Updated to use local model

app.py

@@ -4,7 +4,7 @@ import gradio as gr
 import os
 
 #%%
-whisper = pipeline(model='jlvdoorn/whisper-large-v2-atco2-asr-atcosim')
+whisper = pipeline(model='C:\\Users\\doornj\\Desktop\\WhisperANSP\\Models\\whisper-large-v2-atco2-asr-atcosim-ANSP-3h1m', task='automatic-speech-recognition')
 bert_atco_ner = pipeline(model='Jzuluaga/bert-base-ner-atc-en-atco2-1h')
 
 #%%
@@ -49,8 +49,8 @@ iface = gr.Interface(
         fn=transcribeAndExtract,
         inputs=[gr.Audio(source='upload', type='filepath'), gr.Audio(source='microphone', type='filepath'), gr.Checkbox(label='Transcribe only', default=False)],
         outputs=[gr.Text(label='Transcription'), gr.Text(label='Callsigns, commands and values')],
-        title='Whisper Large v2 - ATCO2-ASR-ATCOSIM',
-        description='This demo will transcribe ATC audio files by using the Whisper Large v2 model fine-tuned on the ATCO2 and ATCOSIM datasets. Further it uses a Named Entity Recognition model to extract callsigns, commands and values from the transcription. This model is based on Google\'s BERT model and fine-tuned on the ATCO2 dataset.',
+        title='Whisper Large v2 - ATCO2-ATCOSIM-ANSP',
+        description='This demo will transcribe ATC audio files by using the Whisper Large v2 model fine-tuned on the ATCO2, ATCOSIM and ANSP datasets. Further it uses a Named Entity Recognition model to extract callsigns, commands and values from the transcription. This model is based on Google\'s BERT model and fine-tuned on the ATCO2 dataset.',
 )
 
 #%%