New Commit

app.py

@@ -1,27 +1,51 @@
-from fastapi import FastAPI, Request
+from fastapi import FastAPI, HTTPException
 from pydantic import BaseModel
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
-import torch
-import os
+import logging
+import sys
+
+# Set up logging to see errors in Space logs
+logging.basicConfig(level=logging.INFO, stream=sys.stdout)
+logger = logging.getLogger(__name__)
 
 app = FastAPI()
 
-# Load the model once at startup
-tokenizer = AutoTokenizer.from_pretrained("offiongbassey/efik-mt")
-model = AutoModelForSeq2SeqLM.from_pretrained("offiongbassey/efik-mt")
+# Declare model/tokenizer globally, load them in startup
+tokenizer = None
+model = None
+
+@app.on_event("startup")
+async def load_model():
+    global tokenizer, model
+    try:
+        logger.info("Loading tokenizer and model...")
+        # Use local model files for faster, more reliable loading
+        tokenizer = AutoTokenizer.from_pretrained("offiongbassey/efik-mt", local_files_only=True)
+        model = AutoModelForSeq2SeqLM.from_pretrained("offiongbassey/efik-mt", local_files_only=True)
+        logger.info("✅ Model loaded successfully!")
+    except Exception as e:
+        logger.error(f"❌ Failed to load model: {e}", exc_info=True)
+        # Don't raise here, let the endpoint handle the error
 
 class TranslateRequest(BaseModel):
     text: str
-    source: str  # "eng_Latn" or "ibo_Latn"
+    source: str  # "eng_Latn"
 
 @app.get("/")
 async def home():
-    return {"message": "Welcome! The Efik translation API is running."}
+    return {"message": "Efik Translation API is running.", "status": "healthy"}
 
 @app.post("/translate")
 async def translate(req: TranslateRequest):
-    input_text = f"{req.source} {req.text}"
-    inputs = tokenizer(input_text, return_tensors="pt")
-    outputs = model.generate(**inputs, max_length=128)
-    translation = tokenizer.decode(outputs[0], skip_special_tokens=True)
-    return {"translation": translation}
+    if tokenizer is None or model is None:
+        raise HTTPException(status_code=503, detail="Model is still loading or failed to load.")
+    
+    try:
+        input_text = f"{req.source} {req.text}"
+        inputs = tokenizer(input_text, return_tensors="pt")
+        outputs = model.generate(**inputs, max_length=128)
+        translation = tokenizer.decode(outputs[0], skip_special_tokens=True)
+        return {"translation": translation}
+    except Exception as e:
+        logger.error(f"Translation error: {e}", exc_info=True)
+        raise HTTPException(status_code=500, detail=f"Translation failed: {str(e)}")

requirements.txt

@@ -1,5 +1,7 @@
-fastapi
-uvicorn[standard]
-transformers
-torch
-huggingface_hub
+fastapi==0.104.1
+uvicorn[standard]==0.24.0
+transformers==4.36.2
+torch==2.1.2
+huggingface_hub==0.20.3
+sentencepiece==0.1.99  # <-- CRITICAL ADDITION
+accelerate==0.26.1     # <-- For efficient model loading