Updated to use larger model

app.py

@@ -2,7 +2,7 @@ import logging
 
 import gradio as gr
 
-# import torch
+import torch
 from transformers import (
     AutoModelForSpeechSeq2Seq,
     AutoProcessor,
@@ -13,7 +13,7 @@ from transformers import (
 
 device = "cpu"
 # device = "cuda:0" if torch.cuda.is_available() else "cpu"
-# torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
+torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
 
 model_id = "openai/whisper-large-v3"
 
@@ -26,20 +26,20 @@ model = AutoModelForSpeechSeq2Seq.from_pretrained(
 )
 model.to(device)
 
-processor = WhisperProcessor.from_pretrained("openai/whisper-base.en")
+# processor = WhisperProcessor.from_pretrained("openai/whisper-base.en")
 
-# processor = AutoProcessor.from_pretrained(model_id)
+processor = AutoProcessor.from_pretrained(model_id)
 
 pipe = pipeline(
     task="automatic-speech-recognition",
     model=model,
     tokenizer=processor.tokenizer,
     feature_extractor=processor.feature_extractor,
-    # max_new_tokens=128,
+    max_new_tokens=128,
     chunk_length_s=30,
     batch_size=8,
-    # return_timestamps=True,
-    # torch_dtype=torch_dtype,
+    return_timestamps=True,
+    torch_dtype=torch_dtype,
     device=device,
 )