debug new training

configs/local/train_config.yaml

@@ -84,7 +84,7 @@ module:
   gf_view_finder:
     t_feat_dim: 128
     pose_feat_dim: 256
-    main_feat_dim: 2048
+    main_feat_dim: 3072
     regression_head: Rx_Ry_and_T
     pose_mode: rot_matrix
     per_point_feature: False

configs/server/server_train_config.yaml

@@ -7,7 +7,7 @@ runner:
     parallel: False
     
   experiment:
-    name: full_w_global_feat_wo_local_pts_feat
+    name: debug
     root_dir: "experiments"
     use_checkpoint: False
     epoch: -1 # -1 stands for last epoch
@@ -28,14 +28,14 @@ runner:
       - OmniObject3d_test
       - OmniObject3d_val
 
-  pipeline: nbv_reconstruction_global_pts_pipeline
+  pipeline: nbv_reconstruction_global_pts_n_num_pipeline
   
 dataset:
   OmniObject3d_train:
-    root_dir: "/home/data/hofee/project/nbv_rec/data/nbv_rec_data_512_preproc_npy"
+    root_dir: "/home/data/hofee/project/nbv_rec/data/sample_for_training_new"
     model_dir: "../data/scaled_object_meshes"
     source: nbv_reconstruction_dataset
-    split_file: "/home/data/hofee/project/nbv_rec/data/OmniObject3d_train.txt"
+    split_file: "/home/data/hofee/project/nbv_rec/data/sample.txt"
     type: train
     cache: True
     ratio: 1
@@ -45,10 +45,10 @@ dataset:
     load_from_preprocess: True
 
   OmniObject3d_test:
-    root_dir: "/home/data/hofee/project/nbv_rec/data/nbv_rec_data_512_preproc_npy"
+    root_dir: "/home/data/hofee/project/nbv_rec/data/sample_for_training_new"
     model_dir: "../data/scaled_object_meshes"
     source: nbv_reconstruction_dataset
-    split_file: "/home/data/hofee/project/nbv_rec/data/OmniObject3d_test.txt"
+    split_file: "/home/data/hofee/project/nbv_rec/data/sample.txt"
     type: test
     cache: True
     filter_degree: 75
@@ -61,10 +61,10 @@ dataset:
     load_from_preprocess: True
 
   OmniObject3d_val:
-    root_dir: "/home/data/hofee/project/nbv_rec/data/nbv_rec_data_512_preproc_npy"
+    root_dir: "/home/data/hofee/project/nbv_rec/data/sample_for_training_new"
     model_dir: "../data/scaled_object_meshes"
     source: nbv_reconstruction_dataset
-    split_file: "/home/data/hofee/project/nbv_rec/data/OmniObject3d_train.txt"
+    split_file: "/home/data/hofee/project/nbv_rec/data/sample.txt"
     type: test
     cache: True
     filter_degree: 75
@@ -96,6 +96,15 @@ pipeline:
     eps: 1e-5
     global_scanned_feat: True
 
+  nbv_reconstruction_global_pts_n_num_pipeline:
+    modules:
+      pts_encoder: pointnet_encoder
+      transformer_seq_encoder: transformer_seq_encoder
+      pose_encoder: pose_encoder
+      view_finder: gf_view_finder
+      pts_num_encoder: pts_num_encoder
+    eps: 1e-5
+    global_scanned_feat: True
 
     
 module:
@@ -107,7 +116,7 @@ module:
     feature_transform: False
 
   transformer_seq_encoder:
-    embed_dim: 1344
+    embed_dim: 384
     num_heads: 4
     ffn_dim: 256
     num_layers: 3
@@ -116,7 +125,7 @@ module:
   gf_view_finder:
     t_feat_dim: 128
     pose_feat_dim: 256
-    main_feat_dim: 2048
+    main_feat_dim: 3072
     regression_head: Rx_Ry_and_T
     pose_mode: rot_matrix
     per_point_feature: False
@@ -128,6 +137,9 @@ module:
     pose_dim: 9
     out_dim: 256
 
