53, 5, 17, 87, 13, 86, 54, 47, 65, 90, 62, 55, 57, 60, 51, 35, 66, 81, 93, 4, 25, 9, 74, 31]]
+Epoch 0: |                                                                          | 1579/? [1:31:38<00:00,  0.29it/s, v_num=xp6d]train step 1580; scene = [['09fef1964d67a64e']]; loss = 0.014640
+Epoch 0: |                                                                          | 1580/? [1:31:41<00:00,  0.29it/s, v_num=xp6d]context = [[24, 32, 64, 72], [45, 88, 92, 104], [15, 29, 51, 89], [29, 63, 69, 82], [14, 59, 81, 103], [10, 28, 48, 64]]target = [[48, 33, 32, 64], [75, 85, 98, 51], [43, 72, 58, 69], [67, 32, 69, 63], [83, 95, 61, 39], [36, 54, 12, 58]]
+Epoch 0: |                                                                          | 1589/? [1:32:13<00:00,  0.29it/s, v_num=xp6d]train step 1590; scene = [['d3917d0a1eda2a1f'], ['8df7ef966383d5ba']]; loss = 0.016701
+Epoch 0: |                                                                          | 1590/? [1:32:16<00:00,  0.29it/s, v_num=xp6d]context = [[20, 23, 36, 43, 44, 51, 68, 70, 71, 74, 77, 84], [6, 8, 19, 38, 39, 48, 55, 56, 63, 64, 70, 71]]target = [[71, 26, 69, 38, 28, 44, 64, 56, 43, 50, 51, 61], [18, 15, 51, 36, 8, 25, 70, 23, 30, 63, 31, 55]]
+Epoch 0: |                                                                          | 1599/? [1:32:47<00:00,  0.29it/s, v_num=xp6d]train step 1600; scene = [['8bb4464a97a1190f'], ['72d6108a977fb29c'], ['f2cc31144c64dc6a']]; loss = 0.011845
+Epoch 0: |                                                                          | 1600/? [1:32:51<00:00,  0.29it/s, v_num=xp6d]context = [[29, 37, 44, 66, 75, 86], [101, 117, 120, 140, 161, 169], [38, 53, 65, 70, 73, 85], [55, 67, 71, 76, 95, 102]]target = [[39, 62, 68, 36, 69, 34], [118, 115, 120, 157, 107, 102], [76, 39, 45, 64, 50, 68], [56, 78, 60, 57, 59, 95]]
+[2026-02-25 06:09:43,025][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1609/? [1:33:22<00:00,  0.29it/s, v_num=xp6d]train step 1610; scene = [['93598ff73741eb47'], ['9c3f970e0901e956'], ['a005c2781ac3e795']]; loss = 0.010217
+Epoch 0: |                                                                          | 1610/? [1:33:26<00:00,  0.29it/s, v_num=xp6d]context = [[22, 23, 31, 33, 35, 40, 41, 43, 44, 53, 62, 63, 67, 72, 84, 94, 95, 97, 102, 106, 109, 113, 116, 119]]target = [[92, 105, 54, 104, 40, 89, 87, 102, 96, 57, 99, 72, 48, 109, 55, 47, 75, 33, 91, 39, 66, 29, 23, 56]]
+Epoch 0: |                                                                          | 1619/? [1:33:58<00:00,  0.29it/s, v_num=xp6d]train step 1620; scene = [['81426f2270c3b08e'], ['f7d839262e5cc647']]; loss = 0.073825
+Epoch 0: |                                                                          | 1620/? [1:34:00<00:00,  0.29it/s, v_num=xp6d]context = [[52, 54, 57, 63, 70, 72, 81, 83, 86, 99, 103, 106], [19, 33, 38, 47, 48, 51, 53, 57, 58, 67, 68, 73]]target = [[64, 79, 60, 103, 102, 88, 70, 97, 56, 95, 82, 86], [72, 61, 54, 32, 44, 50, 69, 56, 37, 70, 28, 68]]
+Epoch 0: |                                                                          | 1629/? [1:34:32<00:00,  0.29it/s, v_num=xp6d]train step 1630; scene = [['4fa0b5f07c785cd6'], ['a3c9ffcea941f481']]; loss = 0.017137
+Epoch 0: |                                                                          | 1630/? [1:34:35<00:00,  0.29it/s, v_num=xp6d]context = [[91, 102, 107, 122, 123, 129, 132, 145], [2, 16, 17, 21, 53, 54, 57, 58], [0, 15, 21, 23, 26, 45, 53, 57]]target = [[119, 141, 113, 107, 127, 132, 133, 110], [38, 32, 45, 46, 12, 51, 52, 49], [37, 6, 43, 20, 28, 36, 29, 15]]
+Epoch 0: |                                                                          | 1639/? [1:35:05<00:00,  0.29it/s, v_num=xp6d]train step 1640; scene = [['a802441c288b598f'], ['8c8616ef6d449859'], ['51386d82afab8d56'], ['a5c1c97ecddf796a'], ['c26c2e3257c2d464'], ['ca35b3630b4f28ec'], ['6c10df89a6e8c8c1'], ['1a5022549d2718ae']]; loss = 0.024422
+Epoch 0: |                                                                          | 1640/? [1:35:08<00:00,  0.29it/s, v_num=xp6d]context = [[97, 98, 101, 102, 108, 115, 128, 139, 142, 143, 144, 156], [3, 5, 6, 7, 13, 20, 33, 45, 48, 49, 68, 77]]target = [[149, 98, 122, 102, 128, 120, 105, 138, 147, 110, 101, 117], [66, 25, 54, 52, 47, 38, 55, 27, 69, 59, 44, 48]]
+Epoch 0: |                                                                          | 1649/? [1:35:39<00:00,  0.29it/s, v_num=xp6d]train step 1650; scene = [['193e4b6a051b246b']]; loss = 0.007202
+Epoch 0: |                                                                          | 1650/? [1:35:42<00:00,  0.29it/s, v_num=xp6d]context = [[83, 89, 99, 100, 105, 107, 108, 118, 119, 120, 129, 134, 137, 138, 139, 140, 144, 145, 163, 164, 168, 174, 179, 180]]target = [[141, 110, 134, 98, 164, 172, 152, 162, 135, 136, 87, 178, 114, 123, 103, 95, 115, 158, 175, 119, 153, 94, 86, 166]]
+Epoch 0: |                                                                          | 1659/? [1:36:14<00:00,  0.29it/s, v_num=xp6d]train step 1660; scene = [['891648f976ef8493'], ['635af1b06fef3489'], ['f682d1c5b8431f46']]; loss = 0.029150
+Epoch 0: |                                                                          | 1660/? [1:36:18<00:00,  0.29it/s, v_num=xp6d]context = [[144, 176, 199, 202], [27, 50, 82, 102], [93, 128, 147, 166], [11, 15, 17, 76], [22, 32, 73, 77], [10, 27, 45, 82]]target = [[147, 178, 157, 197], [78, 79, 88, 80], [137, 159, 98, 119], [14, 22, 32, 43], [76, 32, 33, 27], [56, 60, 26, 13]]
+Epoch 0: |                                                                          | 1669/? [1:36:50<00:00,  0.29it/s, v_num=xp6d]train step 1670; scene = [['e7222b23037f83f7']]; loss = 0.011006
+Epoch 0: |                                                                          | 1670/? [1:36:53<00:00,  0.29it/s, v_num=xp6d]context = [[14, 16, 22, 26, 32, 57, 62, 64], [29, 33, 56, 59, 65, 69, 73, 76], [24, 29, 31, 51, 52, 69, 71, 76]]target = [[51, 47, 56, 55, 63, 50, 46, 21], [59, 65, 69, 45, 32, 52, 51, 33], [40, 51, 38, 67, 43, 61, 44, 69]]
+Epoch 0: |                                                                          | 1679/? [1:37:25<00:00,  0.29it/s, v_num=xp6d]train step 1680; scene = [['9e80f805fd6c7b93'], ['4f37cf41bcd303b6'], ['7cbf9883cb0c72d8'], ['bee292458a6d4570'], ['0ced4934341f8375'], ['b6fd5085cd94fb12'], ['c513a3c2f59aa548'], ['6cfdd5147e9d0d35'], ['60ac2ed64f533db5'], ['5edda72d021a1bc3'], ['ad127c962ce28e63'], ['2a2e3da0444d2ecd']]; loss = 0.021772
+Epoch 0: |                                                                          | 1680/? [1:37:28<00:00,  0.29it/s, v_num=xp6d]context = [[32, 35, 43, 45, 47, 59, 64, 67, 70, 74, 79, 85, 87, 89, 90, 102, 103, 110, 112, 116, 120, 125, 128, 129]]target = [[68, 125, 115, 78, 113, 48, 34, 58, 109, 85, 81, 72, 121, 77, 82, 98, 64, 111, 63, 65, 110, 37, 60, 52]]
+Epoch 0: |                                                                          | 1689/? [1:38:00<00:00,  0.29it/s, v_num=xp6d]train step 1690; scene = [['c537c888e9db11fb'], ['b008df0261874734'], ['63341a860ea3a43a'], ['276825499ea5dd5c'], ['de34678ce0e6b0bf'], ['2b026a3e98536fb5'], ['9d8eb52936b59070'], ['03acfb465cc3a17d']]; loss = 0.017924
+Epoch 0: |                                                                          | 1690/? [1:38:04<00:00,  0.29it/s, v_num=xp6d]context = [[50, 64, 68, 70, 78, 85, 91, 99, 101, 103, 115, 118], [178, 183, 184, 189, 190, 208, 210, 216, 221, 223, 240, 241]]target = [[101, 115, 91, 81, 57, 104, 90, 74, 103, 72, 82, 71], [180, 188, 235, 224, 240, 216, 206, 192, 237, 230, 214, 200]]
+Epoch 0: |                                                                          | 1699/? [1:38:35<00:00,  0.29it/s, v_num=xp6d]train step 1700; scene = [['f7e2aedf86e18a26']]; loss = 0.006839
+Epoch 0: |                                                                          | 1700/? [1:38:39<00:00,  0.29it/s, v_num=xp6d]context = [[37, 38, 42, 75, 80, 86, 100, 117], [127, 141, 150, 155, 156, 161, 169, 172], [0, 39, 45, 51, 65, 72, 73, 75]]target = [[40, 84, 114, 96, 57, 43, 38, 56], [131, 151, 128, 149, 148, 153, 132, 136], [24, 30, 46, 28, 25, 73, 31, 49]]
+Epoch 0: |                                                                          | 1709/? [1:39:11<00:00,  0.29it/s, v_num=xp6d]train step 1710; scene = [['d23860636bd4f6b9'], ['b0c6597c77c51a8c'], ['276c268acac9eca8'], ['2702020b31c28210'], ['ddc9f94d295eec5d'], ['d52a9764036894ac'], ['94034faf0ea1937a'], ['488f101dac781951']]; loss = 0.036832
+Epoch 0: |                                                                          | 1710/? [1:39:15<00:00,  0.29it/s, v_num=xp6d]context = [[89, 105, 115, 118, 121, 124, 125, 128, 129, 136, 157, 168], [32, 35, 38, 54, 56, 65, 68, 73, 86, 94, 98, 101]]target = [[167, 124, 131, 121, 102, 94, 93, 114, 90, 122, 106, 145], [44, 56, 57, 94, 33, 40, 68, 70, 90, 91, 82, 43]]
+Epoch 0: |                                                                          | 1719/? [1:39:45<00:00,  0.29it/s, v_num=xp6d]train step 1720; scene = [['af71d7a4edf61c31']]; loss = 0.010614
+Epoch 0: |                                                                          | 1720/? [1:39:49<00:00,  0.29it/s, v_num=xp6d]context = [[197, 215, 217, 233, 243, 247, 267, 274], [0, 7, 30, 47, 55, 69, 78, 89], [130, 136, 157, 165, 181, 182, 192, 201]]target = [[249, 263, 267, 222, 214, 208, 204, 256], [87, 47, 55, 24, 71, 58, 6, 35], [191, 143, 199, 172, 156, 139, 148, 176]]
+Epoch 0: |                                                                          | 1729/? [1:40:21<00:00,  0.29it/s, v_num=xp6d]train step 1730; scene = [['59ae8e2d2504d760']]; loss = 0.013345
+Epoch 0: |                                                                          | 1730/? [1:40:24<00:00,  0.29it/s, v_num=xp6d]context = [[12, 20, 26, 27, 35, 53, 55, 59, 60, 62, 66, 67], [12, 14, 18, 26, 31, 36, 41, 42, 48, 55, 63, 65]]target = [[24, 19, 58, 35, 48, 41, 63, 56, 62, 29, 28, 42], [42, 35, 18, 13, 39, 38, 20, 56, 52, 64, 19, 53]]
+Epoch 0: |                                                                          | 1739/? [1:40:54<00:00,  0.29it/s, v_num=xp6d]train step 1740; scene = [['af4565fb713ed79f'], ['473a917e0ee91a80']]; loss = 0.006910
+Epoch 0: |                                                                          | 1740/? [1:40:58<00:00,  0.29it/s, v_num=xp6d]context = [[20, 41, 54, 59, 61, 68], [37, 42, 48, 54, 61, 85], [116, 118, 130, 160, 161, 166], [37, 40, 68, 91, 92, 124]]target = [[31, 61, 50, 66, 46, 43], [68, 39, 48, 47, 40, 49], [157, 154, 128, 160, 141, 158], [60, 70, 98, 82, 39, 87]]
+Epoch 0: |                                                                          | 1749/? [1:41:29<00:00,  0.29it/s, v_num=xp6d]train step 1750; scene = [['bbaa2254cc65ceac'], ['8677be0fd77ef664']]; loss = 0.009957
+Epoch 0: |                                                                          | 1750/? [1:41:32<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 1750; scene = ['3b273cb40c55db95'];
+target intrinsic:  tensor(1.0504, device='cuda:0') tensor(1.0506, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(1.0313, device='cuda:0') tensor(1.0262, device='cuda:0')
+[2026-02-25 06:18:21,029][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1750/? [1:41:33<00:00,  0.29it/s, v_num=xp6d]context = [[5, 7, 12, 14, 17, 24, 29, 38, 49, 51, 57, 60], [28, 34, 39, 42, 46, 47, 50, 54, 56, 58, 71, 78]]target = [[10, 54, 15, 44, 52, 56, 6, 20, 35, 27, 30, 46], [55, 44, 50, 53, 49, 54, 62, 46, 35, 48, 61, 70]]
+Epoch 0: |                                                                          | 1759/? [1:42:03<00:00,  0.29it/s, v_num=xp6d]train step 1760; scene = [['380aa8c6d47a59a7'], ['1643b6a11a819f37'], ['c2162040f6db4d2f']]; loss = 0.009241
+Epoch 0: |                                                                          | 1760/? [1:42:07<00:00,  0.29it/s, v_num=xp6d]context = [[99, 104, 109, 112, 114, 119, 122, 130, 133, 137, 152, 161], [33, 40, 56, 59, 63, 67, 76, 84, 91, 94, 96, 97]]target = [[147, 120, 135, 124, 155, 157, 150, 138, 104, 142, 159, 109], [68, 35, 65, 51, 58, 93, 89, 49, 94, 64, 63, 72]]
+Epoch 0: |                                                                          | 1769/? [1:42:37<00:00,  0.29it/s, v_num=xp6d]train step 1770; scene = [['f63d2df8871ce70c'], ['066424b7f19bf6e5'], ['cc966a9b2af2232f']]; loss = 0.012349
+Epoch 0: |                                                                          | 1770/? [1:42:41<00:00,  0.29it/s, v_num=xp6d]context = [[7, 31, 84, 91], [5, 23, 52, 53], [45, 63, 89, 104], [70, 73, 125, 139], [7, 25, 48, 52], [95, 106, 138, 142]]target = [[36, 62, 87, 73], [16, 40, 27, 47], [64, 65, 74, 68], [98, 123, 97, 137], [19, 14, 33, 50], [125, 120, 102, 108]]
+Epoch 0: |                                                                          | 1779/? [1:43:13<00:00,  0.29it/s, v_num=xp6d]train step 1780; scene = [['c8f7ce4900a24c98'], ['966391dedc9b04e1']]; loss = 0.010337
