Upload configs/elastic_latent_4_1_12_33.1487_psnr.yaml with huggingface_hub

configs/elastic_latent_4_1_12_33.1487_psnr.yaml

@@ -0,0 +1,339 @@
+dataset:
+  view_sampler:
+    name: no_replacement
+    num_target_views: -1
+    num_context_views: 4
+    context_views:
+    - - 0
+      - 3
+      - 6
+      - 9
+    - - 1
+      - 4
+      - 7
+      - 10
+    - - 2
+      - 5
+      - 8
+      - 11
+    test_context_views:
+    - 1
+    - 4
+    - 7
+    - 10
+    test_target_views: all
+  name: genesis
+  roots:
+  - /mnt/fast/mzhobro/datasets/soft_genesis_elastic
+  root_depths:
+  - 2
+  make_baseline_1: false
+  augment: false
+  with_mask: true
+  with_seg: true
+  with_depth: false
+  image_shape:
+  - 128
+  - 128
+  background_color:
+  - 0.0
+  - 0.0
+  - 0.0
+  cameras_are_circular: false
+  near: 1.0
+  far: 8.0
+  baseline_scale_bounds: false
+  shuffle_val: false
+  test_len: -1
+  test_chunk_interval: 5
+  test_times_per_scene: 1
+  n_step_state: 4
+  n_step_state_predict: 1
+  n_step_predict_multiple: 1
+  speedup: 2
+  workspace_limits: null
+  test_validity: false
+  overfit_to_scene: null
+model:
+  encoder:
+    name: costvolume
+    opacity_mapping:
+      initial: 0.0
+      final: 0.0
+      warm_up: 1
+    num_depth_candidates: 64
+    num_surfaces: 1
+    gaussians_per_pixel: 1
+    d_feature: 128
+    visualizer:
+      num_samples: 8
+      min_resolution: 256
+      export_ply: false
+    unimatch_weights_path: checkpoints/gmdepth-scale1-resumeflowthings-scannet-5d9d7964.pth
+    multiview_trans_attn_split: 2
+    costvolume_unet_feat_dim: 128
+    costvolume_unet_channel_mult:
+    - 1
+    - 1
+    - 1
+    costvolume_unet_attn_res:
+    - 4
+    depth_unet_feat_dim: 32
+    depth_unet_attn_res:
+    - 16
+    depth_unet_channel_mult:
+    - 1
+    - 1
+    - 1
+    - 1
+    - 1
+    downscale_factor: 4
+    shim_patch_size: 4
+    wo_depth_refine: false
+    wo_cost_volume: false
+    wo_backbone_cross_attn: false
+    wo_cost_volume_refine: false
+    use_epipolar_trans: false
+  dynamic_model:
+    model:
+      in_channels: -1
+      out_channels: -1
+      grid_size: 0.004
+      order:
+      - z
+      - z-trans
+      - hilbert
+      - hilbert-trans
+      mlp_ratio: 4
+      cls_mode: false
+      temporal_merger: true
+      with_time_emb: true
+      merge_window_size:
+      - 2
+      - 2
+      - 2
+      - 8
+      grid_mode: false
+      stride:
+      - 4
+      - 2
+      - 2
+      - 2
+      enc_depths:
+      - 2
+      - 2
+      - 2
+      - 6
+      - 2
+      enc_channels:
+      - 32
+      - 64
+      - 128
+      - 256
+      - 512
+      enc_num_head:
+      - 2
+      - 4
+      - 8
+      - 16
+      - 32
+      enc_patch_size:
+      - 1024
+      - 1024
+      - 1024
+      - 1024
+      - 1024
+      dec_depths:
+      - 2
+      - 2
+      - 2
+      - 2
+      dec_channels:
+      - 64
+      - 64
+      - 128
+      - 256
+      dec_num_head:
+      - 4
+      - 4
+      - 8
+      - 16
+      dec_patch_size:
+      - 1024
+      - 1024
+      - 1024
+      - 1024
+      qkv_bias: true
+      qk_scale: null
+      attn_drop: 0.0
+      proj_drop: 0.0
+      drop_path: 0.3
+      pre_norm: true
+      shuffle_orders: true
+      enable_rpe: false
+      enable_flash: true
+      upcast_attention: false
+      upcast_softmax: false
+      pdnorm_bn: false
