Update app.py

app.py

@@ -6,6 +6,7 @@ import time
 from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
 from happytransformer import HappyTextToText, TTSettings  # Using HappyTransformer
 
+# Load models only once for efficiency
 def load_models():
     model_name = "prithivida/grammar_error_correcter_v1"
     tokenizer = AutoTokenizer.from_pretrained(model_name)
@@ -13,6 +14,9 @@ def load_models():
     happy_tt = HappyTextToText("T5", "prithivida/grammar_error_correcter_v1")  # Using T5-based model
     return tokenizer, model, happy_tt
 
+tokenizer, model, happy_tt = load_models()  # Load models at startup
+
+# Speech-to-text conversion
 def transcribe_audio(audio):
     recognizer = sr.Recognizer()
     with sr.AudioFile(audio) as source:
@@ -25,7 +29,11 @@ def transcribe_audio(audio):
     except sr.RequestError as e:
         return f"Speech recognition error: {e}"
 
-def correct_grammar(text, tokenizer, model, happy_tt):
+# Grammar correction function
+def correct_grammar(text):
+    if not text.strip():
+        return "No input provided.", 0, "No correction available."
+
     inputs = tokenizer.encode("gec: " + text, return_tensors="pt", max_length=128, truncation=True)
     with torch.no_grad():
         outputs = model.generate(inputs, max_length=128, num_return_sequences=1)
@@ -37,33 +45,27 @@ def correct_grammar(text, tokenizer, model, happy_tt):
     
     return corrected_text, grammar_score, correction
 
-def gradio_interface(text):
-    tokenizer, model, happy_tt = load_models()
-    corrected_text, grammar_score, correction = correct_grammar(text, tokenizer, model, happy_tt)
-    return corrected_text, grammar_score, correction
-
-def gradio_audio_interface(audio):
-    text = transcribe_audio(audio)
-    return gradio_interface(text)
+# Unified function for both speech and text input
+def process_input(audio, text):
+    if audio:  
+        text = transcribe_audio(audio)  # If audio is provided, transcribe it
+    return correct_grammar(text)
 
+# Gradio UI
 def main():
     iface = gr.Interface(
-        fn=gradio_interface,
-        inputs=gr.Textbox(placeholder="Enter a sentence..."),
+        fn=process_input,
+        inputs=[
+            gr.Audio(sources=["microphone"], type="filepath", label="Speak your sentence"),
+            gr.Textbox(placeholder="Or type here if not speaking...", label="Text Input"),
+        ],
         outputs=["text", "number", "text"],
         title="AI Grammar Checker",
-        description="Enter text to check grammar, get suggestions, and see a score."
+        description="Speak or type a sentence to check its grammar, get corrections, and see a score.",
+        live=False,  # Only processes when user submits
     )
-    
-    audio_iface = gr.Interface(
-        fn=gradio_audio_interface,
-        inputs=gr.Audio(sources=["microphone"], type="filepath"),  # Fixed here
-        outputs=["text", "text", "number", "text"],
-        title="AI Grammar Checker (Audio)",
-        description="Speak to check grammar, get suggestions, and see a score."
-    )
-    
-    gr.TabbedInterface([iface, audio_iface], ["Text Input", "Speech Input"]).launch()
+
+    iface.launch()
 
 if __name__ == "__main__":
     main()