+Epoch 0: |                                                                          | 1780/? [1:43:16<00:00,  0.29it/s, v_num=xp6d]context = [[137, 146, 150, 157, 165, 169, 173, 179, 183, 189, 194, 199], [12, 14, 17, 25, 29, 30, 33, 36, 45, 66, 77, 93]]target = [[197, 142, 143, 186, 170, 146, 147, 196, 185, 181, 179, 154], [23, 49, 46, 72, 31, 71, 17, 26, 92, 50, 55, 56]]
+Epoch 0: |                                                                          | 1789/? [1:43:47<00:00,  0.29it/s, v_num=xp6d]train step 1790; scene = [['7e934c4d5fc1df83'], ['48d1d822c47b5d0c'], ['afe6b05d0554a880'], ['32b6c71b6f77de55']]; loss = 0.024056
+Epoch 0: |                                                                          | 1790/? [1:43:50<00:00,  0.29it/s, v_num=xp6d]context = [[0, 1, 2, 3, 4, 9, 11, 26, 28, 32, 33, 36, 39, 41, 42, 47, 52, 54, 55, 67, 71, 87, 93, 97]]target = [[40, 57, 2, 6, 92, 63, 4, 13, 65, 71, 84, 66, 38, 16, 78, 55, 10, 79, 26, 34, 74, 46, 51, 14]]
+Epoch 0: |                                                                          | 1799/? [1:44:22<00:00,  0.29it/s, v_num=xp6d]train step 1800; scene = [['7e7eef981c97198b'], ['dc98908eb1cd3761'], ['2ff53fd35bc18dca']]; loss = 0.015797
+Epoch 0: |                                                                          | 1800/? [1:44:26<00:00,  0.29it/s, v_num=xp6d]context = [[17, 18, 24, 26, 35, 46, 49, 50, 60, 62, 69, 70, 72, 74, 76, 84, 85, 86, 89, 90, 104, 105, 110, 114]]target = [[97, 77, 86, 73, 24, 100, 21, 52, 78, 54, 30, 72, 41, 67, 84, 74, 94, 20, 82, 46, 65, 35, 61, 37]]
+[2026-02-25 06:21:18,231][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 1809/? [1:44:58<00:00,  0.29it/s, v_num=xp6d]train step 1810; scene = [['cd7187c5ac3ae71e'], ['8b9ad2078e503344'], ['93416bc86e84f4c5']]; loss = 0.025869
+Epoch 0: |                                                                          | 1810/? [1:45:01<00:00,  0.29it/s, v_num=xp6d]context = [[27, 28, 29, 30, 34, 39, 43, 47, 49, 55, 78, 81, 92, 94, 96, 99, 103, 104, 106, 107, 109, 111, 117, 124]]target = [[99, 84, 100, 68, 103, 70, 49, 36, 120, 65, 78, 92, 46, 58, 80, 74, 121, 57, 95, 39, 28, 59, 67, 62]]
+Epoch 0: |                                                                          | 1819/? [1:45:33<00:00,  0.29it/s, v_num=xp6d]train step 1820; scene = [['54e0c39079f7a30b']]; loss = 0.012097
+Epoch 0: |                                                                          | 1820/? [1:45:36<00:00,  0.29it/s, v_num=xp6d]context = [[47, 53, 56, 59, 61, 65, 68, 74, 76, 81, 86, 91, 92, 93, 94, 98, 106, 114, 123, 128, 129, 137, 140, 144]]target = [[120, 49, 106, 72, 133, 71, 48, 113, 138, 122, 116, 66, 134, 110, 103, 89, 76, 84, 69, 115, 61, 112, 123, 136]]
+Epoch 0: |                                                                          | 1829/? [1:46:07<00:00,  0.29it/s, v_num=xp6d]train step 1830; scene = [['1a0b5e0311249585'], ['f5d05305398238d4']]; loss = 0.030630
+Epoch 0: |                                                                          | 1830/? [1:46:10<00:00,  0.29it/s, v_num=xp6d]context = [[97, 106, 108, 114, 142, 151], [3, 15, 16, 51, 63, 70], [21, 29, 45, 46, 68, 75], [181, 182, 194, 246, 265, 268]]target = [[112, 127, 141, 134, 133, 124], [51, 26, 50, 60, 59, 30], [68, 73, 28, 22, 23, 67], [237, 206, 224, 188, 267, 187]]
+Epoch 0: |                                                                          | 1839/? [1:46:40<00:00,  0.29it/s, v_num=xp6d]train step 1840; scene = [['f30398f000a252c0'], ['2c1b49fe037d6723'], ['4a0f95a3db913b56'], ['e0e41af517c97c32'], ['8d758914077e5926'], ['7887a8b69eef3f6c'], ['11c69796af56d901'], ['9cdbe57a44e11638'], ['f98d6c32648cdc8b'], ['81f8c3d21285d1cc'], ['7391f166ec0c4169'], ['da1f9f2859b59142']]; loss = 0.029404
+Epoch 0: |                                                                          | 1840/? [1:46:44<00:00,  0.29it/s, v_num=xp6d]context = [[36, 49, 51, 53, 69, 74, 75, 80, 90, 94, 101, 113], [68, 78, 86, 87, 89, 90, 95, 113, 127, 133, 136, 143]]target = [[64, 81, 86, 101, 67, 50, 62, 37, 107, 73, 89, 71], [101, 81, 100, 114, 86, 103, 131, 72, 113, 95, 137, 102]]
+Epoch 0: |                                                                          | 1849/? [1:47:16<00:00,  0.29it/s, v_num=xp6d]train step 1850; scene = [['509210a9f0f87e8f']]; loss = 0.008306
+Epoch 0: |                                                                          | 1850/? [1:47:19<00:00,  0.29it/s, v_num=xp6d]context = [[54, 90, 135], [1, 13, 72], [13, 62, 103], [45, 77, 94], [93, 117, 167], [30, 42, 104], [44, 89, 103], [1, 3, 55]]target = [[65, 105, 131], [46, 9, 41], [102, 34, 68], [65, 91, 46], [132, 134, 125], [87, 81, 45], [65, 50, 54], [54, 13, 36]]
+Epoch 0: |                                                                          | 1859/? [1:47:51<00:00,  0.29it/s, v_num=xp6d]train step 1860; scene = [['4486fd0a58b1b39b']]; loss = 0.012648
+Epoch 0: |                                                                          | 1860/? [1:47:54<00:00,  0.29it/s, v_num=xp6d]context = [[4, 8, 15, 23, 29, 38, 46, 50], [140, 156, 174, 199, 200, 209, 215, 220], [2, 30, 36, 42, 45, 49, 62, 65]]target = [[30, 17, 16, 23, 6, 14, 37, 25], [153, 175, 149, 211, 210, 151, 148, 157], [55, 33, 11, 3, 59, 42, 25, 7]]
+Epoch 0: |                                                                          | 1869/? [1:48:25<00:00,  0.29it/s, v_num=xp6d]train step 1870; scene = [['ae9963ff8cd6c4de']]; loss = 0.014648
+Epoch 0: |                                                                          | 1870/? [1:48:29<00:00,  0.29it/s, v_num=xp6d]context = [[133, 135, 136, 144, 163, 166, 167, 172, 175, 176, 181, 189, 191, 192, 194, 195, 197, 201, 202, 204, 216, 224, 227, 230]]target = [[171, 154, 201, 193, 166, 210, 167, 150, 222, 136, 155, 159, 164, 148, 208, 188, 137, 221, 181, 216, 152, 160, 161, 226]]
+Epoch 0: |                                                                          | 1879/? [1:49:00<00:00,  0.29it/s, v_num=xp6d]train step 1880; scene = [['97eab841f52c3532']]; loss = 0.008732
+Epoch 0: |                                                                          | 1880/? [1:49:04<00:00,  0.29it/s, v_num=xp6d]context = [[9, 19, 24, 30, 37, 41, 45, 50, 51, 54, 55, 61, 65, 67, 69, 72, 75, 77, 78, 81, 87, 90, 96, 106]]target = [[95, 39, 90, 41, 67, 94, 89, 84, 66, 81, 29, 45, 40, 104, 20, 34, 74, 26, 17, 103, 102, 33, 68, 22]]
+Epoch 0: |                                                                          | 1889/? [1:49:35<00:00,  0.29it/s, v_num=xp6d]train step 1890; scene = [['c58d526cb2894a17'], ['d3a0a89d951a6101'], ['e8e48c278d3f4624']]; loss = 0.021401
+Epoch 0: |                                                                          | 1890/? [1:49:39<00:00,  0.29it/s, v_num=xp6d]context = [[79, 80, 116, 122, 127, 128, 129, 131, 139, 141, 144, 151], [12, 13, 21, 24, 26, 32, 44, 45, 53, 62, 68, 71]]target = [[94, 147, 101, 142, 150, 136, 104, 80, 140, 125, 144, 96], [31, 26, 46, 41, 39, 18, 53, 24, 23, 13, 63, 21]]
+Epoch 0: |                                                                          | 1899/? [1:50:11<00:00,  0.29it/s, v_num=xp6d]train step 1900; scene = [['7a03c6c69a883bac'], ['9467c3431e0e586c'], ['eb91b53e213504b8'], ['9e496cf3ceac8708'], ['6b01b60bd401ef3e'], ['db6ce23f93a9a0d3'], ['252ab646bcd51b88'], ['752724fcd4a2061f']]; loss = 0.027863
+Epoch 0: |                                                                          | 1900/? [1:50:14<00:00,  0.29it/s, v_num=xp6d]context = [[72, 78, 84, 85, 87, 92, 95, 99, 116, 118, 119, 123], [28, 29, 30, 32, 33, 35, 45, 62, 68, 76, 80, 100]]target = [[97, 77, 113, 99, 74, 92, 81, 90, 114, 80, 83, 119], [77, 85, 58, 45, 82, 49, 97, 80, 99, 93, 79, 53]]
+Epoch 0: |                                                                          | 1909/? [1:50:45<00:00,  0.29it/s, v_num=xp6d]train step 1910; scene = [['5f14e0e70310b263'], ['5c9877200c432073'], ['c0580f743d41ee56']]; loss = 0.007936
+Epoch 0: |                                                                          | 1910/? [1:50:48<00:00,  0.29it/s, v_num=xp6d]context = [[6, 8, 9, 11, 12, 13, 26, 32, 34, 46, 59, 67], [67, 70, 83, 85, 90, 94, 118, 122, 132, 133, 142, 146]]target = [[26, 13, 54, 52, 35, 55, 19, 56, 53, 59, 30, 36], [116, 87, 82, 79, 104, 89, 76, 126, 128, 131, 123, 68]]
+Epoch 0: |                                                                          | 1919/? [1:51:20<00:00,  0.29it/s, v_num=xp6d]train step 1920; scene = [['242ee2c8b1208167']]; loss = 0.009372
+Epoch 0: |                                                                          | 1920/? [1:51:24<00:00,  0.29it/s, v_num=xp6d]context = [[20, 30, 33, 40, 44, 45, 47, 57, 67, 70, 77, 83, 84, 88, 93, 95, 98, 102, 103, 108, 111, 112, 116, 117]]target = [[46, 105, 95, 29, 114, 78, 35, 69, 77, 44, 112, 40, 60, 74, 21, 107, 87, 38, 104, 22, 43, 27, 52, 34]]
+Epoch 0: |                                                                          | 1929/? [1:51:55<00:00,  0.29it/s, v_num=xp6d]train step 1930; scene = [['4e8e3034d0aa307b'], ['da65394ba9286a06'], ['640dafc4b5a3d491'], ['b25a0f4ffca51d79'], ['534ddfd7e12e4188'], ['db19784431c29cae']]; loss = 0.026474
+Epoch 0: |                                                                          | 1930/? [1:51:59<00:00,  0.29it/s, v_num=xp6d]context = [[2, 5, 8, 23, 32, 47, 61, 62], [0, 20, 21, 29, 32, 33, 43, 50], [27, 33, 34, 37, 51, 62, 68, 109]]target = [[13, 49, 20, 39, 23, 4, 42, 36], [14, 6, 25, 39, 20, 44, 43, 47], [105, 37, 104, 72, 75, 55, 34, 64]]
+Epoch 0: |                                                                          | 1939/? [1:52:31<00:00,  0.29it/s, v_num=xp6d]train step 1940; scene = [['861260140edeed05']]; loss = 0.017322
+Epoch 0: |                                                                          | 1940/? [1:52:34<00:00,  0.29it/s, v_num=xp6d]context = [[3, 43, 46, 55, 57, 83], [3, 24, 25, 40, 47, 54], [0, 12, 31, 36, 70, 73], [18, 31, 40, 50, 64, 69]]target = [[74, 48, 79, 6, 51, 82], [9, 10, 33, 23, 21, 36], [8, 65, 61, 53, 46, 66], [23, 19, 40, 22, 55, 47]]
+Epoch 0: |                                                                          | 1949/? [1:53:05<00:00,  0.29it/s, v_num=xp6d]train step 1950; scene = [['d19b2171a8edd830'], ['677817c638b61cbe'], ['ae1ce26f27f55ee0'], ['a54c4e390105d712'], ['9a5201a21290371f'], ['39ed8d2efe760b94'], ['789a1b00d73a2072'], ['9ad88af7e2443b55'], ['bb8df0200c9fa103'], ['e6cd96ceddb9b665'], ['27474a9b81b5b915'], ['20b2bb20840251c6']]; loss = 0.025248
+Epoch 0: |                                                                          | 1950/? [1:53:09<00:00,  0.29it/s, v_num=xp6d]context = [[7, 23, 28, 35, 43, 62], [0, 3, 14, 15, 23, 78], [27, 33, 46, 47, 82, 101], [28, 41, 42, 80, 100, 103]]target = [[55, 17, 13, 47, 33, 28], [1, 74, 3, 5, 20, 7], [63, 74, 85, 73, 41, 90], [35, 36, 99, 78, 88, 57]]
+Epoch 0: |                                                                          | 1959/? [1:53:41<00:00,  0.29it/s, v_num=xp6d]train step 1960; scene = [['c176654607fd9cce']]; loss = 0.013598
+Epoch 0: |                                                                          | 1960/? [1:53:44<00:00,  0.29it/s, v_num=xp6d]context = [[2, 8, 22, 32, 42, 44, 45, 47, 55, 57, 64, 67], [2, 29, 34, 35, 38, 42, 46, 47, 51, 57, 64, 69]]target = [[8, 32, 23, 65, 53, 19, 59, 14, 42, 11, 40, 9], [37, 24, 10, 46, 49, 23, 3, 68, 55, 44, 14, 26]]
+Epoch 0: |                                                                          | 1969/? [1:54:14<00:00,  0.29it/s, v_num=xp6d]train step 1970; scene = [['48cfc82a1cc3ee30'], ['1c92d53dec160984'], ['68f7ee0def2ecdb3'], ['c6b3169c8747d1fd']]; loss = 0.010578
+Epoch 0: |                                                                          | 1970/? [1:54:16<00:00,  0.29it/s, v_num=xp6d]context = [[189, 224, 239], [50, 66, 96], [125, 146, 182], [43, 50, 126], [118, 128, 179], [26, 77, 87], [4, 42, 68], [56, 65, 108]]target = [[236, 195, 197], [94, 79, 78], [170, 167, 143], [120, 92, 49], [169, 142, 172], [62, 75, 80], [43, 35, 40], [98, 95, 70]]
+Epoch 0: |                                                                          | 1979/? [1:54:48<00:00,  0.29it/s, v_num=xp6d]train step 1980; scene = [['01cbdc2687833ad4'], ['ad9a0c2b517fc375'], ['b6d6a1ec65fa5263'], ['64199c57dfb3dc4e'], ['6d0232cfe9cc7adc'], ['e3b385876279e0fc'], ['56c780409feaf822'], ['c841844509970531']]; loss = 0.026427
+Epoch 0: |                                                                          | 1980/? [1:54:52<00:00,  0.29it/s, v_num=xp6d]context = [[84, 88, 91, 109, 117, 140, 144, 149], [162, 178, 182, 184, 189, 210, 235, 240], [7, 18, 23, 32, 42, 68, 75, 78]]target = [[134, 147, 122, 135, 121, 142, 132, 129], [210, 186, 227, 232, 218, 193, 195, 220], [56, 72, 60, 75, 15, 27, 18, 29]]
+Epoch 0: |                                                                          | 1989/? [1:55:23<00:00,  0.29it/s, v_num=xp6d]train step 1990; scene = [['c67ada3a05b5cac2']]; loss = 0.010000