+  pts_num_encoder:
+    out_dim: 64
+
 loss_function:
   gf_loss:
 

core/global_pts_n_num_pipeline.py

@@ -117,22 +117,21 @@ class NBVReconstructionGlobalPointsPipeline(nn.Module):
             for seq_idx in range(seq_len): 
                 partial_idx_in_combined_pts = scanned_mask == seq_idx # Ndarray(V), N->V idx mask
                 partial_perpoint_feat = perpoint_scanned_feat[partial_idx_in_combined_pts] # Ndarray(V x Dl)
-                partial_feat = torch.mean(partial_perpoint_feat, dim=0)[0] # Tensor(Dl)
+                partial_feat = torch.mean(partial_perpoint_feat, dim=0) # Tensor(Dl)
                 partial_feat_seq.append(partial_feat)
                 scanned_target_pts_num.append(partial_perpoint_feat.shape[0])
-                
-            scanned_target_pts_num = torch.tensor(scanned_target_pts_num, dtype=torch.int32).to(device) # Tensor(S)
-            scanned_n_to_world_pose_9d = scanned_n_to_world_pose_9d.to(device)  # Tensor(S x 9)
             
+                
+            scanned_target_pts_num = torch.tensor(scanned_target_pts_num, dtype=torch.float32).unsqueeze(-1).to(device) # Tensor(S x 1)
+            scanned_n_to_world_pose_9d = scanned_n_to_world_pose_9d.to(device)  # Tensor(S x 9)
             pose_feat_seq = self.pose_encoder.encode_pose(scanned_n_to_world_pose_9d)  # Tensor(S x Dp)
             pts_num_feat_seq = self.pts_num_encoder.encode_pts_num(scanned_target_pts_num)  # Tensor(S x Dn)
             partial_feat_seq = torch.stack(partial_feat_seq) # Tensor(S x Dl)
-            
             seq_embedding = torch.cat([pose_feat_seq, pts_num_feat_seq, partial_feat_seq], dim=-1) # Tensor(S x (Dp+Dn+Dl))
             embedding_list_batch.append(seq_embedding) # List(B): Tensor(S x (Dp+Dn+Dl))
 
+        
         seq_feat = self.transformer_seq_encoder.encode_sequence(embedding_list_batch) # Tensor(B x Ds)
-
         main_feat = torch.cat([seq_feat, global_scanned_feat], dim=-1) # Tensor(B x (Ds+Dg))
 
         if torch.isnan(main_feat).any():

core/nbv_dataset.py

@@ -34,7 +34,7 @@ class NBVReconstructionDataset(BaseDataset):
             self.model_dir = config["model_dir"]
             self.filter_degree = config["filter_degree"]
         if self.type == namespace.Mode.TRAIN:
-            scale_ratio = 1
+            scale_ratio = 100
             self.datalist = self.datalist*scale_ratio
         if self.cache:
             expr_root = ConfigManager.get("runner", "experiment", "root_dir")
@@ -122,7 +122,7 @@ class NBVReconstructionDataset(BaseDataset):
             scanned_views_pts,
             scanned_coverages_rate,
             scanned_n_to_world_pose,
-        ) = ([], [], [], [])
+        ) = ([], [], [])
         for view in scanned_views:
             frame_idx = view[0]
             coverage_rate = view[1]
@@ -164,7 +164,7 @@ class NBVReconstructionDataset(BaseDataset):
             combined_scanned_views_pts, self.pts_num, require_idx=True
         )
         
-        combined_scanned_views_pts_mask = np.zeros(len(scanned_views_pts), dtype=np.uint8)
+        combined_scanned_views_pts_mask = np.zeros(len(combined_scanned_views_pts), dtype=np.uint8)
         
         start_idx = 0
         for i in range(len(scanned_views_pts)):
@@ -174,9 +174,6 @@ class NBVReconstructionDataset(BaseDataset):
             
         fps_downsampled_combined_scanned_pts_mask = combined_scanned_views_pts_mask[fps_idx]
         
-        
-        
-        
         data_item = {
             "scanned_pts": np.asarray(scanned_views_pts, dtype=np.float32), # Ndarray(S x Nv x 3)
             "scanned_pts_mask": np.asarray(fps_downsampled_combined_scanned_pts_mask,dtype=np.uint8), # Ndarray(N), range(0, S)