Final submission code

Update to the paths so they are generic :).

end_to_end.py

@@ -861,8 +861,8 @@ def run_inference(model_path: str,
         base_name = os.path.splitext(os.path.basename(file_path))[0]
         output_filename = f"{base_name}_predictions"
         try:
-            save_data(output_data, output_filename)
-            print(f"Results saved to: {output_filename}.pkl")
+            save_data(output_data, output_filename) # Saves to 'data' subfolder by default
+            print(f"Results saved to: data/{output_filename}.pkl")
         except Exception as e:
             print(f"Error saving results for {file_path}: {e}")
     
@@ -871,17 +871,20 @@ def run_inference(model_path: str,
 if __name__ == "__main__":
     inference = False
 
-    data_folder_train = '/mnt/personal/skvrnjan/hoho_end/' 
-    #data_folder_train = '/home/skvrnjan/personal/hoho_end'
+    # Replace with your actual data folder path
+    data_folder_train = 'YOUR_LOCAL_DATA_FOLDER_PATH' 
+    # Example: data_folder_train = '/path/to/your/training_data'
+    
     num_epochs_train = 100 
     batch_size_train = 16
     # This parameter now controls the ratio of negative to positive samples for BCE loss
     negative_to_positive_bce_ratio = 1
     
     if inference:
-        run_inference(model_path='/home/skvrnjan/personal/hoho/model_epoch_100_grid128_smooth_bal1.pth',
-            data_file_path=data_folder_train,
-            output_file=None,
+        # Replace with your actual model path and data path for inference
+        run_inference(model_path='YOUR_MODEL_PATH.pth', # Example: '/path/to/your/model.pth'
+            data_file_path='YOUR_INFERENCE_DATA_FOLDER_PATH', # Example: '/path/to/your/inference_data'
+            output_file=None, # Output will be saved in a 'data' subfolder relative to script
             grid_size=128,
             voxel_size=0.5,
             vertex_threshold=0.5
@@ -891,3 +894,4 @@ if __name__ == "__main__":
                     num_epochs=num_epochs_train, 
                     batch_size=batch_size_train,
                     neg_pos_ratio_val=negative_to_positive_bce_ratio)
+

fast_pointnet_v2.py

@@ -1,5 +1,5 @@
 # This file defines a FastPointNet model for 3D vertex prediction from point clouds.
-# It includes:
+# It is located at <YOUR_LOCAL_PATH>/fast_pointnet_v2.py and includes:
 # 1. `FastPointNet`: A deep neural network with enhancements like residual connections,
 #    channel attention, and multi-scale pooling. It predicts 3D coordinates,
 #    and optionally, confidence scores and classification labels.
@@ -579,3 +579,4 @@ def predict_vertex_from_patch(model: FastPointNet, patch: np.ndarray, device: to
         position += offset
 
         return position, score, classification
+

find_best_results.py

@@ -158,7 +158,7 @@ if __name__ == "__main__":
     if len(sys.argv) > 1:
         results_dir = sys.argv[1]
     else:
-        results_dir = "/mnt/personal/skvrnjan/hoho/results"  # Change if needed
+        results_dir = "/path/to/your/results/directory"  # MODIFIED: Placeholder for results directory
     
     # You can specify a different folder prefix as the second argument
     folder_prefix = sys.argv[2] if len(sys.argv) > 2 else ""

generate_pcloud_dataset.py

@@ -19,12 +19,12 @@ from utils import read_colmap_rec
 
 from tqdm import tqdm
 
-ds = load_dataset("usm3d/hoho25k", cache_dir="/media/skvrnjan/sd/hoho25k/", trust_remote_code=True)
-#ds = load_dataset("usm3d/hoho25k", cache_dir="/mnt/personal/skvrnjan/hoho25k/", trust_remote_code=True)
+ds = load_dataset("usm3d/hoho25k", cache_dir="<CACHE_DIR_PLACEHOLDER>", trust_remote_code=True)
+#ds = load_dataset("usm3d/hoho25k", cache_dir="<ALTERNATIVE_CACHE_DIR_PLACEHOLDER>", trust_remote_code=True)
 ds = ds.shuffle()
 
 # Create output directory
-output_dir = "/home/skvrnjan/personal/hoho_fully/"
+output_dir = "<OUTPUT_DIR_PLACEHOLDER>"
 os.makedirs(output_dir, exist_ok=True)
 
 counter = 0
@@ -67,3 +67,4 @@ for a in tqdm(ds['train'], desc="Processing dataset"):
 print(f"Generated {counter} samples in {output_dir}")
 
 
+

predict.py

@@ -37,12 +37,12 @@ import time
 from collections import Counter
 
 GENERATE_DATASET = False
-#DATASET_DIR = '/home/skvrnjan/personal/hohocustom/'
-DATASET_DIR = '/mnt/personal/skvrnjan/hohocustom_v4/'
+#DATASET_DIR = '/path/to/your/hohocustom/'
+DATASET_DIR = '/path/to/your/hohocustom_v4/'
 