+Epoch 0: |                                                                          | 1990/? [1:55:26<00:00,  0.29it/s, v_num=xp6d]context = [[22, 30, 31, 35, 52, 59, 65, 69], [19, 37, 47, 48, 53, 64, 87, 102], [8, 19, 20, 21, 27, 52, 53, 54]]target = [[47, 45, 37, 25, 67, 29, 54, 59], [61, 20, 72, 71, 53, 36, 98, 39], [19, 39, 11, 52, 44, 13, 16, 21]]
+Epoch 0: |                                                                          | 1999/? [1:55:58<00:00,  0.29it/s, v_num=xp6d]train step 2000; scene = [['af11d275959fa201'], ['770a3d6cc6eb482b'], ['e2f2a27bfce53270']]; loss = 0.008729
+Epoch 0: |                                                                          | 2000/? [1:56:02<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 2000; scene = ['be75142d4652fe3e'];
+target intrinsic:  tensor(0.9402, device='cuda:0') tensor(0.9404, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8902, device='cuda:0') tensor(0.8910, device='cuda:0')
+[2026-02-25 06:32:50,731][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2000/? [1:56:03<00:00,  0.29it/s, v_num=xp6d]context = [[9, 26, 82], [130, 132, 192], [35, 48, 110], [83, 145, 172], [27, 95, 113], [94, 101, 144], [23, 43, 75], [8, 57, 96]]target = [[51, 71, 22], [186, 183, 166], [71, 76, 77], [84, 95, 134], [49, 43, 91], [116, 113, 131], [32, 54, 45], [65, 25, 14]]
+[2026-02-25 06:32:53,993][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2009/? [1:56:33<00:00,  0.29it/s, v_num=xp6d]train step 2010; scene = [['9f2fffeb41e0834f'], ['e718b9c40932334a'], ['8b9d2f4e94b006f6']]; loss = 0.011564
+Epoch 0: |                                                                          | 2010/? [1:56:36<00:00,  0.29it/s, v_num=xp6d]context = [[12, 38, 73, 93], [0, 7, 44, 46], [70, 79, 112, 160], [12, 48, 49, 81], [75, 78, 104, 122], [16, 67, 78, 83]]target = [[81, 53, 22, 16], [39, 32, 23, 13], [110, 107, 96, 155], [62, 26, 59, 17], [117, 100, 101, 82], [42, 34, 57, 29]]
+Epoch 0: |                                                                          | 2019/? [1:57:08<00:00,  0.29it/s, v_num=xp6d]train step 2020; scene = [['4b4f4736ef565c77'], ['0be2e01e27a683b0'], ['3c1aa6e11ad2bb0c'], ['90998c5f495333d2']]; loss = 0.015989
+Epoch 0: |                                                                          | 2020/? [1:57:11<00:00,  0.29it/s, v_num=xp6d]context = [[28, 29, 33, 46, 48, 49, 50, 56, 67, 71, 85, 86, 94, 98, 99, 104, 105, 109, 111, 116, 119, 120, 124, 125]]target = [[105, 112, 57, 45, 65, 70, 86, 82, 46, 73, 40, 94, 35, 121, 72, 120, 110, 118, 117, 56, 39, 90, 101, 98]]
+Epoch 0: |                                                                          | 2029/? [1:57:42<00:00,  0.29it/s, v_num=xp6d]train step 2030; scene = [['aa5f0f9be3bd3b2a'], ['7e8bc02377c148f9'], ['12a521cf027c35f6'], ['d8940a999b9fc9d8']]; loss = 0.014668
+Epoch 0: |                                                                          | 2030/? [1:57:46<00:00,  0.29it/s, v_num=xp6d]context = [[40, 44, 53, 57, 59, 67, 72, 76, 81, 83, 88, 91], [139, 140, 148, 154, 159, 162, 165, 171, 173, 179, 184, 193]]target = [[51, 68, 56, 70, 65, 55, 49, 48, 52, 67, 54, 69], [181, 180, 169, 185, 140, 161, 182, 192, 144, 143, 142, 148]]
+Epoch 0: |                                                                          | 2039/? [1:58:18<00:00,  0.29it/s, v_num=xp6d]train step 2040; scene = [['f6d34faab262bd11'], ['b3eb7d9489b53fa6'], ['3562ce0ddbf1d45e'], ['56b9fc4f422e8019'], ['245ec87f817cf1e4'], ['417bfbed2fe3e456'], ['5dedc6ab8fbe23d7'], ['480231b83d000bf7']]; loss = 0.016455
+Epoch 0: |                                                                          | 2040/? [1:58:21<00:00,  0.29it/s, v_num=xp6d]context = [[143, 153, 158, 159, 162, 163, 165, 167, 173, 185, 189, 198, 200, 205, 219, 221, 222, 223, 227, 230, 231, 235, 236, 240]]target = [[176, 193, 207, 188, 229, 218, 204, 236, 238, 239, 185, 152, 196, 197, 223, 150, 186, 172, 233, 162, 227, 170, 232, 160]]
+Epoch 0: |                                                                          | 2049/? [1:58:51<00:00,  0.29it/s, v_num=xp6d]train step 2050; scene = [['a660d0f19f982875'], ['09f2ad8e87f8f42f'], ['3530096ad3087c6c'], ['3c79b442b329338e']]; loss = 0.012365
+Epoch 0: |                                                                          | 2050/? [1:58:55<00:00,  0.29it/s, v_num=xp6d]context = [[47, 51, 53, 60, 61, 68, 70, 92, 98, 105, 107, 114], [3, 15, 16, 23, 27, 28, 29, 31, 38, 46, 58, 61]]target = [[108, 92, 82, 67, 68, 93, 103, 52, 107, 94, 72, 113], [52, 4, 33, 35, 54, 55, 11, 5, 18, 7, 42, 45]]
+Epoch 0: |                                                                          | 2059/? [1:59:27<00:00,  0.29it/s, v_num=xp6d]train step 2060; scene = [['a2e2c2e66db9a05c']]; loss = 0.010659
+Epoch 0: |                                                                          | 2060/? [1:59:30<00:00,  0.29it/s, v_num=xp6d]context = [[3, 31, 34, 41, 47, 57, 60, 62, 75, 76, 81, 89], [30, 40, 43, 44, 52, 69, 74, 80, 87, 89, 100, 102]]target = [[75, 37, 83, 16, 78, 52, 19, 67, 86, 57, 72, 69], [80, 42, 38, 49, 90, 50, 46, 63, 97, 70, 71, 94]]
+Epoch 0: |                                                                          | 2069/? [2:00:00<00:00,  0.29it/s, v_num=xp6d]train step 2070; scene = [['a16a0940ea913a28']]; loss = 0.008825
+Epoch 0: |                                                                          | 2070/? [2:00:04<00:00,  0.29it/s, v_num=xp6d]context = [[105, 108, 109, 111, 116, 117, 120, 122, 132, 133, 135, 139, 145, 147, 154, 155, 164, 170, 182, 184, 187, 188, 196, 202]]target = [[162, 107, 121, 131, 127, 173, 161, 134, 159, 199, 182, 133, 128, 146, 115, 189, 144, 168, 157, 178, 148, 152, 196, 117]]
+Epoch 0: |                                                                          | 2079/? [2:00:36<00:00,  0.29it/s, v_num=xp6d]train step 2080; scene = [['06ed257e33ae67f5']]; loss = 0.009111
+Epoch 0: |                                                                          | 2080/? [2:00:39<00:00,  0.29it/s, v_num=xp6d]context = [[99, 102, 114, 128, 136, 149, 150, 180], [59, 61, 86, 103, 114, 123, 125, 130], [0, 4, 7, 8, 14, 27, 36, 45]]target = [[158, 155, 114, 148, 120, 102, 178, 171], [121, 92, 93, 106, 61, 72, 91, 68], [8, 33, 39, 11, 22, 13, 38, 7]]
+Epoch 0: |                                                                          | 2089/? [2:01:11<00:00,  0.29it/s, v_num=xp6d]train step 2090; scene = [['ca0ef4335afc5644'], ['31ff05dbf772323c'], ['2f3a17c1963f8747'], ['ba57ad8ee0354f53'], ['222ef9ba2f4b2abb'], ['e98d838758c7b400'], ['c5536f755d325407'], ['204c24d2d8cc0ef9']]; loss = 0.024320
+Epoch 0: |                                                                          | 2090/? [2:01:14<00:00,  0.29it/s, v_num=xp6d]context = [[141, 178, 179, 181, 199, 203, 207, 214, 217, 218, 220, 221], [9, 14, 18, 27, 30, 33, 34, 40, 43, 52, 55, 61]]target = [[199, 202, 218, 214, 142, 197, 187, 180, 155, 144, 165, 154], [47, 21, 13, 57, 59, 52, 34, 23, 40, 31, 12, 49]]
+Epoch 0: |                                                                          | 2099/? [2:01:46<00:00,  0.29it/s, v_num=xp6d]train step 2100; scene = [['83222e191a9467b0']]; loss = 0.010271
+Epoch 0: |                                                                          | 2100/? [2:01:50<00:00,  0.29it/s, v_num=xp6d]context = [[3, 27, 56, 60], [11, 39, 51, 61], [13, 14, 58, 73], [33, 78, 94, 95], [29, 75, 81, 94], [65, 97, 98, 111]]target = [[43, 28, 40, 32], [23, 39, 45, 37], [70, 36, 20, 49], [89, 92, 76, 68], [85, 39, 55, 32], [88, 81, 104, 71]]
+Epoch 0: |                                                                          | 2109/? [2:02:20<00:00,  0.29it/s, v_num=xp6d]train step 2110; scene = [['1d4ffe48542eb14b'], ['851dd950cdbcea17'], ['235b7fe59778e1e0'], ['013ec74a4fde6737']]; loss = 0.021984
+Epoch 0: |                                                                          | 2110/? [2:02:24<00:00,  0.29it/s, v_num=xp6d]context = [[21, 30, 31, 33, 34, 35, 36, 37, 43, 45, 46, 48, 54, 55, 56, 70, 71, 76, 81, 90, 94, 100, 104, 118]]target = [[73, 52, 77, 71, 36, 69, 80, 108, 43, 81, 117, 40, 44, 115, 64, 74, 109, 32, 113, 79, 45, 85, 104, 70]]
+Epoch 0: |                                                                          | 2119/? [2:02:55<00:00,  0.29it/s, v_num=xp6d]train step 2120; scene = [['434bd83c81eae45d'], ['52e13717e8b9393d'], ['90e4f7b0d3147f4b']]; loss = 0.016336
+Epoch 0: |                                                                          | 2120/? [2:02:59<00:00,  0.29it/s, v_num=xp6d]context = [[197, 204, 207, 208, 226, 229, 232, 235, 247, 255, 266, 268], [7, 9, 31, 32, 41, 42, 46, 48, 50, 65, 77, 79]]target = [[258, 238, 222, 245, 227, 265, 246, 242, 253, 211, 229, 202], [72, 25, 51, 45, 63, 39, 68, 46, 67, 15, 21, 69]]
+Epoch 0: |                                                                          | 2129/? [2:03:31<00:00,  0.29it/s, v_num=xp6d]train step 2130; scene = [['50ad33ea0f572d25']]; loss = 0.007659
+Epoch 0: |                                                                          | 2130/? [2:03:34<00:00,  0.29it/s, v_num=xp6d]context = [[5, 30, 45, 61], [44, 45, 73, 106], [7, 27, 42, 85], [112, 122, 170, 196], [14, 29, 32, 67], [69, 71, 126, 129]]target = [[19, 47, 42, 58], [85, 65, 105, 72], [39, 71, 24, 13], [161, 172, 165, 126], [40, 18, 16, 34], [107, 109, 74, 84]]
+Epoch 0: |                                                                          | 2139/? [2:04:05<00:00,  0.29it/s, v_num=xp6d]train step 2140; scene = [['b4caf9dbe386c507'], ['7f441f51d44ecaf7'], ['db84572032e69dc0'], ['4a18f49c0e887115'], ['5799a001baf71ce5'], ['7430821f0f20c61a'], ['63232e078598c9c8'], ['f291b18d16c3b10c']]; loss = 0.021496
+Epoch 0: |                                                                          | 2140/? [2:04:08<00:00,  0.29it/s, v_num=xp6d]context = [[27, 34, 53, 54, 56, 63, 72, 84], [8, 9, 11, 15, 24, 34, 50, 61], [18, 38, 43, 45, 73, 79, 93, 101]]target = [[77, 63, 61, 46, 64, 35, 73, 31], [53, 25, 59, 13, 58, 34, 37, 56], [97, 22, 71, 90, 91, 84, 50, 72]]
+Epoch 0: |                                                                          | 2149/? [2:04:40<00:00,  0.29it/s, v_num=xp6d]train step 2150; scene = [['916b86f95631b480'], ['a2e4b97b2c9a0ae6'], ['ed22594386b563a1']]; loss = 0.010075
+Epoch 0: |                                                                          | 2150/? [2:04:43<00:00,  0.29it/s, v_num=xp6d]context = [[74, 91, 116, 149], [40, 50, 66, 115], [158, 196, 225, 245], [11, 33, 48, 65], [19, 40, 79, 98], [33, 63, 77, 80]]target = [[88, 108, 102, 117], [104, 113, 80, 96], [176, 227, 159, 162], [14, 31, 46, 63], [60, 96, 49, 70], [37, 64, 46, 48]]
+Epoch 0: |                                                                          | 2159/? [2:05:15<00:00,  0.29it/s, v_num=xp6d]train step 2160; scene = [['037b77f7cc0565e4'], ['648fc6db158f6e55'], ['465fa8314b741006']]; loss = 0.010648
+Epoch 0: |                                                                          | 2160/? [2:05:19<00:00,  0.29it/s, v_num=xp6d]context = [[48, 71, 110], [66, 105, 134], [16, 74, 75], [11, 51, 63], [23, 70, 83], [70, 79, 140], [83, 105, 130], [28, 38, 84]]target = [[89, 72, 55], [67, 99, 95], [36, 35, 17], [19, 26, 32], [45, 67, 72], [134, 91, 139], [113, 123, 112], [70, 62, 78]]
+Epoch 0: |                                                                          | 2169/? [2:05:50<00:00,  0.29it/s, v_num=xp6d]train step 2170; scene = [['a56e2c19e91b75a5'], ['d82aac6bf67b7d17'], ['d46599d6e4a2b451']]; loss = 0.010974
+Epoch 0: |                                                                          | 2170/? [2:05:54<00:00,  0.29it/s, v_num=xp6d]context = [[12, 15, 26, 27, 32, 40, 57, 64, 70, 71, 74, 78], [0, 8, 15, 21, 30, 41, 49, 50, 77, 81, 85, 88]]target = [[63, 74, 18, 46, 77, 35, 76, 68, 31, 32, 48, 65], [66, 18, 4, 85, 83, 31, 63, 24, 34, 19, 58, 23]]
+Epoch 0: |                                                                          | 2179/? [2:06:24<00:00,  0.29it/s, v_num=xp6d]train step 2180; scene = [['c31a1cb20a151f9f'], ['e45fd910f1e78bb9'], ['1241275438f193c1'], ['008ec1473e4ce029'], ['69937c2efa2c110b'], ['e0f2adb45811eca4'], ['1ceacc053b551f8d'], ['4be7b2dbfbb85134']]; loss = 0.018359
+Epoch 0: |                                                                          | 2180/? [2:06:27<00:00,  0.29it/s, v_num=xp6d]context = [[31, 39, 43, 48, 54, 56, 58, 61, 62, 66, 72, 73, 74, 75, 80, 81, 88, 93, 95, 101, 104, 115, 125, 128]]target = [[72, 114, 58, 78, 68, 125, 55, 53, 50, 77, 123, 61, 90, 115, 33, 83, 127, 102, 120, 57, 41, 98, 79, 95]]
+Epoch 0: |                                                                          | 2189/? [2:06:59<00:00,  0.29it/s, v_num=xp6d]train step 2190; scene = [['aff3adf77641ddcb'], ['58cdd230b4f5b590']]; loss = 0.014425
+Epoch 0: |                                                                          | 2190/? [2:07:03<00:00,  0.29it/s, v_num=xp6d]context = [[0, 6, 8, 12, 14, 18, 25, 28, 37, 41, 45, 46, 48, 53, 54, 58, 68, 74, 76, 85, 93, 94, 96, 97]]target = [[9, 35, 80, 72, 4, 81, 22, 86, 10, 64, 19, 93, 34, 12, 50, 82, 38, 3, 53, 66, 8, 76, 71, 90]]
