Update app.py

app.py

@@ -1,25 +1,45 @@
-import torch
-import torchaudio
+from transformers import pipeline
 import gradio as gr
-from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
+import time
+import unicodedata
+p = pipeline("automatic-speech-recognition",model="kingabzpro/wav2vec2-large-xls-r-300m-Urdu")
 
-# Load the model and processor
-model_name = "kingabzpro/wav2vec2-large-xlsr-300m-urdu"
-model = Wav2Vec2ForCTC.from_pretrained(model_name)
-processor = Wav2Vec2Processor.from_pretrained(model_name)
+def transcribe(audio, state=""):
+    time.sleep(2)
+    text = p(audio)["text"]
+    state += unicodedata.normalize("NFC",text) + " "
+    return state, state
 
-# Define the transcribe function
-def transcribe(audio):
-    waveform, sample_rate = torchaudio.load(audio.name)
-    input_dict = processor(waveform, return_tensors="pt", padding=True)
-    logits = model(input_dict.input_values).logits
-    predicted_ids = torch.argmax(logits, dim=-1).squeeze()
-    transcription = processor.decode(predicted_ids)
-    return transcription
+################### Gradio Web APP ################################
 
-# Define the interface
-audio_input = gr.inputs.Audio(source="upload", type="file", label="Upload audio file")
-text_output = gr.outputs.Textbox(label="Transcription")
+title = "Real-Time Urdu ASR"
 
-interface = gr.Interface(fn=transcribe, inputs=audio_input, outputs=text_output, title="Urdu Speech Recognition")
-interface.launch(share=True)
+description = """
+<p>
+<center>
+This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice dataset.
+</center>
+</p>
+<center>
+<img src="https://huggingface.co/spaces/kingabzpro/real-time-Urdu-ASR/resolve/main/Images/cover.jpg" alt="logo" width="550"/>
+</center>
+"""
+
+article = "<p style='text-align: center'><a href='https://dagshub.com/kingabzpro/Urdu-ASR-SOTA' target='_blank'>Source Code on DagsHub</a></p><p style='text-align: center'><a href='https://huggingface.co/blog/fine-tune-xlsr-wav2vec2' target='_blank'>Fine-tuning XLS-R for Multi-Lingual ASR with 🤗 Transformers</a></p></center><center><img src='https://visitor-badge.glitch.me/badge?page_id=kingabzpro/real-time-Urdu-ASR' alt='visitor badge'></center></p>"
+
+
+gr.Interface(
+    fn=transcribe, 
+    inputs=[
+        gr.Audio(source="microphone", type="filepath", streaming=True), 
+        "state"
+    ],
+    outputs=[
+        "textbox",
+        "state"
+    ],
+    title=title,
+    description=description,
+    article=article,
+    theme='EveryPizza/Cartoony-Gradio-Theme',
+    live=True).launch()