Add application code

ModelName

@@ -0,0 +1 @@
+model: facebook/wav2vec2-base-960h

app.py

@@ -0,0 +1,34 @@
+import warnings
+import torch
+import gradio as gr
+import librosa
+from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
+
+warnings.filterwarnings("ignore")
+processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-base-960h")
+model     = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-960h")
+
+def transcribe_audio(audio_path):
+    try:
+        audio,sampling_rate = librosa.load(audio_path, sampling_rate=16000)
+        input_values = processor(audio, return_tensors='pt', sampling_rate=16000).input_values
+
+        with torch.no_grad():
+            logits = model(input_values).logits
+
+        predicted_ids = torch.argmax(logits, dim=-1)
+        transcriptions = processor.batch_decode(predicted_ids)[0]
+        return transcriptions
+    except Exception as e:
+        return str(e)
+
+demo = gr.Interface(
+    fn = transcribe_audio,
+    inputs= gr.Audio(type='filepath'),
+    outputs='text',
+    title="Subtitle Genarator",
+    description="This tool transcribes audio file into text"
+)
+demo.launch()
+
+

requirments.txt

@@ -0,0 +1,5 @@
+torch 
+librosa //for audio
+numpy
+transformers
+gradio //making interface for a AI app with this package