+Epoch 0: |                                                                          | 2199/? [2:07:33<00:00,  0.29it/s, v_num=xp6d]train step 2200; scene = [['7d2402b99eed8e47']]; loss = 0.014506
+Epoch 0: |                                                                          | 2200/? [2:07:37<00:00,  0.29it/s, v_num=xp6d]context = [[106, 109, 113, 117, 130, 132, 142, 162, 164, 171, 175, 183], [0, 8, 10, 12, 19, 20, 22, 23, 30, 39, 58, 74]]target = [[121, 161, 131, 125, 182, 166, 158, 132, 113, 118, 139, 140], [12, 46, 34, 26, 72, 1, 27, 36, 13, 64, 17, 49]]
+[2026-02-25 06:44:28,349][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2209/? [2:08:08<00:00,  0.29it/s, v_num=xp6d]train step 2210; scene = [['0296e319701adeab']]; loss = 0.008394
+Epoch 0: |                                                                          | 2210/? [2:08:12<00:00,  0.29it/s, v_num=xp6d]context = [[77, 79, 91, 107, 136, 157], [2, 15, 19, 26, 69, 78], [6, 9, 42, 47, 56, 60], [18, 59, 62, 68, 75, 86]]target = [[84, 112, 128, 103, 87, 96], [40, 38, 45, 3, 30, 43], [37, 25, 44, 13, 50, 12], [81, 51, 56, 31, 41, 21]]
+Epoch 0: |                                                                          | 2219/? [2:08:43<00:00,  0.29it/s, v_num=xp6d]train step 2220; scene = [['f0f9dc20febf63c6'], ['247393e72009dab0'], ['85b95906da927213'], ['8b80a0150e20f2f2'], ['069c4f8d7a3225b4'], ['e92f5355816a5aef'], ['73089f2cde9f9cd0'], ['8812c0981c6f30a8']]; loss = 0.014124
+Epoch 0: |                                                                          | 2220/? [2:08:47<00:00,  0.29it/s, v_num=xp6d]context = [[146, 166, 170, 186, 196, 205, 219, 233], [3, 25, 36, 44, 46, 48, 55, 63], [12, 20, 33, 55, 68, 74, 97, 98]]target = [[169, 183, 158, 191, 198, 205, 227, 171], [48, 33, 26, 39, 28, 21, 58, 61], [63, 83, 48, 97, 20, 80, 66, 85]]
+Epoch 0: |                                                                          | 2229/? [2:09:18<00:00,  0.29it/s, v_num=xp6d]train step 2230; scene = [['873b6566b29196f3'], ['18061c7463438d89'], ['3055585671a8da3f'], ['d6a1f3e13c45df99']]; loss = 0.014487
+Epoch 0: |                                                                          | 2230/? [2:09:22<00:00,  0.29it/s, v_num=xp6d]context = [[106, 110, 132, 133, 135, 147, 162, 165], [17, 23, 50, 60, 61, 64, 66, 71], [68, 70, 78, 90, 113, 128, 132, 146]]target = [[137, 109, 143, 111, 121, 152, 139, 107], [51, 57, 48, 20, 18, 59, 24, 70], [92, 98, 89, 79, 125, 145, 134, 91]]
+Epoch 0: |                                                                          | 2239/? [2:09:54<00:00,  0.29it/s, v_num=xp6d]train step 2240; scene = [['513826433660cb1c'], ['53d5559c250a0f44'], ['f44cc142d9796ff7'], ['f199ea57262f903a'], ['bf00a8f83ba6fb09'], ['6b618bf721772750']]; loss = 0.023518
+Epoch 0: |                                                                          | 2240/? [2:09:57<00:00,  0.29it/s, v_num=xp6d]context = [[33, 41, 43, 61, 63, 102, 107, 116], [215, 228, 229, 235, 240, 266, 269, 274], [19, 36, 38, 42, 61, 89, 94, 97]]target = [[114, 45, 55, 110, 100, 79, 63, 66], [251, 235, 227, 267, 219, 257, 269, 258], [72, 29, 43, 69, 55, 88, 58, 76]]
+Epoch 0: |                                                                          | 2249/? [2:10:29<00:00,  0.29it/s, v_num=xp6d]train step 2250; scene = [['75cfeb226b47a5a1'], ['ff70172dd4ad43ec'], ['496e2618d0fb9556']]; loss = 0.012653
+Epoch 0: |                                                                          | 2250/? [2:10:32<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 2250; scene = ['651a7f83ed093001'];
+target intrinsic:  tensor(0.8796, device='cuda:0') tensor(0.8798, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9353, device='cuda:0') tensor(0.9337, device='cuda:0')
+[2026-02-25 06:47:21,393][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2250/? [2:10:33<00:00,  0.29it/s, v_num=xp6d]context = [[11, 13, 18, 19, 21, 29, 35, 39, 41, 44, 59, 61, 65, 66, 69, 70, 73, 84, 86, 91, 94, 95, 100, 108]]target = [[51, 83, 56, 90, 33, 46, 48, 26, 84, 28, 80, 105, 77, 69, 42, 41, 63, 76, 15, 64, 13, 87, 27, 47]]
+Epoch 0: |                                                                          | 2259/? [2:11:05<00:00,  0.29it/s, v_num=xp6d]train step 2260; scene = [['7277bcf45f9b2f21'], ['a4ccb53922783c60'], ['83bf7f4e33c13400']]; loss = 0.009372
+Epoch 0: |                                                                          | 2260/? [2:11:09<00:00,  0.29it/s, v_num=xp6d]context = [[27, 37, 42, 64, 74, 83, 100, 107], [13, 15, 21, 23, 24, 37, 39, 65], [5, 11, 13, 17, 28, 32, 51, 71]]target = [[96, 45, 92, 58, 29, 79, 90, 97], [28, 50, 36, 23, 49, 62, 37, 64], [58, 23, 41, 15, 64, 27, 12, 7]]
+Epoch 0: |                                                                          | 2269/? [2:11:39<00:00,  0.29it/s, v_num=xp6d]train step 2270; scene = [['68530d6f86185a8c']]; loss = 0.015185
+Epoch 0: |                                                                          | 2270/? [2:11:43<00:00,  0.29it/s, v_num=xp6d]context = [[167, 189, 197, 200, 204, 216, 219, 220, 225, 227, 234, 235], [8, 12, 16, 22, 23, 31, 33, 35, 39, 41, 42, 61]]target = [[184, 220, 193, 233, 225, 170, 215, 180, 186, 227, 207, 206], [27, 40, 18, 35, 38, 37, 45, 24, 22, 25, 59, 36]]
+Epoch 0: |                                                                          | 2279/? [2:12:15<00:00,  0.29it/s, v_num=xp6d]train step 2280; scene = [['f1b09696ccbe7ed0']]; loss = 0.016914
+Epoch 0: |                                                                          | 2280/? [2:12:18<00:00,  0.29it/s, v_num=xp6d]context = [[9, 26, 30, 50, 58, 73, 83, 84], [45, 46, 76, 80, 81, 88, 93, 103], [5, 26, 29, 33, 37, 50, 52, 75]]target = [[43, 51, 31, 78, 76, 27, 15, 74], [55, 57, 94, 70, 99, 46, 81, 53], [10, 47, 74, 20, 27, 61, 56, 14]]
+Epoch 0: |                                                                          | 2289/? [2:12:49<00:00,  0.29it/s, v_num=xp6d]train step 2290; scene = [['077a53a763c9b806'], ['a5d23080c6b858b0']]; loss = 0.016943
+Epoch 0: |                                                                          | 2290/? [2:12:52<00:00,  0.29it/s, v_num=xp6d]context = [[23, 31, 35, 42, 46, 48, 49, 51, 56, 61, 62, 65, 75, 77, 79, 84, 85, 86, 93, 100, 103, 106, 118, 120]]target = [[42, 67, 85, 32, 113, 77, 84, 44, 103, 81, 73, 27, 75, 35, 41, 87, 61, 53, 48, 37, 99, 110, 24, 86]]
+Epoch 0: |                                                                          | 2299/? [2:13:24<00:00,  0.29it/s, v_num=xp6d]train step 2300; scene = [['1f6b10c1bfeab825'], ['fefab354938651ee'], ['773592c80a3cbf02']]; loss = 0.011730
+Epoch 0: |                                                                          | 2300/? [2:13:27<00:00,  0.29it/s, v_num=xp6d]context = [[11, 29, 34, 37, 43, 44, 45, 67, 80, 81, 85, 87], [25, 33, 39, 42, 53, 54, 66, 67, 69, 70, 76, 78]]target = [[58, 78, 21, 29, 35, 65, 76, 43, 64, 12, 42, 15], [57, 34, 75, 45, 35, 60, 77, 67, 76, 62, 50, 36]]
+Epoch 0: |                                                                          | 2309/? [2:13:57<00:00,  0.29it/s, v_num=xp6d]train step 2310; scene = [['6349ab522cc11c42'], ['96bab6e15eb77d4e'], ['3d5b0940a28bb67d'], ['1d58a359467ff24f'], ['60c37b519a01205d'], ['6aeff4800dc767c8'], ['56a0da3c4605e004'], ['3c52bbe20a514f48'], ['f4e639c0c0c58d2f'], ['ebab00ac8428f90a'], ['57a8005cadd17369'], ['ac5159b7b9d8ab53']]; loss = 0.060998
+Epoch 0: |                                                                          | 2310/? [2:14:01<00:00,  0.29it/s, v_num=xp6d]context = [[73, 83, 87, 89, 112, 133], [121, 130, 190, 192, 197, 201], [140, 172, 174, 179, 185, 190], [22, 37, 52, 58, 63, 78]]target = [[129, 119, 117, 112, 74, 118], [186, 161, 184, 197, 195, 156], [177, 159, 184, 147, 143, 163], [39, 53, 41, 25, 42, 77]]
+Epoch 0: |                                                                          | 2319/? [2:14:33<00:00,  0.29it/s, v_num=xp6d]train step 2320; scene = [['ce50e60b3d231911'], ['b43b9870fd778691'], ['7f52ee494a9f69f1'], ['c48ef8fab1483416']]; loss = 0.008982
+Epoch 0: |                                                                          | 2320/? [2:14:36<00:00,  0.29it/s, v_num=xp6d]context = [[48, 49, 54, 64, 69, 71, 76, 86, 89, 92, 93, 95, 97, 105, 108, 109, 113, 119, 120, 123, 125, 126, 133, 145]]target = [[134, 94, 61, 54, 51, 119, 72, 68, 100, 50, 79, 104, 85, 52, 128, 89, 103, 116, 70, 57, 115, 130, 66, 62]]
+Epoch 0: |                                                                          | 2329/? [2:15:08<00:00,  0.29it/s, v_num=xp6d]train step 2330; scene = [['52b984bea1f78ad6'], ['f5d47cbf373ef48c']]; loss = 0.012128
+Epoch 0: |                                                                          | 2330/? [2:15:11<00:00,  0.29it/s, v_num=xp6d]context = [[12, 33, 40, 47, 53, 54, 63, 66], [55, 74, 105, 106, 114, 115, 117, 137], [41, 46, 55, 59, 63, 73, 85, 97]]target = [[14, 56, 35, 30, 42, 57, 40, 53], [118, 82, 89, 122, 59, 87, 79, 92], [93, 70, 85, 86, 79, 73, 88, 84]]
+Epoch 0: |                                                                          | 2339/? [2:15:42<00:00,  0.29it/s, v_num=xp6d]train step 2340; scene = [['f353f0555d221d44'], ['871ab234bc798889'], ['58a049a2126413d5'], ['6f29c6e5920ebf58'], ['6b549d67d37b3d6a'], ['4c768903212b223d'], ['0e7f80803adb73f5'], ['e39974ca56f48929'], ['7e8beb820f189792'], ['da3a6901f12a6e59'], ['de1c9674255ba611'], ['af1aad7fedbf1fc7']]; loss = 0.020657
+Epoch 0: |                                                                          | 2340/? [2:15:46<00:00,  0.29it/s, v_num=xp6d]context = [[0, 8, 17, 20, 27, 44, 47, 49], [30, 43, 45, 48, 72, 79, 90, 99], [33, 35, 45, 47, 54, 61, 66, 85]]target = [[34, 28, 7, 10, 9, 32, 2, 3], [57, 45, 73, 41, 78, 39, 53, 51], [45, 81, 42, 47, 50, 59, 44, 62]]
+Epoch 0: |                                                                          | 2349/? [2:16:17<00:00,  0.29it/s, v_num=xp6d]train step 2350; scene = [['77542628c9955b43'], ['b6f9cfe435a0fde7']]; loss = 0.009000
+Epoch 0: |                                                                          | 2350/? [2:16:21<00:00,  0.29it/s, v_num=xp6d]context = [[19, 22, 24, 28, 31, 35, 36, 37, 38, 43, 48, 54, 58, 59, 60, 62, 71, 74, 79, 80, 82, 88, 94, 116]]target = [[80, 48, 21, 102, 57, 91, 39, 74, 89, 67, 60, 87, 35, 27, 43, 85, 77, 61, 76, 45, 41, 23, 33, 81]]
+Epoch 0: |                                                                          | 2359/? [2:16:53<00:00,  0.29it/s, v_num=xp6d]train step 2360; scene = [['a3dee13fa0216d57'], ['f77ef6219492d962'], ['5bfcfd9def7a2b51'], ['fd1737c9d5f1cb37'], ['62e8f32c902fe339'], ['fb32daf720b64da7'], ['f46d9b174fbfd403'], ['3828ad130b1c93b4']]; loss = 0.016803
+Epoch 0: |                                                                          | 2360/? [2:16:56<00:00,  0.29it/s, v_num=xp6d]context = [[13, 14, 15, 20, 28, 33, 43, 70, 72, 73, 76, 78], [126, 134, 137, 165, 167, 168, 174, 187, 193, 197, 202, 212]]target = [[52, 44, 31, 72, 30, 50, 70, 62, 16, 76, 69, 15], [192, 186, 199, 161, 138, 194, 145, 183, 191, 147, 140, 170]]
+Epoch 0: |                                                                          | 2369/? [2:17:28<00:00,  0.29it/s, v_num=xp6d]train step 2370; scene = [['0a13c96ef5627832']]; loss = 0.013124
+Epoch 0: |                                                                          | 2370/? [2:17:31<00:00,  0.29it/s, v_num=xp6d]context = [[89, 91, 103, 138], [5, 7, 31, 67], [29, 47, 58, 102], [16, 45, 49, 61], [3, 13, 40, 51], [3, 17, 52, 54]]target = [[113, 98, 128, 99], [40, 56, 10, 52], [89, 55, 53, 97], [26, 39, 28, 41], [24, 23, 40, 12], [20, 48, 30, 35]]
+Epoch 0: |                                                                          | 2379/? [2:18:03<00:00,  0.29it/s, v_num=xp6d]train step 2380; scene = [['43e59b21a3805c73']]; loss = 0.016114
+Epoch 0: |                                                                          | 2380/? [2:18:06<00:00,  0.29it/s, v_num=xp6d]context = [[1, 7, 38, 48, 73, 75, 87, 91], [3, 6, 10, 20, 22, 40, 47, 48], [136, 152, 165, 167, 168, 169, 189, 195]]target = [[66, 3, 70, 59, 50, 65, 2, 86], [26, 22, 24, 35, 21, 37, 10, 6], [153, 190, 167, 166, 192, 172, 143, 183]]
+Epoch 0: |                                                                          | 2389/? [2:18:38<00:00,  0.29it/s, v_num=xp6d]train step 2390; scene = [['1559e264c5f4c481'], ['6db9c9051381e481'], ['89fc915d613ff37f'], ['f5f5c6090e031091']]; loss = 0.012018
+Epoch 0: |                                                                          | 2390/? [2:18:42<00:00,  0.29it/s, v_num=xp6d]context = [[138, 141, 142, 146, 151, 160, 166, 181, 187, 191, 199, 202], [5, 9, 17, 18, 29, 35, 39, 40, 47, 52, 70, 79]]target = [[196, 162, 177, 163, 155, 181, 161, 146, 178, 199, 190, 154], [58, 48, 23, 36, 42, 18, 13, 60, 49, 69, 63, 71]]
