Update README.md

README.md

@@ -136,8 +136,8 @@ def collate(batch: list[dict[str, str | Image]]) -> tuple[list[str], list[str],
 def process_imgs(imgs: torch.Tensor, tile_ids: list[str]) -> torch.Tensor:
     """Perform inference on input (already transformed) images."""
     with torch.inference_mode():
-        batch_features = model(imgs.to(device)).squeeze()
-        batch_tiles_coordinates = np.array([tile_id.split("_")[1:] for tile_id in tile_ids]).astype(int)
+        batch_features = model(imgs.to(device)).squeeze() # (N_tiles, d) numpy array
+        batch_tiles_coordinates = np.array([tile_id.split("_")[1:] for tile_id in tile_ids]).astype(int) # (N_tiles, 3) numpy array
     batch_stack = np.concatenate([batch_tiles_coordinates, batch_features], axis=1)
     return batch_stack
 
@@ -159,38 +159,41 @@ dataloader = DataLoader(
 )
 
 # Iterate over the full dataset and store features each time 16278 input images have been processed
+   
 slide_features = []
+current_num_tiles = 0
 
 for i, (slide_ids, tile_ids, imgs) in tqdm(
     enumerate(dataloader),
     total=ceil(num_slides * num_tiles / batch_size),
     desc="Extracting features"
 ):
-    # Detect the current slide
-    current_slide_id = slide_ids[0]
+    reference_slide_id = slide_ids[0]
 
-    # If all tiles from the current batch belongs to the same slide,
-    # we just add the batch features to the running list
-    if all(slide_id == current_slide_id for slide_id in slide_ids):
+    # If we're on the same slide, we just add the batch features to the running list
+    if all(slide_id == reference_slide_id for slide_id in slide_ids):
         batch_stack = process_imgs(imgs, tile_ids)
         slide_features.append(batch_stack)
-        # If `slide_features` contains 16,278 tiles features, we export it
-        if len(slide_features) == ceil(num_tiles / batch_size):
-            save_features(slide_features, slide_id=current_slide_id)
-            # And initialize `slide_features` as an empty list
+        current_num_tiles += batch_size
+        # If the current batch contains exactly the last `batch_size` tile features for the slide,
+        # export the slide features and reset `slide_features` and `current_num_tiles`
+        if current_num_tiles == num_tiles:
+            save_features(slide_features, slide_id=reference_slide_id)
             slide_features = []
-            continue
-    # The current batch contains tiles from current (N) and N+1 slide
+            current_num_tiles = 0
+    # The current batch contains tiles from slide N (`reference_slide_id`) and slide N+1
     else:
-        # We get the maximum index at which tiles all belong to slide N
-        mask = (np.array(slide_ids) != current_slide_id)
+        # We retrieve the maximum index at which all tiles in the batch comes from slide N
+        mask = (np.array(slide_ids) != reference_slide_id)
         idx = mask.argmax()
-        # Then export the slide features with the remaining idx features
-        batch_stack = process_imgs(imgs[:idx], tile_ids[:idx])
+        # And only process the later, then export the slides features
+        batch_stack = process_imgs(imgs[:idx], tile_ids[:idx]
         slide_features.append(batch_stack)
-        save_features(slide_features, slide_id=current_slide_id)
-        # And initialize with (batch_size - idx) tiles features from slide N+1
+        save_features(slide_features, slide_id=reference_slide_id)
+        # We initialize `slide_features` and `current_num_tiles` with respectively
+        # the tile features from slide N+1
         slide_features = [process_imgs(imgs[idx:], tile_ids[idx:])]
+        current_num_tiles = batch_size - idx
 ```
 
 # License