app.py

import gradio as gr
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
import torch
import librosa

# Load your model
processor = Wav2Vec2Processor.from_pretrained("jayarizco101/fuller-finetuned-wav2vec2")
model = Wav2Vec2ForCTC.from_pretrained("jayarizco101/fuller-finetuned-wav2vec2")
model.eval()

def transcribe(audio):
    # audio is automatically 16kHz float32 from Gradio
    inputs = processor(audio, sampling_rate=16000, return_tensors="pt", padding=True)
    with torch.no_grad():
        logits = model(**inputs).logits
    pred_ids = torch.argmax(logits, dim=-1)
    transcription = processor.batch_decode(pred_ids)[0]
    return transcription

iface = gr.Interface(
    fn=transcribe,
    inputs=gr.Audio(sources=["upload"], type="numpy"),
    outputs="text"
)

iface.launch()