+Epoch 0: |                                                                          | 2399/? [2:19:12<00:00,  0.29it/s, v_num=xp6d]train step 2400; scene = [['c411d27c23509ccb'], ['521aa01f117a1f3d'], ['b65dd487d8a9df81'], ['a30e87ca123b7dcc']]; loss = 0.012006
+Epoch 0: |                                                                          | 2400/? [2:19:15<00:00,  0.29it/s, v_num=xp6d]context = [[132, 142, 149, 151, 153, 158, 162, 166, 179, 180, 184, 187, 190, 191, 192, 200, 203, 208, 212, 218, 220, 221, 222, 229]]target = [[203, 210, 195, 163, 183, 187, 226, 212, 179, 142, 177, 216, 190, 139, 213, 169, 161, 154, 135, 196, 151, 224, 137, 145]]
+[2026-02-25 06:56:07,075][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2409/? [2:19:47<00:00,  0.29it/s, v_num=xp6d]train step 2410; scene = [['d35650707515c2cf'], ['3c60078735491108'], ['7f6de4b0a9fc2031']]; loss = 0.010393
+Epoch 0: |                                                                          | 2410/? [2:19:50<00:00,  0.29it/s, v_num=xp6d]context = [[20, 21, 32, 37, 38, 39, 42, 43, 61, 65, 66, 67, 68, 72, 74, 75, 78, 83, 86, 102, 103, 108, 116, 117]]target = [[27, 41, 58, 112, 75, 82, 94, 61, 96, 116, 30, 56, 69, 109, 50, 64, 44, 86, 79, 104, 59, 68, 46, 93]]
+Epoch 0: |                                                                          | 2419/? [2:20:21<00:00,  0.29it/s, v_num=xp6d]train step 2420; scene = [['894851ce05febce5']]; loss = 0.012209
+Epoch 0: |                                                                          | 2420/? [2:20:25<00:00,  0.29it/s, v_num=xp6d]context = [[41, 58, 59, 65, 68, 71, 80, 129], [176, 180, 184, 192, 209, 215, 216, 230], [105, 107, 111, 113, 115, 153, 154, 161]]target = [[42, 95, 110, 106, 113, 82, 47, 46], [205, 198, 190, 183, 203, 206, 201, 180], [134, 142, 114, 111, 150, 135, 120, 115]]
+Epoch 0: |                                                                          | 2429/? [2:20:56<00:00,  0.29it/s, v_num=xp6d]train step 2430; scene = [['8a6f6565aa136bfd']]; loss = 0.009340
+Epoch 0: |                                                                          | 2430/? [2:21:00<00:00,  0.29it/s, v_num=xp6d]context = [[16, 25, 29, 30, 46, 63, 68, 69], [20, 24, 48, 50, 55, 64, 80, 81], [212, 220, 236, 237, 241, 250, 257, 261]]target = [[37, 67, 36, 34, 27, 33, 23, 35], [41, 44, 56, 25, 30, 21, 52, 75], [227, 220, 237, 215, 247, 239, 260, 236]]
+Epoch 0: |                                                                          | 2439/? [2:21:30<00:00,  0.29it/s, v_num=xp6d]train step 2440; scene = [['d2bf3cc61876b88a'], ['41d454a6a4ccc3c6'], ['50c8a233bd82a613'], ['1b8238f8057a7975'], ['5e89c7650c97cdf1'], ['62b4841181d253e7'], ['49328b9c29331060'], ['77c6dd845fd239ec']]; loss = 0.020521
+Epoch 0: |                                                                          | 2440/? [2:21:34<00:00,  0.29it/s, v_num=xp6d]context = [[12, 13, 20, 27, 30, 31, 39, 40, 41, 43, 44, 53, 56, 57, 60, 66, 75, 79, 82, 83, 86, 100, 107, 109]]target = [[106, 96, 48, 79, 36, 108, 37, 25, 43, 88, 38, 84, 83, 21, 76, 58, 33, 94, 73, 78, 30, 53, 64, 99]]
+Epoch 0: |                                                                          | 2449/? [2:22:05<00:00,  0.29it/s, v_num=xp6d]train step 2450; scene = [['a8abd4eeea08e1ad']]; loss = 0.008887
+Epoch 0: |                                                                          | 2450/? [2:22:09<00:00,  0.29it/s, v_num=xp6d]context = [[15, 21, 22, 26, 29, 34, 40, 41, 46, 49, 51, 52, 55, 62, 63, 64, 65, 68, 70, 77, 78, 84, 110, 112]]target = [[83, 49, 68, 55, 107, 44, 81, 102, 97, 89, 86, 19, 32, 21, 52, 62, 95, 64, 61, 88, 59, 101, 22, 18]]
+Epoch 0: |                                                                          | 2459/? [2:22:41<00:00,  0.29it/s, v_num=xp6d]train step 2460; scene = [['9d3f87d62d49025d'], ['c152ab577a837dac']]; loss = 0.013133
+Epoch 0: |                                                                          | 2460/? [2:22:44<00:00,  0.29it/s, v_num=xp6d]context = [[176, 177, 180, 212, 215, 228], [18, 27, 52, 75, 98, 101], [17, 38, 52, 81, 85, 104], [7, 36, 82, 90, 92, 95]]target = [[200, 210, 201, 198, 191, 216], [40, 53, 56, 91, 23, 48], [67, 37, 35, 81, 39, 41], [73, 92, 61, 29, 46, 39]]
+Epoch 0: |                                                                          | 2469/? [2:23:16<00:00,  0.29it/s, v_num=xp6d]train step 2470; scene = [['5614d74255c9c07c'], ['abe30f78eedbc519'], ['8ef9ff3189c85eee'], ['ff66444e1858620a'], ['4976860ffb8e6b02'], ['4af1f08d7b5d1523']]; loss = 0.014806
+Epoch 0: |                                                                          | 2470/? [2:23:20<00:00,  0.29it/s, v_num=xp6d]context = [[12, 14, 18, 30, 40, 48, 49, 67], [57, 60, 67, 84, 88, 97, 121, 124], [143, 160, 169, 181, 185, 190, 198, 212]]target = [[15, 13, 65, 59, 38, 63, 48, 20], [72, 87, 80, 69, 114, 93, 81, 99], [168, 172, 179, 187, 173, 149, 180, 189]]
+Epoch 0: |                                                                          | 2479/? [2:23:51<00:00,  0.29it/s, v_num=xp6d]train step 2480; scene = [['1a402532663d05ad'], ['36b937cc3684eddb'], ['08da23838ee6e23b'], ['ef29eabcdae21636']]; loss = 0.014571
+Epoch 0: |                                                                          | 2480/? [2:23:54<00:00,  0.29it/s, v_num=xp6d]context = [[29, 84, 88], [169, 172, 215], [20, 54, 69], [25, 69, 84], [211, 272, 277], [48, 95, 133], [112, 176, 185], [85, 102, 149]]target = [[59, 75, 32], [185, 192, 191], [34, 39, 56], [75, 51, 31], [256, 233, 240], [66, 98, 119], [122, 149, 170], [144, 138, 104]]
+Epoch 0: |                                                                          | 2489/? [2:24:25<00:00,  0.29it/s, v_num=xp6d]train step 2490; scene = [['efbdf67e0c80c27b'], ['982da008194b287c'], ['4eb66b49aeb1d641'], ['9c500c3d949c224e']]; loss = 0.020473
+Epoch 0: |                                                                          | 2490/? [2:24:28<00:00,  0.29it/s, v_num=xp6d]context = [[0, 5, 14, 16, 23, 25, 26, 28, 30, 34, 42, 49], [46, 60, 65, 67, 74, 75, 86, 88, 89, 90, 93, 117]]target = [[29, 26, 30, 16, 31, 11, 28, 36, 45, 18, 21, 40], [56, 104, 112, 73, 102, 57, 68, 52, 74, 91, 109, 61]]
+Epoch 0: |                                                                          | 2499/? [2:24:59<00:00,  0.29it/s, v_num=xp6d]train step 2500; scene = [['a48e4e90b76cc3a3'], ['715e8695976cdb61'], ['40b14e9ca06271a0'], ['f46f73c6b994a630'], ['4d32322df2d07217'], ['faa760d7d2c034ec']]; loss = 0.015561
+Epoch 0: |                                                                          | 2500/? [2:25:02<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 2500; scene = ['97ef4323919c5e8a'];
+target intrinsic:  tensor(0.8889, device='cuda:0') tensor(0.8892, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9282, device='cuda:0') tensor(0.9287, device='cuda:0')
+[2026-02-25 07:01:50,987][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2500/? [2:25:03<00:00,  0.29it/s, v_num=xp6d]context = [[135, 137, 148, 151, 152, 157, 185, 186, 192, 198, 200, 204, 208, 212, 214, 215, 217, 220, 221, 222, 223, 225, 226, 232]]target = [[228, 186, 188, 214, 163, 148, 155, 141, 146, 149, 164, 198, 194, 205, 145, 176, 151, 213, 229, 174, 201, 199, 181, 224]]
+Epoch 0: |                                                                          | 2509/? [2:25:35<00:00,  0.29it/s, v_num=xp6d]train step 2510; scene = [['62bfd0a1ecd6a33c'], ['86fa33e82678f2b4'], ['a7e5c2b2c8fe38eb'], ['313c2ce83a17b22d']]; loss = 0.008729
+Epoch 0: |                                                                          | 2510/? [2:25:38<00:00,  0.29it/s, v_num=xp6d]context = [[2, 3, 27, 39, 52, 60, 66, 88], [6, 22, 31, 32, 38, 39, 46, 53], [32, 41, 51, 59, 63, 65, 77, 101]]target = [[60, 18, 51, 6, 77, 10, 8, 87], [9, 46, 34, 15, 43, 31, 44, 35], [78, 55, 48, 100, 77, 84, 57, 82]]
+Epoch 0: |                                                                          | 2519/? [2:26:10<00:00,  0.29it/s, v_num=xp6d]train step 2520; scene = [['3fc603b4c4531c11']]; loss = 0.018603
+Epoch 0: |                                                                          | 2520/? [2:26:14<00:00,  0.29it/s, v_num=xp6d]context = [[22, 24, 38, 48, 62, 78, 83, 95], [5, 11, 30, 31, 37, 58, 70, 72], [23, 32, 33, 59, 94, 97, 101, 104]]target = [[94, 35, 69, 29, 28, 40, 72, 37], [42, 57, 40, 56, 18, 34, 11, 37], [97, 70, 59, 44, 51, 52, 34, 35]]
+Epoch 0: |                                                                          | 2529/? [2:26:45<00:00,  0.29it/s, v_num=xp6d]train step 2530; scene = [['7dc5c394263df267'], ['b113c611f9029c49']]; loss = 0.013811
+Epoch 0: |                                                                          | 2530/? [2:26:49<00:00,  0.29it/s, v_num=xp6d]context = [[144, 162, 163, 167, 185, 191], [44, 48, 58, 102, 106, 107], [23, 30, 71, 84, 86, 89], [1, 13, 23, 28, 54, 83]]target = [[176, 153, 187, 165, 152, 157], [76, 101, 48, 62, 78, 67], [40, 37, 41, 33, 34, 73], [51, 39, 8, 24, 70, 79]]
+Epoch 0: |                                                                          | 2539/? [2:27:21<00:00,  0.29it/s, v_num=xp6d]train step 2540; scene = [['85b281f0e77a632f'], ['66ae84fb53b12759']]; loss = 0.012758
+Epoch 0: |                                                                          | 2540/? [2:27:25<00:00,  0.29it/s, v_num=xp6d]context = [[11, 17, 21, 24, 33, 34, 50, 63, 68, 69, 72, 88], [5, 13, 17, 19, 22, 24, 31, 33, 35, 36, 53, 54]]target = [[54, 39, 52, 20, 16, 23, 77, 65, 87, 43, 66, 78], [9, 33, 36, 44, 52, 11, 6, 51, 14, 29, 22, 30]]
+Epoch 0: |                                                                          | 2549/? [2:27:56<00:00,  0.29it/s, v_num=xp6d]train step 2550; scene = [['b35b0b431fab2105']]; loss = 0.011811
+Epoch 0: |                                                                          | 2550/? [2:28:00<00:00,  0.29it/s, v_num=xp6d]context = [[96, 100, 101, 114, 118, 121, 122, 123, 124, 136, 140, 142, 149, 150, 151, 159, 160, 162, 163, 167, 169, 173, 190, 193]]target = [[165, 189, 177, 184, 168, 112, 131, 182, 138, 164, 101, 111, 192, 121, 128, 120, 132, 125, 135, 167, 163, 178, 142, 170]]
+Epoch 0: |                                                                          | 2559/? [2:28:32<00:00,  0.29it/s, v_num=xp6d]train step 2560; scene = [['987ba4e94414a901'], ['4a74477406314bef'], ['e1bfe1e13278c747'], ['917b6ab7c8384c3b'], ['7eb0dffcddc1722c'], ['0fa6bc2796cbba41'], ['a039939946c82ed7'], ['478e07af7840c08a'], ['395fd5b1237500c6'], ['50468b60ba969ac2'], ['ee7cb189dff51d83'], ['aed762fe049b1f86']]; loss = 0.021171
+Epoch 0: |                                                                          | 2560/? [2:28:35<00:00,  0.29it/s, v_num=xp6d]context = [[27, 31, 41, 55, 59, 67, 70, 73], [47, 52, 65, 80, 90, 92, 94, 100], [51, 66, 72, 99, 109, 113, 122, 124]]target = [[70, 71, 35, 43, 64, 67, 53, 52], [53, 94, 82, 67, 55, 63, 85, 88], [84, 116, 57, 61, 104, 88, 80, 79]]
+Epoch 0: |                                                                          | 2569/? [2:29:07<00:00,  0.29it/s, v_num=xp6d]train step 2570; scene = [['b7873e1ebdb5721f']]; loss = 0.014287
+Epoch 0: |                                                                          | 2570/? [2:29:10<00:00,  0.29it/s, v_num=xp6d]context = [[4, 6, 12, 16, 31, 37, 39, 40, 47, 50, 52, 63], [3, 4, 14, 22, 23, 35, 37, 38, 40, 51, 64, 74]]target = [[41, 62, 34, 35, 5, 31, 18, 24, 47, 55, 29, 10], [52, 40, 60, 63, 19, 11, 36, 66, 4, 16, 42, 18]]
+Epoch 0: |                                                                          | 2579/? [2:29:42<00:00,  0.29it/s, v_num=xp6d]train step 2580; scene = [['70b45bbd1147fbf0'], ['1b2bfb2e03827c26'], ['9ade5b1dc78259b9'], ['d79b084da2f40032']]; loss = 0.015479
+Epoch 0: |                                                                          | 2580/? [2:29:45<00:00,  0.29it/s, v_num=xp6d]context = [[0, 68], [30, 89], [53, 123], [12, 87], [0, 80], [135, 188], [11, 71], [9, 74], [1, 88], [4, 58], [4, 54], [7, 94]]target = [[53, 1], [83, 49], [118, 84], [63, 81], [25, 71], [143, 139], [66, 33], [56, 18], [46, 35], [48, 8], [23, 11], [53, 32]]
+Epoch 0: |                                                                          | 2589/? [2:30:16<00:00,  0.29it/s, v_num=xp6d]train step 2590; scene = [['21a170c902c43a97'], ['0e57d2d8655afebb'], ['362a71463cb49249'], ['02843207f75c20d3']]; loss = 0.010578
+Epoch 0: |                                                                          | 2590/? [2:30:20<00:00,  0.29it/s, v_num=xp6d]context = [[119, 139, 141, 145, 167, 171, 173, 177], [0, 3, 9, 22, 36, 42, 71, 81], [11, 16, 22, 26, 29, 30, 41, 58]]target = [[128, 160, 123, 142, 139, 157, 151, 164], [10, 34, 9, 21, 4, 65, 15, 7], [25, 14, 22, 51, 33, 45, 15, 21]]
+Epoch 0: |                                                                          | 2599/? [2:30:51<00:00,  0.29it/s, v_num=xp6d]train step 2600; scene = [['31c1855b8ad30220']]; loss = 0.007095
