Update app.py

app.py

@@ -1,38 +1,44 @@
+import os
+os.environ["TRANSFORMERS_CACHE"] = "/tmp/hf-cache"  # Important for Docker
+
 from fastapi import FastAPI, UploadFile, File
 from fastapi.middleware.cors import CORSMiddleware
 import torchaudio
 import torch
-from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
-import uvicorn
+from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
 import io
 
 app = FastAPI()
 
-# Allow requests from Flutter (localhost or any domain)
+# Allow all origins (for Flutter)
 app.add_middleware(
     CORSMiddleware,
     allow_origins=["*"],
-    allow_credentials=True,
     allow_methods=["*"],
     allow_headers=["*"],
 )
 
-# Load model and processor once at startup
+# Load model
 processor = Wav2Vec2Processor.from_pretrained("Mustafaa4a/ASR-Somali")
 model = Wav2Vec2ForCTC.from_pretrained("Mustafaa4a/ASR-Somali")
 
-@app.post("/transcribe")
-async def transcribe_audio(file: UploadFile = File(...)):
-    contents = await file.read()
-    audio_bytes = io.BytesIO(contents)
+@app.get("/")
+async def root():
+    return {"message": "Somali Speech-to-Text API is running."}
 
-    waveform, sample_rate = torchaudio.load(audio_bytes)
+@app.post("/transcribe")
+async def transcribe(file: UploadFile = File(...)):
+    audio_bytes = await file.read()
+    audio_stream = io.BytesIO(audio_bytes)
+    
+    waveform, sample_rate = torchaudio.load(audio_stream)
 
     if sample_rate != 16000:
         resampler = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)
         waveform = resampler(waveform)
 
     inputs = processor(waveform.squeeze(), sampling_rate=16000, return_tensors="pt")
+    
     with torch.no_grad():
         logits = model(**inputs).logits