Update custom model files, README, and requirements

asr_modeling.py

@@ -708,16 +708,22 @@ class ASRModel(PreTrainedModel):
         print(f"DEBUG generate (non-streaming): task={task}, system_prompt={system_prompt}, user_prompt={user_prompt}", file=sys.stderr)
         print(f"DEBUG generate (non-streaming): generate_kwargs={generate_kwargs}", file=sys.stderr)
 
-        # Add streamer if provided
+        # Generate with or without streamer
         if streamer is not None:
-            generate_kwargs["streamer"] = streamer
-
-        generated_ids = self.decoder.generate(
-            input_ids=expanded_prompt_ids,
-            inputs_embeds=inputs_embeds,
-            attention_mask=attention_mask,
-            **generate_kwargs,
-        )
+            generated_ids = self.decoder.generate(
+                input_ids=expanded_prompt_ids,
+                inputs_embeds=inputs_embeds,
+                attention_mask=attention_mask,
+                streamer=streamer,
+                **generate_kwargs,
+            )
+        else:
+            generated_ids = self.decoder.generate(
+                input_ids=expanded_prompt_ids,
+                inputs_embeds=inputs_embeds,
+                attention_mask=attention_mask,
+                **generate_kwargs,
+            )
 
         return generated_ids[:, prompt_length:]