+      pdnorm_ln: false
+      pdnorm_decouple: true
+      pdnorm_adaptive: false
+      pdnorm_affine: true
+      pdnorm_conditions:
+      - ScanNet
+      - S3DIS
+      - Structured3D
+    name: ptv3
+    no_xyz: true
+  decoder:
+    name: mvsplat
+  state_adapter:
+    fps_to_num_points: -1
+    gaussians:
+      sh_degree: 3
+      rotations: 4
+      scale_min: 0.0002
+      scale_max: 0.1
+      cam_scale_min: 0.5
+      cam_scale_max: 15.0
+    state_3dgs_params:
+    - means
+    - opacities
+    dynamic_latent_feature_dim: 32
+    static_latent_feature_dim: 32
+    learn_static_mask: false
+    learn_background_mask: false
+    pred_modality:
+      means: cx_x
+      harmonics: z_x
+      opacities: cx_x
+      rotations: cx_x
+      scales: cx_x
+      dynamic_latent_features: z_x
+    dynamic_state_keys:
+    - means
+    metric: 1.0
+    immune_to_static:
+    - harmonics
+    pred_inductive_bias: true
+  dynamic_model.model:
+    with_time_emb: true
+    merge_window_size: '[2,2,2,8]'
+    stride: '[4,2,2,2]'
+    grid_size: 0.004
+loss:
+  mse:
+    weight: 1.0
+  lpips:
+    weight: 0.05
+    apply_after_step: 0
+wandb:
+  project: genesis_end2end_iccv
+  entity: mzhobro
+  name: ind_4-1-1-12_c_uniform_t_-1_with_bckg-mask_with_seg-mask_regr-m_o_dyn-m_from_scratch_end2end
+  mode: online
+  log_model: false
+  id: null
+  tags:
+  - 4-1-1-12
+  - c_uniform_t_-1
+  - with_bckg-mask
+  - with_seg-mask
+  - regr-m_o
+  - dyn-m
+  - from_scratch
+  - end2end
+  - speedup_2
+  - ind
+mode: train
+data_loader:
+  dynamic_train_batch_size: false
+  train:
+    num_workers: 10
+    persistent_workers: true
+    batch_size: 4
+    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: 2000
+  cosine_lr: true
+  cosine_annealing_warmup: false
+checkpointing:
+  load: null
+  every_n_train_steps: 2000
+  save_top_k: 3
+  pretrained_model: null
+  resume: true
+train:
+  freeze_encoder: false
+  depth_mode: null
+  extended_visualization: false
+  print_log_every_n_steps: 20
+  gradient_checkpointing: true
+  dyn_model_schedule:
+    N0: -1
+    reg_weight_dict:
+      means: 0.0
+      rotations: 0.0
+      scales: 0.0
+      harmonics: 0.0
+      opacities: 0.0
+      features: 0.0
+    N1: 1000
+    N2: 10000
+    n_max_pred: 1
+    n_max_grad_pred: 12
+    probability: 0.0
+test:
+  output_path: null
+  compute_scores: true
+  eval_time_skip_steps: 0
+  save_image: true
+  save_video: true
+seed: 111123
+trainer:
+  max_epochs: -1
+  max_steps: 140001
+  val_check_interval: 400
+  gradient_clip_val: 0.5
+  num_sanity_val_steps: 0
+  num_nodes: 1
+  detect_anomaly: false
+  use_ema: false
+  log_every_n_steps: 50
+  accumulate_grad_batches: 1
+output_dir: null
+im_shape: 128
+name0: 4-1-1-12
+name1: c_uniform_t_-1
+name2: with_bckg-mask
+name3: with_seg-mask
+name4: regr-m_o
+name5: dyn-m
+name6: from_scratch
+name7: end2end
+name8: speedup_2
+name9: ind
+context: uniform
+learn_static_float: false
+b: 4
+n_state: 4
+n_state_pred: 1
+n_pred_dim_multiple: 1
+n_max_grad_pred: 12
+dyn_l: 32
+stat_l: 32
+job_id: 1
+results_dir: /lustre/fast/fast/mzhobro/results/genesis_end2end/2025-02-22/00-17-20/job_1
+project_dir: /fast/mzhobro/.cache/clus/genesis_end2end-1740179840.9840438-project
+job_dir: /lustre/fast/fast/mzhobro/results/genesis_end2end/2025-02-22/00-17-20/job_1/rerun_2025-02-28/22-02-06/job_1