Update custom model files, README, and requirements

asr_modeling.py

@@ -781,6 +781,14 @@ class ASRModel(PreTrainedModel):
                 if chunk:
                     output_token_count += 1
                     yield StreamChunk(chunk)
+        except Exception as e:
+            # Check if it's the Empty exception from queue
+            if e.__class__.__name__ == "Empty":
+                # This is expected when streaming completes quickly
+                pass
+            else:
+                # Re-raise other exceptions
+                raise
         finally:
             # Wait for generation to complete
             thread.join()