Update app.py

updated transcription model for timestamp feature

app.py

@@ -1,7 +1,10 @@
 import torch
-from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
+# from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
+from transformers import pipeline
 import gradio as gr
+import datetime
 
+"""
 device = "cuda:0" if torch.cuda.is_available() else "cpu"
 torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
 
@@ -23,9 +26,35 @@ pipe = pipeline(
     torch_dtype=torch_dtype,
     device=device,
 )
+"""
+# call a text generation model to display the audio content after identifying the word(s) in the text output
 
-def audio2text(audio_file):
-  output=pipe(audio_file)
-  return output['text']
+#import torch
+#from transformers import pipeline
+#from datasets import load_dataset
 
-gr.Interface(fn=audio2text, inputs=[gr.Audio(label='upload your audio file', sources='upload', type='filepath')], outputs=[gr.Textbox(label="transcription")]).launch()
+device = "cuda:0" if torch.cuda.is_available() else "cpu"
+
+pipe = pipeline(
+  "automatic-speech-recognition",
+  # model="openai/whisper-base",
+    model = "microsoft/whisper-base-webnn",
+    chunk_length_s=30,
+    device=device,
+)
+
+# ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
+# sample = ds[0]["audio"]
+
+# prediction = pipe(sample.copy(), batch_size=8)["text"]
+
+# we can also return timestamps for the predictions
+prediction = pipe(sample.copy(), batch_size=8, return_timestamps=True)["chunks"]
+
+
+def audio2text(audio_file, prompt : str | list):
+    prediction = pipe(audio_file, batch_size=8, return_timestamps=True)["chunks"]
+    #prediction=pipe(audio_file)
+    return prediction['text']
+
+gr.Interface(fn=audio2text, inputs=[gr.Audio(label='upload your audio file', sources='upload', type='filepath'), gr.Textbox(label="provide word(s) to search for")], outputs=[gr.Textbox(label="transcription")]).launch()