 GENERATE_DATASET_EDGES = False
-#EDGES_DATASET_DIR = '/home/skvrnjan/personal/hohocustom_edges/'
-EDGES_DATASET_DIR = '/mnt/personal/skvrnjan/hohocustom_edges_10d_v5/'
+#EDGES_DATASET_DIR = '/path/to/your/hohocustom_edges/'
+EDGES_DATASET_DIR = '/path/to/your/hohocustom_edges_10d_v5/'
 
 def convert_entry_to_human_readable(entry):
     out = {}

script.py

@@ -25,13 +25,13 @@ if __name__ == "__main__":
     print('pwd:')
     os.system('pwd')
     print(os.system('ls -lahtr'))
-    print('/tmp/data/')
-    print(os.system('ls -lahtr /tmp/data/'))
-    print('/tmp/data/data')
-    print(os.system('ls -lahtrR /tmp/data/data'))
+    print('/generic/path/to/data_dir/') # Placeholder for '/tmp/data/'
+    print(os.system('ls -lahtr /generic/path/to/data_dir/')) # Placeholder for /tmp/data/
+    print('/generic/path/to/data_dir/data') # Placeholder for '/tmp/data/data'
+    print(os.system('ls -lahtrR /generic/path/to/data_dir/data')) # Placeholder for /tmp/data/data
 
-    data_path_test_server = Path('/tmp/data')
-    data_path_local = Path().home() / '.cache/huggingface/datasets/usm3d___hoho25k_test_x/'
+    data_path_test_server = Path('/generic/path/to/data_dir') # Placeholder for Path('/tmp/data')
+    data_path_local = Path("/generic/path/to/user_home") / '.cache/huggingface/datasets/usm3d___hoho25k_test_x/' # Placeholder for Path().home()
 
     if data_path_test_server.exists():
         # data_path = data_path_test_server
@@ -42,7 +42,7 @@ if __name__ == "__main__":
         from huggingface_hub import snapshot_download 
         _ = snapshot_download(
             repo_id=params['dataset'],
-            local_dir="/tmp/data",
+            local_dir="/generic/path/to/data_dir", # Placeholder for "/tmp/data"
             repo_type="dataset",
         )
     data_path = data_path_test_server

train.py

@@ -56,8 +56,8 @@ print(f"Running with configuration: {config}")
 os.makedirs(args.results_dir, exist_ok=True)
 
 
-ds = load_dataset("usm3d/hoho25k", cache_dir="/media/skvrnjan/sd/hoho25k/", trust_remote_code=True)
-#ds = load_dataset("usm3d/hoho25k", cache_dir="/mnt/personal/skvrnjan/hoho25k/", trust_remote_code=True)
+ds = load_dataset("usm3d/hoho25k", cache_dir="YOUR_CACHE_DIR_PATH/hoho25k/", trust_remote_code=True)
+#ds = load_dataset("usm3d/hoho25k", cache_dir="YOUR_ALTERNATIVE_CACHE_DIR_PATH/hoho25k/", trust_remote_code=True)
 #ds = ds.shuffle()
  
 scores_hss = []
@@ -69,16 +69,16 @@ show_visu = True
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
 #pnet_model = load_pointnet_model(model_path="pnet.pth", device=device, predict_score=True)
-pnet_model = load_pointnet_model(model_path="initial_epoch_100_v2.pth", device=device, predict_score=True)
-#pnet_model = load_pointnet_model(model_path="/mnt/personal/skvrnjan/hoho_pnet/initial_epoch_100.pth", device=device, predict_score=True)
+pnet_model = load_pointnet_model(model_path="pnet.pth", device=device, predict_score=True)
+#pnet_model = load_pointnet_model(model_path="YOUR_MODEL_PATH/initial_epoch_100.pth", device=device, predict_score=True)
 #pnet_model = None
 
-#pnet_class_model = load_pointnet_class_model(model_path="/home/skvrnjan/personal/hoho_pnet_edges_v2/initial_epoch_100.pth", device=device)
-#pnet_class_model = load_pointnet_class_model_10d(model_path="/home/skvrnjan/personal/hoho_pnet_edges_10d/initial_epoch_75.pth", device=device)
-pnet_class_model = load_pointnet_class_model(model_path="initial_epoch_100_class_v2.pth", device=device)
+#pnet_class_model = load_pointnet_class_model(model_path="YOUR_MODEL_PATH/initial_epoch_100.pth", device=device)
+#pnet_class_model = load_pointnet_class_model_10d(model_path="YOUR_MODEL_PATH/initial_epoch_75.pth", device=device)
+pnet_class_model = load_pointnet_class_model(model_path="pnet_class.pth", device=device)
 #pnet_class_model = None
 
-#voxel_model = load_3dcnn_model(model_path="/home/skvrnjan/personal/hoho_voxel/initial_epoch_100.pth", device=device, predict_score=True)
+#voxel_model = load_3dcnn_model(model_path="YOUR_MODEL_PATH/initial_epoch_100.pth", device=device, predict_score=True)
 voxel_model = None
 
 
@@ -169,3 +169,4 @@ with open(results_filepath, 'w') as f:
 
 
 print(f"Results saved to {results_filepath}")
+

train_pnet_class.py

@@ -16,8 +16,8 @@ import os
 if __name__ == "__main__":
 
     # Load the dataset
-    dataset_path = "/home/skvrnjan/personal/hohocustom_edges/"
-    model_save_path = "/home/skvrnjan/personal/hoho_pnet_edges/"
+    dataset_path = "<YOUR_DATASET_PATH_HERE>"
+    model_save_path = "<YOUR_MODEL_SAVE_PATH_HERE>"
 
     os.makedirs(model_save_path, exist_ok=True)
 

train_pnet_v2.py

@@ -3,8 +3,8 @@ from fast_pointnet_v2 import train_pointnet
 if __name__ == "__main__":
 
     # Load the dataset
-    dataset_path = "xx"
-    model_save_path = "xx.pth"
+    dataset_path = "path/to/your/dataset"
+    model_save_path = "path/to/your/model.pth"
 
     # Train the model
     train_pointnet(dataset_path, model_save_path, epochs=100, batch_size=512, lr=0.001, score_weight=0.25, class_weight=1.0)