Update app.py

app.py

@@ -18,19 +18,25 @@ from transformers import AutoProcessor, pipeline
 
 
 # Initialize processor and pipeline
-processor = AutoProcessor.from_pretrained("NbAiLabBeta/nb-whisper-large")
-device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-torch_dtype = torch.float32
 
-# Set distinct pad and eos tokens
-if processor.tokenizer.pad_token_id is None:
-    processor.tokenizer.pad_token_id = processor.tokenizer.convert_tokens_to_ids("[PAD]")
-if processor.tokenizer.eos_token_id is None:
-    processor.tokenizer.eos_token_id = processor.tokenizer.convert_tokens_to_ids("[EOS]")
+# torch_dtype = torch.float32
 
-pipe = pipeline("automatic-speech-recognition", model="NbAiLabBeta/nb-whisper-large", torch_dtype=torch_dtype)
 
-#language = "no"
+MODEL_NAME = "NbAiLabBeta/nb-whisper-large"
+
+language = "no"
+
+device = "cuda" if torch.cuda.is_available() else "cpu"
+pipe = pipeline(
+    task="automatic-speech-recognition",
+    model=MODEL_NAME,
+    chunk_length_s=30,
+    device=device,
+)
+
+
+pipe.model.config.forced_decoder_ids = pipe.tokenizer.get_decoder_prompt_ids(language=lang, task="transcribe")
+
 # task = "transcribe"
 
 def transcribe_audio(audio_file):
@@ -185,7 +191,8 @@ def text_rank_summary(text, num_paragraphs=3):
 
 banner_html = """
 <div style="text-align: center;">
-    <img src="https://raw.huggingface.co/spaces/camparchimedes/transcription_app/blob/main/banner_trans.png" alt="" width="100%" height="auto">
+    <img src="https://raw.huggingface.co/spaces/camparchimedes/transcription_app/blob/main/banner_trans.png
+" alt="" width="100%" height="auto">
 </div>
 """
 # https://raw.huggingface.co/spaces/camparchimedes/transcription_app/blob/main/banner_trans.png