Update app.py

app.py

@@ -1,25 +1,10 @@
 import os
 import gradio as gr
 from gtts import gTTS
-from transformers import pipeline, AutoProcessor, WhisperForConditionalGeneration
 from openai import OpenAI
 
 client = OpenAI()
 
-# Initialize the processor and model separately for better control
-processor = AutoProcessor.from_pretrained("seeafricatz/kiaziboraasr")
-model = WhisperForConditionalGeneration.from_pretrained("seeafricatz/kiaziboraasr")
-
-pipe = pipeline(
-    "automatic-speech-recognition",
-    model=model,
-    tokenizer=processor.tokenizer,
-    feature_extractor=processor.feature_extractor,
-    chunk_length_s=30,
-    return_timestamps=False,
-    generate_kwargs={"language": "<|swahili|>", "task": "transcribe"}
-)
-
 def transcribe(audio):
     try:
         if audio is None:
@@ -30,23 +15,13 @@ def transcribe(audio):
         if not os.path.exists(audio_path):
             return "Audio file not found"
             
-        result = pipe(
-            audio_path,
-            return_timestamps=False,
-            generate_kwargs={
-                "language": "<|swahili|>",
-                "task": "transcribe",
-                "num_beams": 5,
-                "temperature": 0
-            }
-        )
-        
-        if isinstance(result, dict) and "text" in result:
-            return result["text"]
-        elif isinstance(result, str):
-            return result
-        else:
-            return "Error in transcription format"
+        with open(audio_path, "rb") as audio_file:
+            transcription = client.audio.transcriptions.create(
+                model="whisper-1",
+                file=audio_file
+            )
+            
+        return transcription.text
             
     except Exception as e:
         print(f"Transcription error: {str(e)}")