fix: cleanup typos

Author: drbh

server/text_generation_server/utils/logits_process.py

@@ -627,7 +627,7 @@ def filter(self, indices):
 
 class LogitBiasProcessor(LogitsProcessor):
     """
-    `LogitsProcessor` creates a bias tensor from a dictionary of token IDs and their
+    `LogitBiasProcessor` creates a bias tensor from a dictionary of token IDs and their
     corresponding bias values. Bias are applied to the logits during each forward pass.
 
     Supports token IDs provided as strings (e.g., {"9707": -100}).
@@ -656,7 +656,7 @@ def __init__(
     def __call__(self, input_ids: torch.Tensor, scores: torch.Tensor) -> torch.Tensor:
         # Apply bias tensor as a broadcasted addition
         if self.bias_tensor.shape[0] != scores.shape[1]:
-            # Fix if the bias tensor is smaller than the scores
+            # Pad the bias matrix to match the scores if it's smaller
             self.bias_tensor = torch.nn.functional.pad(
                 self.bias_tensor, (0, scores.shape[1] - self.bias_tensor.shape[0])
             )
@@ -699,7 +699,7 @@ def __init__(
     def __call__(self, input_ids: torch.Tensor, scores: torch.Tensor) -> torch.Tensor:
         # Apply bias matrix as a broadcasted addition
         if self.bias_matrix.shape[1] != scores.shape[1]:
-            # Fix if the bias matrix is smaller than the scores
+            # Pad the bias matrix to match the scores if it's smaller
             self.bias_matrix = torch.nn.functional.pad(
                 self.bias_matrix, (0, scores.shape[1] - self.bias_matrix.shape[1])
             )