+Epoch 0: |                                                                          | 2600/? [2:30:55<00:00,  0.29it/s, v_num=xp6d]context = [[5, 10, 18, 19, 22, 28, 29, 34, 44, 51, 54, 60], [81, 82, 84, 90, 93, 95, 99, 109, 112, 124, 130, 131]]target = [[51, 48, 26, 14, 22, 54, 50, 23, 46, 6, 52, 53], [127, 85, 93, 121, 94, 88, 90, 101, 82, 104, 109, 108]]
+[2026-02-25 07:07:47,162][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2609/? [2:31:26<00:00,  0.29it/s, v_num=xp6d]train step 2610; scene = [['fef928bdd59753c7'], ['d715711625070e85'], ['3b7c67b6632b1c74']]; loss = 0.011501
+Epoch 0: |                                                                          | 2610/? [2:31:29<00:00,  0.29it/s, v_num=xp6d]context = [[18, 20, 39, 41, 42, 43, 51, 55, 59, 61, 62, 67], [6, 14, 16, 35, 47, 53, 64, 68, 72, 77, 82, 85]]target = [[61, 23, 64, 31, 32, 55, 41, 66, 38, 50, 63, 49], [66, 15, 52, 48, 21, 80, 51, 56, 54, 62, 32, 61]]
+Epoch 0: |                                                                          | 2619/? [2:32:01<00:00,  0.29it/s, v_num=xp6d]train step 2620; scene = [['8481bc2172478625']]; loss = 0.019770
+Epoch 0: |                                                                          | 2620/? [2:32:04<00:00,  0.29it/s, v_num=xp6d]context = [[151, 154, 158, 161, 167, 174, 180, 185, 189, 191, 193, 204, 209, 210, 211, 216, 224, 229, 237, 240, 242, 243, 244, 248]]target = [[179, 180, 165, 183, 223, 207, 176, 156, 194, 178, 243, 159, 214, 184, 210, 191, 168, 211, 237, 195, 170, 224, 166, 225]]
+Epoch 0: |                                                                          | 2629/? [2:32:34<00:00,  0.29it/s, v_num=xp6d]train step 2630; scene = [['d8ad0faac7caff87'], ['673c44a2c8176eb6'], ['02a58849f00b13fc'], ['d5f11c320ee101a1']]; loss = 0.022233
+Epoch 0: |                                                                          | 2630/? [2:32:38<00:00,  0.29it/s, v_num=xp6d]context = [[32, 77, 80], [67, 92, 114], [55, 108, 118], [153, 192, 205], [12, 21, 97], [214, 219, 273], [29, 81, 91], [145, 182, 222]]target = [[63, 68, 40], [102, 85, 77], [117, 89, 102], [169, 171, 167], [75, 68, 31], [242, 253, 252], [58, 59, 75], [198, 172, 197]]
+Epoch 0: |                                                                          | 2639/? [2:33:09<00:00,  0.29it/s, v_num=xp6d]train step 2640; scene = [['08c09a60291a2f5b'], ['210ab3c1a9d655e5'], ['44f828162649e72a'], ['363ee2c7afc81e84'], ['78b83f9c9174e8cc'], ['d8a3b176b0529293'], ['f276b95e48af6e36'], ['5ae12f0ee3f7ad6f']]; loss = 0.015007
+Epoch 0: |                                                                          | 2640/? [2:33:13<00:00,  0.29it/s, v_num=xp6d]context = [[178, 186, 187, 190, 197, 200, 203, 204, 221, 224, 227, 234], [0, 5, 15, 21, 28, 29, 33, 35, 37, 38, 49, 50]]target = [[209, 190, 203, 187, 191, 214, 205, 201, 202, 179, 212, 215], [25, 21, 46, 27, 44, 1, 40, 41, 16, 49, 35, 3]]
+Epoch 0: |                                                                          | 2649/? [2:33:45<00:00,  0.29it/s, v_num=xp6d]train step 2650; scene = [['376c7ff59babc257']]; loss = 0.011251
+Epoch 0: |                                                                          | 2650/? [2:33:48<00:00,  0.29it/s, v_num=xp6d]context = [[3, 18, 29, 58], [0, 17, 61, 84], [60, 99, 117, 131], [0, 7, 12, 45], [37, 60, 81, 101], [183, 190, 222, 244]]target = [[49, 31, 35, 25], [40, 50, 67, 10], [100, 115, 120, 107], [41, 34, 1, 29], [39, 43, 78, 67], [232, 242, 223, 191]]
+Epoch 0: |                                                                          | 2659/? [2:34:19<00:00,  0.29it/s, v_num=xp6d]train step 2660; scene = [['5d50f21bfe94df8c'], ['a0f0af3e118b6ca6'], ['1ed3d0dda9649020'], ['0e613a7c35450462'], ['985af15298bdcea5'], ['fd859fb51b16cf6e'], ['c35de22a7238f9a5'], ['4e588d9d58d1ec06']]; loss = 0.021400
+Epoch 0: |                                                                          | 2660/? [2:34:22<00:00,  0.29it/s, v_num=xp6d]context = [[12, 16, 20, 21, 23, 26, 27, 29, 32, 36, 46, 49, 55, 62, 71, 83, 84, 90, 95, 99, 100, 102, 103, 109]]target = [[34, 86, 61, 97, 67, 73, 17, 43, 72, 98, 92, 80, 62, 87, 70, 42, 85, 90, 44, 41, 77, 75, 21, 57]]
+Epoch 0: |                                                                          | 2669/? [2:34:54<00:00,  0.29it/s, v_num=xp6d]train step 2670; scene = [['ac16e227bc0144d6'], ['6b1950140a598578']]; loss = 0.010016
+Epoch 0: |                                                                          | 2670/? [2:34:58<00:00,  0.29it/s, v_num=xp6d]context = [[0, 6, 9, 19, 23, 31, 41, 48], [9, 21, 22, 23, 24, 33, 70, 72], [51, 58, 66, 72, 73, 75, 81, 97]]target = [[39, 40, 27, 21, 18, 10, 12, 5], [53, 64, 18, 47, 26, 11, 43, 30], [89, 59, 82, 80, 93, 88, 77, 79]]
+Epoch 0: |                                                                          | 2679/? [2:35:29<00:00,  0.29it/s, v_num=xp6d]train step 2680; scene = [['b6dd69cee72df5a3'], ['1fb562b09fc361ea'], ['46ed182c11fd6b04']]; loss = 0.007058
+Epoch 0: |                                                                          | 2680/? [2:35:32<00:00,  0.29it/s, v_num=xp6d]context = [[82, 83, 86, 89, 94, 96, 99, 106, 107, 109, 114, 120, 122, 130, 138, 140, 142, 145, 149, 156, 165, 170, 171, 179]]target = [[175, 93, 160, 127, 130, 125, 110, 117, 142, 153, 152, 103, 158, 85, 138, 155, 89, 119, 98, 118, 96, 154, 88, 116]]
+Epoch 0: |                                                                          | 2689/? [2:36:02<00:00,  0.29it/s, v_num=xp6d]train step 2690; scene = [['8d7519b2e98e73b0'], ['9a89163b62f0a058']]; loss = 0.013940
+Epoch 0: |                                                                          | 2690/? [2:36:06<00:00,  0.29it/s, v_num=xp6d]context = [[46, 47, 48, 93, 113, 133], [118, 122, 129, 135, 158, 166], [1, 6, 36, 44, 47, 69], [73, 102, 110, 130, 141, 142]]target = [[91, 96, 65, 88, 86, 66], [139, 120, 125, 135, 148, 153], [32, 65, 39, 55, 18, 36], [125, 86, 94, 91, 141, 83]]
+Epoch 0: |                                                                          | 2699/? [2:36:38<00:00,  0.29it/s, v_num=xp6d]train step 2700; scene = [['1b24cf4a586a15e7'], ['58901334e2d813d9']]; loss = 0.007601
+Epoch 0: |                                                                          | 2700/? [2:36:41<00:00,  0.29it/s, v_num=xp6d]context = [[3, 5, 8, 12, 15, 21, 27, 28, 29, 36, 37, 40, 41, 44, 46, 56, 58, 60, 71, 80, 82, 92, 99, 100]]target = [[9, 60, 19, 69, 72, 79, 63, 59, 55, 54, 75, 21, 29, 82, 46, 98, 40, 91, 86, 4, 71, 83, 48, 85]]
+Epoch 0: |                                                                          | 2709/? [2:37:13<00:00,  0.29it/s, v_num=xp6d]train step 2710; scene = [['ef380960d4d8a206'], ['d351ba344c572b7c'], ['4beb0dd348a2c905'], ['2f6a5dd7b6a7e992']]; loss = 0.023489
+Epoch 0: |                                                                          | 2710/? [2:37:17<00:00,  0.29it/s, v_num=xp6d]context = [[145, 155, 166, 167, 176, 194, 199, 201, 203, 204, 206, 209], [0, 5, 14, 19, 26, 29, 36, 37, 41, 43, 45, 50]]target = [[163, 201, 195, 199, 181, 185, 151, 194, 190, 154, 171, 198], [31, 29, 43, 23, 22, 34, 9, 11, 5, 44, 47, 7]]
+Epoch 0: |                                                                          | 2719/? [2:37:48<00:00,  0.29it/s, v_num=xp6d]train step 2720; scene = [['76b44b96d0d32f80'], ['a9df659c4acffb49'], ['759e2542191f378d'], ['b7003ac834dc298b'], ['4d836c051f02de01'], ['107fabf0d1dce254']]; loss = 0.016485
+Epoch 0: |                                                                          | 2720/? [2:37:52<00:00,  0.29it/s, v_num=xp6d]context = [[30, 33, 45, 50, 62, 67, 91, 97, 98, 106, 107, 115], [95, 98, 102, 104, 105, 109, 113, 116, 133, 136, 138, 147]]target = [[31, 88, 109, 102, 63, 68, 65, 93, 40, 35, 38, 39], [124, 140, 128, 139, 123, 136, 103, 134, 127, 135, 113, 121]]
+Epoch 0: |                                                                          | 2729/? [2:38:24<00:00,  0.29it/s, v_num=xp6d]train step 2730; scene = [['5c0ddb9de8c16f05'], ['e82ae746fe86d59a'], ['ea02d0f42c603c21']]; loss = 0.024763
+Epoch 0: |                                                                          | 2730/? [2:38:27<00:00,  0.29it/s, v_num=xp6d]context = [[42, 78, 108, 126], [5, 26, 68, 94], [194, 228, 242, 267], [107, 124, 139, 154], [10, 21, 31, 62], [3, 15, 31, 68]]target = [[116, 59, 85, 117], [16, 79, 78, 41], [248, 204, 196, 264], [113, 147, 111, 134], [30, 27, 31, 41], [17, 25, 7, 30]]
+Epoch 0: |                                                                          | 2739/? [2:38:59<00:00,  0.29it/s, v_num=xp6d]train step 2740; scene = [['9eba6b410f166fe0'], ['3bcf3bcfc5d5c365'], ['7d8c26c8ac910aa8'], ['571879c55dd10963'], ['0e0e4a867359f360'], ['1e7d7ef1404597f0'], ['43205c12fd83c588'], ['f25aed34a7d73d42']]; loss = 0.012241
+Epoch 0: |                                                                          | 2740/? [2:39:03<00:00,  0.29it/s, v_num=xp6d]context = [[159, 160, 163, 168, 170, 171, 179, 190, 193, 202, 214, 217, 219, 222, 229, 230, 234, 236, 238, 241, 245, 248, 250, 256]]target = [[218, 230, 175, 234, 202, 170, 210, 206, 200, 197, 231, 205, 193, 225, 166, 223, 228, 209, 211, 195, 188, 215, 199, 173]]
+Epoch 0: |                                                                          | 2749/? [2:39:32<00:00,  0.29it/s, v_num=xp6d]train step 2750; scene = [['a9cd1a8fc1fa2269']]; loss = 0.017153
+Epoch 0: |                                                                          | 2750/? [2:39:36<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 2750; scene = ['3e07add8413f8157'];
+target intrinsic:  tensor(0.8521, device='cuda:0') tensor(0.8523, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.8829, device='cuda:0') tensor(0.8795, device='cuda:0')
+[2026-02-25 07:16:25,080][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2750/? [2:39:37<00:00,  0.29it/s, v_num=xp6d]context = [[2, 14, 28, 57], [18, 38, 82, 100], [12, 34, 76, 92], [0, 46, 61, 82], [78, 86, 110, 131], [4, 6, 63, 81]]target = [[17, 12, 31, 19], [48, 27, 88, 71], [58, 73, 82, 35], [32, 23, 40, 38], [119, 105, 130, 91], [34, 32, 40, 59]]
+Epoch 0: |                                                                          | 2759/? [2:40:08<00:00,  0.29it/s, v_num=xp6d]train step 2760; scene = [['88c708a41a1fb05d']]; loss = 0.007121
+Epoch 0: |                                                                          | 2760/? [2:40:12<00:00,  0.29it/s, v_num=xp6d]context = [[15, 19, 26, 33, 35, 37, 39, 42, 44, 48, 51, 64], [13, 21, 46, 47, 52, 62, 63, 68, 72, 76, 91, 103]]target = [[60, 51, 18, 36, 23, 31, 27, 19, 39, 55, 62, 52], [44, 37, 66, 31, 32, 42, 15, 51, 74, 79, 96, 73]]
+Epoch 0: |                                                                          | 2769/? [2:40:44<00:00,  0.29it/s, v_num=xp6d]train step 2770; scene = [['0ca3f423c85f9681'], ['ecbf048504e09969'], ['edaf19c032358e0d']]; loss = 0.008473
+Epoch 0: |                                                                          | 2770/? [2:40:47<00:00,  0.29it/s, v_num=xp6d]context = [[57, 62, 65, 71, 73, 76, 86, 87, 89, 91, 92, 113, 118, 119, 121, 123, 141, 142, 144, 146, 148, 151, 152, 154]]target = [[94, 88, 60, 148, 117, 113, 66, 120, 114, 103, 112, 131, 119, 108, 78, 137, 123, 90, 99, 130, 67, 118, 129, 59]]
+Epoch 0: |                                                                          | 2779/? [2:41:19<00:00,  0.29it/s, v_num=xp6d]train step 2780; scene = [['c97fb236ee21af38'], ['4a0f95a3db913b56'], ['cbe7d9bfe38d2de8'], ['6f1508676b76c4f7'], ['3ad6f38502e90d67'], ['04c740af3de5ae4c']]; loss = 0.030128
+Epoch 0: |                                                                          | 2780/? [2:41:22<00:00,  0.29it/s, v_num=xp6d]context = [[2, 6, 7, 8, 11, 13, 14, 15, 28, 59, 60, 63, 64, 67, 68, 77, 79, 81, 84, 85, 87, 91, 93, 99]]target = [[67, 33, 38, 74, 94, 54, 47, 35, 69, 55, 97, 84, 34, 61, 22, 60, 44, 36, 26, 86, 90, 50, 30, 75]]
+Epoch 0: |                                                                          | 2789/? [2:41:54<00:00,  0.29it/s, v_num=xp6d]train step 2790; scene = [['0666c63c6c8d6a9b'], ['d0aafe6b7593a8c6'], ['5682eadfab7a6bcd'], ['08f41adb663ab4f4'], ['d8d10da8948e5676'], ['665db13a67f47d42']]; loss = 0.018275
+Epoch 0: |                                                                          | 2790/? [2:41:58<00:00,  0.29it/s, v_num=xp6d]context = [[30, 55, 75, 84, 96, 101], [22, 32, 46, 65, 102, 103], [3, 27, 42, 50, 52, 58], [11, 47, 55, 89, 90, 91]]target = [[34, 39, 69, 58, 40, 91], [100, 47, 87, 95, 77, 28], [17, 51, 34, 28, 26, 29], [41, 50, 42, 27, 86, 61]]
+Epoch 0: |                                                                          | 2799/? [2:42:29<00:00,  0.29it/s, v_num=xp6d]train step 2800; scene = [['be6ab5e68b93a77e'], ['bea0e295ee56d42c'], ['8b420b4f59fb2756'], ['c49e7882e04566c0']]; loss = 0.010730
+Epoch 0: |                                                                          | 2800/? [2:42:33<00:00,  0.29it/s, v_num=xp6d]context = [[10, 12, 13, 14, 20, 24, 25, 34, 35, 36, 44, 46, 47, 50, 55, 63, 66, 70, 79, 93, 99, 103, 106, 107]]target = [[78, 50, 100, 34, 13, 21, 89, 20, 60, 40, 88, 86, 63, 77, 106, 95, 61, 94, 66, 91, 52, 49, 82, 83]]
