Update app.py

app.py

@@ -27,6 +27,7 @@ pipe = pipeline(
     task="automatic-speech-recognition",
     model=MODEL_NAME,
     chunk_length_s=CHUNK_LENGTH_S,
+    batch_size=BATCH_SIZE,
     torch_dtype=torch_dtype,
     device=device,
     model_kwargs=model_kwargs
@@ -40,7 +41,7 @@ def transcribe(inputs, prompt):
     generate_kwargs = {"language": "japanese", "task": "transcribe"}
     prompt = "。" if not prompt else prompt
     generate_kwargs['prompt_ids'] = pipe.tokenizer.get_prompt_ids(prompt, return_tensors='pt').to(device)
-    result = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs=generate_kwargs)["text"]
+    result = pipe(inputs, generate_kwargs=generate_kwargs)["text"]
     return result['text'][1 + len(prompt) + 1:] 
 
 def _return_yt_html_embed(yt_url):
@@ -85,7 +86,7 @@ def yt_transcribe(yt_url, prompt, max_filesize=75.0):
     generate_kwargs = {"language": "japanese", "task": "transcribe"}
     prompt = "。" if not prompt else prompt
     generate_kwargs['prompt_ids'] = pipe.tokenizer.get_prompt_ids(prompt, return_tensors='pt').to(device)
-    result = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs=generate_kwargs)["text"]
+    result = pipe(inputs, generate_kwargs=generate_kwargs)["text"]
     return html_embed_str, result['text'][1 + len(prompt) + 1:]