app.py

This is the test, lets see if this would be perfect.

app.py

@@ -0,0 +1,30 @@
+import gradio as gr
+from transformers import AutoProcessor, AutoModelForCTC
+import torch
+import soundfile as sf
+
+# Load the FastConformer model and processor
+processor = AutoProcessor.from_pretrained("nvidia/stt_en_fastconformer_hybrid_large_pc")
+model = AutoModelForCTC.from_pretrained("nvidia/stt_en_fastconformer_hybrid_large_pc")
+
+# Function to transcribe audio
+def transcribe_audio(audio_file):
+    audio_input, sample_rate = sf.read(audio_file)
+    inputs = processor(audio_input, sampling_rate=sample_rate, return_tensors="pt")
+    with torch.no_grad():
+        logits = model(**inputs).logits
+    predicted_ids = torch.argmax(logits, dim=-1)
+    transcription = processor.batch_decode(predicted_ids)[0]
+    return transcription
+
+# Create a Gradio interface
+iface = gr.Interface(
+    fn=transcribe_audio,
+    inputs=gr.Audio(type="filepath"),
+    outputs="text",
+    title="Real-Time Transcription with FastConformer",
+    description="Upload an audio file to transcribe it using NVIDIA FastConformer."
+)
+
+# Launch the app
+iface.launch()