model updated

app/routers/V1/voice/voice_router.py

@@ -23,25 +23,25 @@ async def transcribe_audio(
 ):
     try:
         # Validate URL
-        if not original_url.endswith(".wav"):
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail="Invalid URL. Please provide a URL pointing to a wav file.",
-            )
+        # if not original_url.endswith(".wav"):
+        #     raise HTTPException(
+        #         status_code=status.HTTP_400_BAD_REQUEST,
+        #         detail="Invalid URL. Please provide a URL pointing to a wav file.",
+        #     )
         
-        # Download the audio file from the URL
-        response = requests.get(original_url)
-        if response.status_code != 200:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail="Unable to download the audio file from the URL.",
-            )
+        # # Download the audio file from the URL
+        # response = requests.get(original_url)
+        # if response.status_code != 200:
+        #     raise HTTPException(
+        #         status_code=status.HTTP_400_BAD_REQUEST,
+        #         detail="Unable to download the audio file from the URL.",
+        #     )
 
-        filename_original = f"audio_{int(time.time())}_original.wav"
+        # filename_original = f"audio_{int(time.time())}_original.wav"
         
-        # Save the downloaded file temporarily
-        with open(filename_original, "wb") as buffer:
-            buffer.write(response.content)
+        # # Save the downloaded file temporarily
+        # with open(filename_original, "wb") as buffer:
+        #     buffer.write(response.content)
 
         # Read file bytes
         recorded_bytes = await recorded.read()
@@ -65,7 +65,7 @@ async def transcribe_audio(
             )
         finally:
             # Clean up the temporary file
-            os.remove(filename_original)
+            # os.remove(filename_original)
             os.remove(filename_recorded)
 
     except Exception as e:

app/transcriber.py

@@ -6,8 +6,9 @@ from datasets import load_dataset
 device = "cuda:0" if torch.cuda.is_available() else "cpu"
 torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
 
-model_id = "openai/whisper-large-v3"
+# model_id = "openai/whisper-large-v3"
 # model_id = "MothersTongue/mother_tongue_model"
+model_id = "MothersTongue/mother_tongue_model_v3"
 
 model = AutoModelForSpeechSeq2Seq.from_pretrained(
     model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True