Update app.py

app.py

@@ -30,15 +30,21 @@ else:
 # Move model to the selected device (either GPU or CPU)
 pipe = pipe.to(device)
 
-# Function to process audio (transcription, translation, image generation)
+# Function to transcribe, translate, and analyze sentiment
 def process_audio(audio_path, image_option):
     if audio_path is None:
         return "Please upload an audio file.", None, None, None
 
-    # Step 1: Transcribe audio using Whisper
+    # Step 1: Transcribe audio
     try:
-        transcription = whisper_model.transcribe(audio_path)
-        tamil_text = transcription['text']
+        with open(audio_path, "rb") as file:
+            transcription = client.audio.transcriptions.create(
+                file=(os.path.basename(audio_path), file.read()),
+                model="whisper-large-v3",
+                language="ta",
+                response_format="verbose_json",
+            )
+        tamil_text = transcription.text
     except Exception as e:
         return f"An error occurred during transcription: {str(e)}", None, None, None
 
@@ -53,7 +59,9 @@ def process_audio(audio_path, image_option):
     image = None
     if image_option == "Generate Image":
         try:
-            # Ensure the model runs on the correct device (GPU/CPU)
+            model_id1 = "dreamlike-art/dreamlike-diffusion-1.0"
+            pipe = StableDiffusionPipeline.from_pretrained(model_id1, torch_dtype=torch.float16, use_safetensors=True)
+            pipe = pipe.to("cuda")
             image = pipe(translation).images[0]
         except Exception as e:
             return tamil_text, translation, f"An error occurred during image generation: {str(e)}"