+[2026-02-25 07:19:25,492][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 2809/? [2:43:05<00:00,  0.29it/s, v_num=xp6d]train step 2810; scene = [['b60f8abb905bb1e1']]; loss = 0.013085
+Epoch 0: |                                                                          | 2810/? [2:43:09<00:00,  0.29it/s, v_num=xp6d]context = [[3, 19, 89], [188, 252, 275], [14, 25, 101], [3, 21, 83], [2, 56, 92], [0, 11, 58], [22, 42, 91], [4, 57, 81]]target = [[47, 18, 22], [197, 239, 260], [47, 63, 65], [73, 77, 49], [48, 9, 14], [43, 12, 10], [60, 32, 86], [20, 78, 75]]
+Epoch 0: |                                                                          | 2819/? [2:43:41<00:00,  0.29it/s, v_num=xp6d]train step 2820; scene = [['d3dddf816450d1d0'], ['795b958ce87d556a']]; loss = 0.047324
+Epoch 0: |                                                                          | 2820/? [2:43:44<00:00,  0.29it/s, v_num=xp6d]context = [[87, 93, 95, 97, 105, 110, 127, 129, 131, 132, 133, 135, 139, 140, 147, 149, 159, 161, 166, 168, 180, 181, 182, 184]]target = [[97, 127, 143, 95, 121, 96, 169, 157, 168, 142, 122, 131, 149, 136, 100, 125, 166, 116, 111, 172, 93, 163, 118, 153]]
+Epoch 0: |                                                                          | 2829/? [2:44:16<00:00,  0.29it/s, v_num=xp6d]train step 2830; scene = [['bce06efba2b1d37c']]; loss = 0.006203
+Epoch 0: |                                                                          | 2830/? [2:44:20<00:00,  0.29it/s, v_num=xp6d]context = [[3, 5, 15, 18, 38, 39, 42, 43, 53, 56, 57, 61], [11, 27, 36, 39, 45, 47, 56, 58, 69, 81, 83, 84]]target = [[5, 10, 56, 28, 23, 57, 8, 17, 20, 60, 53, 35], [39, 44, 40, 55, 12, 48, 19, 22, 58, 49, 60, 57]]
+Epoch 0: |                                                                          | 2839/? [2:44:50<00:00,  0.29it/s, v_num=xp6d]train step 2840; scene = [['489e61a289f59044'], ['35b4f262621408f8'], ['a17672177fe8f694']]; loss = 0.013650
+Epoch 0: |                                                                          | 2840/? [2:44:54<00:00,  0.29it/s, v_num=xp6d]context = [[49, 52, 55, 59, 61, 71, 83, 88, 92, 95, 99, 116], [14, 24, 39, 50, 57, 63, 66, 80, 83, 87, 100, 103]]target = [[70, 96, 60, 86, 113, 58, 85, 110, 81, 51, 89, 65], [34, 96, 39, 72, 65, 91, 24, 60, 28, 57, 19, 90]]
+Epoch 0: |                                                                          | 2849/? [2:45:26<00:00,  0.29it/s, v_num=xp6d]train step 2850; scene = [['eea9a40d34d9e8f3'], ['da887eb3ac647f2c'], ['4e6ba49f61d7c52a'], ['f1d926a3302e8534']]; loss = 0.011709
+Epoch 0: |                                                                          | 2850/? [2:45:29<00:00,  0.29it/s, v_num=xp6d]context = [[30, 31, 32, 34, 45, 51, 53, 59, 71, 72, 74, 79, 85, 88, 93, 104, 106, 112, 115, 118, 119, 123, 124, 127]]target = [[61, 117, 83, 66, 57, 48, 126, 54, 119, 112, 107, 79, 80, 74, 47, 114, 121, 78, 64, 71, 96, 116, 88, 62]]
+Epoch 0: |                                                                          | 2859/? [2:46:01<00:00,  0.29it/s, v_num=xp6d]train step 2860; scene = [['f6a68923ac68bc4a'], ['c008476d54d1ee07'], ['8572c96093dc0b71']]; loss = 0.055335
+Epoch 0: |                                                                          | 2860/? [2:46:04<00:00,  0.29it/s, v_num=xp6d]context = [[39, 41, 43, 47, 52, 54, 58, 63, 64, 66, 78, 89], [23, 29, 31, 38, 46, 48, 58, 70, 73, 79, 80, 86]]target = [[82, 66, 77, 72, 73, 69, 64, 52, 62, 68, 67, 51], [68, 33, 76, 45, 53, 55, 44, 64, 31, 37, 47, 79]]
+Epoch 0: |                                                                          | 2869/? [2:46:35<00:00,  0.29it/s, v_num=xp6d]train step 2870; scene = [['0abbdaad60ccd753']]; loss = 0.014306
+Epoch 0: |                                                                          | 2870/? [2:46:39<00:00,  0.29it/s, v_num=xp6d]context = [[44, 49, 51, 57, 92, 105], [2, 10, 30, 50, 58, 59], [68, 77, 104, 142, 145, 148], [4, 16, 32, 39, 85, 94]]target = [[65, 56, 101, 92, 49, 76], [34, 15, 13, 24, 54, 52], [143, 102, 129, 120, 130, 85], [29, 52, 8, 59, 28, 27]]
+Epoch 0: |                                                                          | 2879/? [2:47:11<00:00,  0.29it/s, v_num=xp6d]train step 2880; scene = [['8f811c66bfac4e2b'], ['f388608cf7295e88'], ['3241037fc88ad609']]; loss = 0.034105
+Epoch 0: |                                                                          | 2880/? [2:47:14<00:00,  0.29it/s, v_num=xp6d]context = [[26, 38, 39, 48, 50, 53, 54, 61, 77, 84, 94, 101], [67, 68, 71, 76, 79, 85, 97, 104, 111, 112, 118, 121]]target = [[82, 38, 79, 30, 47, 75, 78, 68, 88, 57, 98, 42], [87, 83, 104, 101, 112, 116, 105, 93, 76, 111, 113, 88]]
+Epoch 0: |                                                                          | 2889/? [2:47:44<00:00,  0.29it/s, v_num=xp6d]train step 2890; scene = [['f8b0d1e280daed5d'], ['5cd3756227c3a8c9'], ['5746d03325bf70d3'], ['c146f23a3704fb63']]; loss = 0.020344
+Epoch 0: |                                                                          | 2890/? [2:47:48<00:00,  0.29it/s, v_num=xp6d]context = [[175, 182, 188, 191, 197, 203, 205, 206, 212, 213, 220, 221, 227, 233, 234, 237, 241, 243, 248, 257, 258, 261, 271, 272]]target = [[227, 176, 252, 248, 190, 203, 228, 194, 255, 225, 205, 182, 207, 217, 238, 185, 257, 208, 268, 221, 242, 188, 231, 232]]
+Epoch 0: |                                                                          | 2899/? [2:48:20<00:00,  0.29it/s, v_num=xp6d]train step 2900; scene = [['f1a0ce57e7071dac'], ['450f13d2be008e30'], ['60e0925e79a6c253'], ['a67ed4a351f4a8c2']]; loss = 0.025384
+Epoch 0: |                                                                          | 2900/? [2:48:23<00:00,  0.29it/s, v_num=xp6d]context = [[107, 112, 118, 127, 129, 135, 143, 150, 151, 154, 163, 166], [91, 100, 105, 106, 107, 111, 118, 120, 123, 130, 132, 145]]target = [[144, 156, 129, 161, 130, 140, 121, 143, 165, 135, 146, 164], [137, 103, 130, 143, 109, 115, 140, 97, 114, 111, 112, 141]]
+Epoch 0: |                                                                          | 2909/? [2:48:55<00:00,  0.29it/s, v_num=xp6d]train step 2910; scene = [['60ebe67fcdbb0767'], ['36f89cf6ea6d7736'], ['3d88d591a0323964'], ['fef48b769b17f0ed']]; loss = 0.014394
+Epoch 0: |                                                                          | 2910/? [2:48:58<00:00,  0.29it/s, v_num=xp6d]context = [[0, 2, 56, 60, 64, 65, 76, 90], [23, 24, 26, 37, 52, 54, 67, 69], [26, 44, 56, 57, 60, 66, 80, 85]]target = [[45, 5, 35, 84, 86, 88, 12, 62], [24, 64, 47, 67, 52, 62, 44, 33], [33, 75, 62, 36, 28, 38, 43, 71]]
+Epoch 0: |                                                                          | 2919/? [2:49:30<00:00,  0.29it/s, v_num=xp6d]train step 2920; scene = [['f0f815ffa7581003']]; loss = 0.009390
+Epoch 0: |                                                                          | 2920/? [2:49:33<00:00,  0.29it/s, v_num=xp6d]context = [[17, 39, 70, 73], [53, 88, 112, 125], [8, 29, 36, 89], [77, 81, 82, 144], [15, 20, 39, 63], [122, 124, 156, 205]]target = [[43, 67, 71, 62], [101, 68, 60, 61], [75, 83, 38, 29], [79, 112, 114, 90], [30, 33, 39, 45], [152, 201, 162, 182]]
+Epoch 0: |                                                                          | 2929/? [2:50:04<00:00,  0.29it/s, v_num=xp6d]train step 2930; scene = [['f78ae8f99b60b424'], ['73aee8654106974f']]; loss = 0.013587
+Epoch 0: |                                                                          | 2930/? [2:50:08<00:00,  0.29it/s, v_num=xp6d]context = [[3, 18, 30, 37, 50, 53], [162, 163, 185, 195, 198, 216], [1, 15, 42, 45, 59, 74], [6, 27, 34, 45, 51, 59]]target = [[52, 51, 49, 50, 8, 20], [204, 164, 190, 174, 170, 166], [63, 9, 16, 72, 27, 54], [38, 20, 42, 28, 46, 19]]
+Epoch 0: |                                                                          | 2939/? [2:50:39<00:00,  0.29it/s, v_num=xp6d]train step 2940; scene = [['6e7a6ed3e8593e75'], ['4099447674fb0515']]; loss = 0.006874
+Epoch 0: |                                                                          | 2940/? [2:50:43<00:00,  0.29it/s, v_num=xp6d]context = [[13, 20, 28, 34, 50, 54, 84, 86], [15, 16, 36, 46, 52, 59, 95, 104], [58, 64, 83, 87, 103, 112, 113, 116]]target = [[41, 54, 44, 58, 42, 29, 34, 31], [79, 32, 46, 67, 78, 33, 99, 22], [68, 95, 70, 88, 104, 59, 80, 81]]
+Epoch 0: |                                                                          | 2949/? [2:51:13<00:00,  0.29it/s, v_num=xp6d]train step 2950; scene = [['2cff2a771002750a'], ['6f4cc17690dcdd2e']]; loss = 0.010025
+Epoch 0: |                                                                          | 2950/? [2:51:17<00:00,  0.29it/s, v_num=xp6d]context = [[0, 3, 7, 9, 16, 19, 20, 28, 29, 30, 35, 52, 55, 57, 58, 64, 71, 78, 79, 81, 84, 87, 93, 97]]target = [[70, 51, 67, 42, 87, 35, 3, 84, 96, 89, 46, 73, 32, 66, 12, 38, 13, 5, 14, 72, 80, 33, 23, 85]]
+Epoch 0: |                                                                          | 2959/? [2:51:47<00:00,  0.29it/s, v_num=xp6d]train step 2960; scene = [['e7fad36853161638'], ['abc9cfcdc52bb1ea'], ['b054fb5ce2f8e1f2'], ['6a94bfa75e7988c8'], ['94d919d55f8b5d90'], ['34fa2f4f1daa9b1b']]; loss = 0.021910
+Epoch 0: |                                                                          | 2960/? [2:51:51<00:00,  0.29it/s, v_num=xp6d]context = [[1, 19, 46, 59, 62, 83], [27, 46, 61, 65, 76, 86], [9, 13, 18, 52, 65, 71], [137, 139, 141, 142, 155, 182]]target = [[8, 46, 52, 39, 66, 32], [40, 59, 47, 60, 45, 50], [66, 25, 24, 36, 49, 53], [171, 152, 173, 150, 161, 144]]
+Epoch 0: |                                                                          | 2969/? [2:52:23<00:00,  0.29it/s, v_num=xp6d]train step 2970; scene = [['f8b97071c3db77f7'], ['2f42d8f78b745b8f'], ['3a19113b55068671']]; loss = 0.012855
+Epoch 0: |                                                                          | 2970/? [2:52:26<00:00,  0.29it/s, v_num=xp6d]context = [[1, 23, 50, 58], [77, 96, 101, 136], [0, 31, 43, 74], [87, 127, 137, 157], [34, 74, 83, 109], [7, 38, 39, 68]]target = [[43, 48, 2, 21], [87, 86, 106, 88], [29, 34, 39, 17], [95, 114, 115, 102], [66, 63, 81, 107], [40, 58, 21, 66]]
+Epoch 0: |                                                                          | 2979/? [2:52:56<00:00,  0.29it/s, v_num=xp6d]train step 2980; scene = [['41b80c5fb43019e5'], ['64554b0854be0a81'], ['10c467dbab0f5134']]; loss = 0.014321
+Epoch 0: |                                                                          | 2980/? [2:53:00<00:00,  0.29it/s, v_num=xp6d]context = [[58, 67, 72, 74, 75, 76, 79, 98, 99, 107, 132, 136], [11, 13, 22, 51, 53, 54, 64, 71, 72, 82, 94, 97]]target = [[87, 130, 123, 111, 120, 86, 73, 59, 95, 126, 66, 116], [16, 91, 76, 73, 35, 12, 57, 29, 56, 15, 44, 82]]
+Epoch 0: |                                                                          | 2989/? [2:53:32<00:00,  0.29it/s, v_num=xp6d]train step 2990; scene = [['5f2d539c6dcc0a82'], ['78ca16a59dd41338']]; loss = 0.009872
+Epoch 0: |                                                                          | 2990/? [2:53:35<00:00,  0.29it/s, v_num=xp6d]context = [[47, 53, 59, 63, 65, 67, 68, 69, 74, 76, 85, 86, 91, 93, 102, 103, 108, 117, 122, 131, 138, 139, 140, 144]]target = [[118, 141, 73, 69, 106, 52, 81, 102, 133, 49, 79, 88, 48, 87, 124, 132, 59, 51, 140, 100, 57, 50, 103, 56]]
+Epoch 0: |                                                                          | 2999/? [2:54:06<00:00,  0.29it/s, v_num=xp6d]train step 3000; scene = [['86276412bbdb6b7a'], ['9a94e71cea472790'], ['aa2197dadd5f5f8d'], ['7d900c809e896e32']]; loss = 0.016700
+Epoch 0: |                                                                          | 3000/? [2:54:10<00:00,  0.29it/s, v_num=xp6d]Validation epoch start on rank 0
+Validation: |                                                                                                | 0/? [00:00<?, ?it/s]validation step 3000; scene = ['1072aae07584e091'];
+target intrinsic:  tensor(0.9886, device='cuda:0') tensor(0.9889, device='cuda:0')                           | 0/1 [00:00<?, ?it/s]
+pred intrinsic:  tensor(0.9628, device='cuda:0') tensor(0.9610, device='cuda:0')
+[2026-02-25 07:31:11,035][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 3000/? [2:54:23<00:00,  0.29it/s, v_num=xp6d]context = [[61, 62, 64, 68, 78, 82, 84, 87, 90, 91, 96, 107, 109, 116, 117, 126, 130, 135, 137, 143, 150, 153, 157, 158]]target = [[155, 85, 107, 142, 83, 78, 68, 126, 117, 133, 125, 99, 121, 141, 97, 81, 67, 100, 153, 146, 84, 118, 86, 150]]
+[2026-02-25 07:31:15,205][py.warnings][WARNING] - /workspace/code/CVPR2026/src/visualization/layout.py:105: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:316.)
+  result[selector] = overlay
+
+Epoch 0: |                                                                          | 3001/? [2:54:28<00:00,  0.29it/s, v_num=xp6d]
+`Trainer.fit` stopped: `max_steps=3001` reached.
+Peak VRAM: 78.070 GB (allocated), 132.943 GB (reserved)
+Total elapsed: 2.93 hours
+Saved memory info to: /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain_SSR/peak_vram_memory.json
diff --git a/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/files/requirements.txt b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/files/requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..7fbf9096f92b53f8bb2a7e5467c79ecbe64faca5
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/files/requirements.txt
@@ -0,0 +1,172 @@
+wheel==0.45.1
+pytz==2025.2
+easydict==1.13
+antlr4-python3-runtime==4.9.3
+wadler_lindig==0.1.7
+urllib3==2.5.0
+tzdata==2025.2
+typing-inspection==0.4.1
+tabulate==0.9.0
+smmap==5.0.2
+kornia_rs==0.1.9
+setuptools==78.1.1
+safetensors==0.5.3
+PyYAML==6.0.2
+PySocks==1.7.1
+pyparsing==3.2.5
+pydantic_core==2.33.2
+pycparser==2.23
+protobuf==6.32.1
+propcache==0.3.2
+proglog==0.1.12
+fsspec==2024.6.1
+platformdirs==4.4.0
+pip==25.2
+pillow==10.4.0
+frozenlist==1.7.0
+packaging==24.2
+opt_einsum==3.4.0
+numpy==1.26.4
+ninja==1.13.0
+fonttools==4.60.0
+networkx==3.4.2
+multidict==6.6.4
+mdurl==0.1.2
+MarkupSafe==3.0.2
+kiwisolver==1.4.9
+imageio-ffmpeg==0.6.0
+idna==3.7
+hf-xet==1.1.10
+gmpy2==2.2.1
+einops==0.8.1
+filelock==3.17.0
+decorator==4.4.2
+dacite==1.9.2
+cycler==0.12.1
+colorama==0.4.6
+click==8.3.0
+nvidia-nvtx-cu12==12.8.90
+charset-normalizer==3.3.2
+certifi==2025.8.3
+beartype==0.19.0
+attrs==25.3.0
+async-timeout==5.0.1
+annotated-types==0.7.0
+aiohappyeyeballs==2.6.1
+yarl==1.20.1
+tifffile==2025.5.10
+sentry-sdk==2.39.0
+scipy==1.15.3
+pydantic==2.11.9
+pandas==2.3.2
+opencv-python==4.11.0.86
+omegaconf==2.3.0
+markdown-it-py==4.0.0
+lightning-utilities==0.14.3
+lazy_loader==0.4
+jaxtyping==0.2.37
+imageio==2.37.0
+gitdb==4.0.12
+contourpy==1.3.2
+colorspacious==1.1.2
+cffi==1.17.1
+aiosignal==1.4.0
+scikit-video==1.1.11
+scikit-image==0.25.2
+rich==14.1.0
+moviepy==1.0.3
+matplotlib==3.10.6
+hydra-core==1.3.2
+nvidia-nccl-cu12==2.27.3
+huggingface-hub==0.35.1
+GitPython==3.1.45
+brotlicffi==1.0.9.2
+aiohttp==3.12.15
+torchmetrics==1.8.2
+opt-einsum-fx==0.1.4
+kornia==0.8.1
+pytorch-lightning==2.5.1
+lpips==0.1.4
+e3nn==0.6.0
+lightning==2.5.1
+nvidia-cusparselt-cu12==0.7.1
+triton==3.4.0
+nvidia-nvjitlink-cu12==12.8.93
+nvidia-curand-cu12==10.3.9.90
+nvidia-cufile-cu12==1.13.1.3
+nvidia-cuda-runtime-cu12==12.8.90
+nvidia-cuda-nvrtc-cu12==12.8.93
+nvidia-cuda-cupti-cu12==12.8.90
+nvidia-cublas-cu12==12.8.4.1
+nvidia-cusparse-cu12==12.5.8.93
+nvidia-cufft-cu12==11.3.3.83
+nvidia-cudnn-cu12==9.10.2.21
+nvidia-cusolver-cu12==11.7.3.90
+torch==2.8.0+cu128
+torchvision==0.23.0+cu128
+torchaudio==2.8.0+cu128
+torch_scatter==2.1.2+pt28cu128
+gsplat==1.5.3
+wandb==0.25.0
+cuda-bindings==13.0.3
+cuda-pathfinder==1.3.3
+Jinja2==3.1.6
+mpmath==1.3.0
+nvidia-cublas==13.1.0.3
+nvidia-cuda-cupti==13.0.85
+nvidia-cuda-nvrtc==13.0.88
+nvidia-cuda-runtime==13.0.96
+nvidia-cudnn-cu13==9.15.1.9
+nvidia-cufft==12.0.0.61
+nvidia-cufile==1.15.1.6
+nvidia-curand==10.4.0.35
+nvidia-cusolver==12.0.4.66
+nvidia-cusparse==12.6.3.3
+nvidia-cusparselt-cu13==0.8.0
+nvidia-nccl-cu13==2.28.9
+nvidia-nvjitlink==13.0.88
+nvidia-nvshmem-cu13==3.4.5
+nvidia-nvtx==13.0.85
+requests==2.32.5
+sentencepiece==0.2.1
+sympy==1.14.0
+torchcodec==0.10.0
+torchdata==0.10.0
+torchtext==0.6.0
+anyio==4.12.0
+asttokens==3.0.1
+comm==0.2.3
+debugpy==1.8.19
+executing==2.2.1
+h11==0.16.0
+httpcore==1.0.9
+httpx==0.28.1
+ipykernel==7.1.0
+ipython==9.8.0
+ipython_pygments_lexers==1.1.1
+ipywidgets==8.1.8
+jedi==0.19.2
+jupyter_client==8.7.0
+jupyter_core==5.9.1
+jupyterlab_widgets==3.0.16
+matplotlib-inline==0.2.1
+nest-asyncio==1.6.0
+parso==0.8.5
+pexpect==4.9.0
+prompt_toolkit==3.0.52
+psutil==7.2.1
+ptyprocess==0.7.0
+pure_eval==0.2.3
+Pygments==2.19.2
+python-dateutil==2.9.0.post0
+pyzmq==27.1.0
+shellingham==1.5.4
+six==1.17.0
+stack-data==0.6.3
+tornado==6.5.4
+tqdm==4.67.1
+traitlets==5.14.3
+typer-slim==0.21.0
+typing_extensions==4.15.0
+wcwidth==0.2.14
+widgetsnbextension==4.0.15
diff --git a/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/files/wandb-summary.json b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/files/wandb-summary.json
new file mode 100644
index 0000000000000000000000000000000000000000..ebf2106f7a27f2523763e53b56f890ba52e721c3
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/files/wandb-summary.json
@@ -0,0 +1 @@
+{"val/lpips":0.3168953061103821,"train/scene_scale":1.0007801055908203,"loss/total":0.016700398176908493,"active_mask_imgs":{"_type":"images/separated","width":536,"height":800,"format":"png","count":1,"filenames":["media/images/active_mask_imgs_183_ed823f1ed9087fd38b4b.png"],"captions":["1072aae07584e091"]},"lr-AdamW/pg1":2.003594834351718e-05,"epoch":0,"lr-AdamW/pg1-momentum":0.9,"lr-AdamW/pg2":2e-05,"loss/final_3dgs/error_score":0.2064337432384491,"val/ssim":0.5433365106582642,"train/psnr_probabilistic":24.096214294433594,"_wandb":{"runtime":10484},"val/gaussian_num_ratio":0.3970947265625,"_timestamp":1.7720046765923991e+09,"train/comparison":{"_type":"images/separated","width":536,"height":6378,"format":"png","count":1,"filenames":["media/images/train/comparison_186_026c0431b4edc567ec5f.png"],"captions":[["c270572a7f5ea828"]]},"trainer/global_step":3001,"comparison":{"format":"png","count":1,"filenames":["media/images/comparison_182_cd35353e49c3d5be124c.png"],"captions":["1072aae07584e091"],"_type":"images/separated","width":1064,"height":1098},"loss/final_3dgs/mse":0.004912285134196281,"val/psnr":17.298194885253906,"error_scores":{"width":800,"height":536,"format":"png","count":1,"filenames":["media/images/error_scores_184_c5b37d084a0db7e2c0dd.png"],"captions":["1072aae07584e091"],"_type":"images/separated"},"lr-AdamW/pg2-momentum":0.9,"_step":186,"info/global_step":3000,"loss/final_3dgs/lpips":0.007674907799810171,"_runtime":10484,"loss/camera":0.00019935148884542286,"loss/scene_scale_reg":5.535334275919013e-05}
\ No newline at end of file
diff --git a/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug-core.log b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug-core.log
new file mode 100644
index 0000000000000000000000000000000000000000..16344c8ca3a1dcd53f9b6a3b948c23f6ca85f11d
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug-core.log
@@ -0,0 +1,15 @@
+{"time":"2026-02-25T04:36:37.449439041Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmp8nx4wfma/port-129608.txt","pid":129608,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-02-25T04:36:37.450239664Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":129608}
+{"time":"2026-02-25T04:36:37.450217214Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-129608-132040-797610036/socket","Net":"unix"}}
+{"time":"2026-02-25T04:36:37.621608922Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-02-25T04:36:37.632379853Z","level":"INFO","msg":"handleInformInit: received","streamId":"hluaxp6d","id":"1(@)"}
+{"time":"2026-02-25T04:36:38.083827094Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"hluaxp6d","id":"1(@)"}
+{"time":"2026-02-25T04:36:43.990186089Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"cx117qjbz6uv"}
+{"time":"2026-02-25T07:31:23.586075636Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
+{"time":"2026-02-25T07:31:23.586156297Z","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-02-25T07:31:23.586153587Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
+{"time":"2026-02-25T07:31:23.586279729Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-129608-132040-797610036/socket","Net":"unix"}}
+{"time":"2026-02-25T07:31:23.586300079Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
+{"time":"2026-02-25T07:31:24.657389232Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
+{"time":"2026-02-25T07:31:24.657433263Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
+{"time":"2026-02-25T07:31:24.657457653Z","level":"INFO","msg":"server is closed"}
diff --git a/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug-internal.log b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..23b5edfa711dc7262f610c0b8f4f02b4d40c1dff
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug-internal.log
@@ -0,0 +1,11 @@
+{"time":"2026-02-25T04:36:37.632656448Z","level":"INFO","msg":"stream: starting","core version":"0.25.0"}
+{"time":"2026-02-25T04:36:38.083360107Z","level":"INFO","msg":"stream: created new stream","id":"hluaxp6d"}
+{"time":"2026-02-25T04:36:38.083494289Z","level":"INFO","msg":"handler: started","stream_id":"hluaxp6d"}
+{"time":"2026-02-25T04:36:38.083807484Z","level":"INFO","msg":"stream: started","id":"hluaxp6d"}
+{"time":"2026-02-25T04:36:38.083822184Z","level":"INFO","msg":"writer: started","stream_id":"hluaxp6d"}
+{"time":"2026-02-25T04:36:38.083845724Z","level":"INFO","msg":"sender: started","stream_id":"hluaxp6d"}
+{"time":"2026-02-25T07:31:23.586161027Z","level":"INFO","msg":"stream: closing","id":"hluaxp6d"}
+{"time":"2026-02-25T07:31:24.4560367Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-02-25T07:31:24.656085012Z","level":"INFO","msg":"handler: closed","stream_id":"hluaxp6d"}
+{"time":"2026-02-25T07:31:24.656271365Z","level":"INFO","msg":"sender: closed","stream_id":"hluaxp6d"}
+{"time":"2026-02-25T07:31:24.656292415Z","level":"INFO","msg":"stream: closed","id":"hluaxp6d"}
diff --git a/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug.log b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..af6bab8cdbb57431c1b5af071523f96b7f0c5d86
--- /dev/null
+++ b/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug.log
@@ -0,0 +1,21 @@
+2026-02-25 04:36:37,352 INFO    MainThread:129608 [wandb_setup.py:_flush():81] Current SDK version is 0.25.0
+2026-02-25 04:36:37,352 INFO    MainThread:129608 [wandb_setup.py:_flush():81] Configure stats pid to 129608
+2026-02-25 04:36:37,352 INFO    MainThread:129608 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-02-25 04:36:37,352 INFO    MainThread:129608 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug.log
+2026-02-25 04:36:37,353 INFO    MainThread:129608 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /workspace/code/CVPR2026/outputs/ablation/re10k/ABLATION_0225_targetTrain_SSR/wandb/run-20260225_043637-hluaxp6d/logs/debug-internal.log
+2026-02-25 04:36:37,353 INFO    MainThread:129608 [wandb_init.py:init():844] calling init triggers
+2026-02-25 04:36:37,353 INFO    MainThread:129608 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
+config: {'model': {'encoder': {'name': 'dcsplat', 'input_image_shape': [518, 518], 'head_mode': 'pcd', 'num_level': 3, 'gs_param_dim': 256, 'align_corners': False, 'use_voxelize': True}, 'decoder': {'name': 'splatting_cuda', 'background_color': [0.0, 0.0, 0.0], 'make_scale_invariant': False}, 'density_control': {'name': 'density_control_module', 'mean_dim': 32, 'gs_param_dim': 256, 'refinement_layer_num': 1, 'num_level': 3, 'grad_mode': 'absgrad', 'use_mean_features': True, 'refinement_type': 'voxelize', 'refinement_hidden_dim': 32, 'aggregation_mode': 'mean', 'num_heads': 1, 'score_mode': 'absgrad', 'latent_dim': 128, 'num_latents': 64, 'num_self_attn_per_block': 2, 'voxel_size': 0.001, 'aux_refine': False, 'refine_error': False, 'use_refine_module': True, 'voxelize_activate': True, 'use_depth': False}}, 'render_loss': {'mse': {'weight': 1.0}, 'lpips': {'weight': 0.05, 'apply_after_step': 0}}, 'density_control_loss': {'error_score': {'weight': 0.01, 'log_scale': False, 'grad_scale': 10000.0, 'mode': 'original'}}, 'direct_loss': {'l1': {'weight': 0.8}, 'ssim': {'weight': 0.2}}, 'wandb': {'project': 'DCSplat', 'entity': 'scene-representation-group', 'name': 'ABLATION_0225_targetTrain_SSR', 'mode': 'online', 'tags': ['re10k', '256x256']}, 'mode': 'train', 'data_loader': {'train': {'num_workers': 16, 'persistent_workers': True, 'batch_size': 16, 'seed': 1234}, 'test': {'num_workers': 4, 'persistent_workers': False, 'batch_size': 1, 'seed': 2345}, 'val': {'num_workers': 1, 'persistent_workers': True, 'batch_size': 1, 'seed': 3456}}, 'optimizer': {'lr': 0.0002, 'warm_up_steps': 25, 'backbone_lr_multiplier': 0.1, 'backbone_trainable': 'T+H', 'accumulate': 1}, 'checkpointing': {'load': None, 'every_n_train_steps': 1500, 'save_top_k': 2, 'save_weights_only': False}, 'train': {'extended_visualization': False, 'print_log_every_n_steps': 10, 'camera_loss': 10.0, 'one_sample_validation': None, 'align_corners': False, 'intrinsic_scaling': False, 'verbose': False, 'beta_dist_param': [0.5, 4.0], 'use_refine_aux': False, 'train_target_set': True, 'train_gs_num': 1, 'ext_scale_detach': False, 'cam_scale_mode': 'sum', 'scene_scale_reg_loss': 0.01, 'train_aux': False, 'vggt_cam_loss': True, 'vggt_distil': False, 'context_view_train': False}, 'test': {'output_path': 'test/ablation/re10k', 'align_pose': False, 'pose_align_steps': 100, 'rot_opt_lr': 0.005, 'trans_opt_lr': 0.005, 'compute_scores': True, 'save_image': False, 'save_video': False, 'save_active_mask_image': False, 'save_error_score_image': False, 'save_compare': False, 'pred_intrinsic': False, 'error_threshold': 0.4, 'error_threshold_list': [0.2, 0.4, 0.6, 0.8, 1.0], 'threshold_mode': 'ratio', 'nvs_view_N_list': [3, 6, 16, 32, 64]}, 'seed': 111123, 'trainer': {'max_steps': 3001, 'val_check_interval': 250, 'gradient_clip_val': 0.5, 'num_nodes': 1}, 'dataset': {'re10k': {'make_baseline_1': True, 'relative_pose': True, 'augment': True, 'background_color': [0.0, 0.0, 0.0], 'overfit_to_scene': None, 'skip_bad_shape': True, 'view_sampler': {'name': 'bounded', 'num_target_views': 4, 'num_context_views': 2, 'min_distance_between_context_views': 45, 'max_distance_between_context_views': 90, 'min_distance_to_context_views': 0, 'warm_up_steps': 1000, 'initial_min_distance_between_context_views': 25, 'initial_max_distance_between_context_views': 25, 'same_target_gap': False, 'num_target_set': 3}, 'name': 're10k', 'roots': ['datasets/re10k'], 'input_image_shape': [256, 256], 'original_image_shape': [360, 640], 'cameras_are_circular': False, 'baseline_min': 0.001, 'baseline_max': 10000000000.0, 'max_fov': 100.0, 'dynamic_context_views': True, 'max_context_views_per_gpu': 24}}, '_wandb': {}}
+2026-02-25 04:36:37,353 INFO    MainThread:129608 [wandb_init.py:init():892] starting backend
+2026-02-25 04:36:37,621 INFO    MainThread:129608 [wandb_init.py:init():895] sending inform_init request
+2026-02-25 04:36:37,629 INFO    MainThread:129608 [wandb_init.py:init():903] backend started and connected
+2026-02-25 04:36:37,637 INFO    MainThread:129608 [wandb_init.py:init():973] updated telemetry
+2026-02-25 04:36:37,646 INFO    MainThread:129608 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
+2026-02-25 04:36:38,625 INFO    MainThread:129608 [wandb_init.py:init():1042] starting run threads in backend
+2026-02-25 04:36:38,750 INFO    MainThread:129608 [wandb_run.py:_console_start():2524] atexit reg
+2026-02-25 04:36:38,750 INFO    MainThread:129608 [wandb_run.py:_redirect():2373] redirect: wrap_raw
+2026-02-25 04:36:38,750 INFO    MainThread:129608 [wandb_run.py:_redirect():2442] Wrapping output streams.
+2026-02-25 04:36:38,750 INFO    MainThread:129608 [wandb_run.py:_redirect():2465] Redirects installed.
+2026-02-25 04:36:38,753 INFO    MainThread:129608 [wandb_init.py:init():1082] run started, returning control to user process
+2026-02-25 07:31:23,586 INFO    wandb-AsyncioManager-main:129608 [service_client.py:_forward_responses():134] Reached EOF.
+2026-02-25 07:31:23,586 INFO    wandb-AsyncioManager-main:129608 [mailbox.py:close():155] Closing mailbox, abandoning 1 handles.