Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,53 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
import logging
|
3 |
+
from fastapi import FastAPI, HTTPException
|
4 |
+
from pydantic import BaseModel
|
5 |
+
from transformers import AutoAdapterModel, AutoTokenizer
|
6 |
+
|
7 |
+
# Initialize the app
|
8 |
+
app = FastAPI()
|
9 |
+
logging.basicConfig(level=logging.INFO)
|
10 |
+
|
11 |
+
# Load model and tokenizer once on startup
|
12 |
+
MODEL_NAME = os.getenv("MODEL_NAME", "bert-base-uncased") # Set default model
|
13 |
+
ADAPTER_NAME = os.getenv("ADAPTER_NAME", "Canstralian/RabbitRedux") # Adapter name
|
14 |
+
|
15 |
+
try:
|
16 |
+
logging.info("Loading model and adapter...")
|
17 |
+
model = AutoAdapterModel.from_pretrained(MODEL_NAME)
|
18 |
+
model.load_adapter(ADAPTER_NAME, set_active=True)
|
19 |
+
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
|
20 |
+
logging.info("Model and adapter loaded successfully.")
|
21 |
+
except Exception as e:
|
22 |
+
logging.error("Error loading model or adapter:", exc_info=True)
|
23 |
+
raise RuntimeError("Model or adapter loading failed.") from e
|
24 |
+
|
25 |
+
# Define request and response data structures
|
26 |
+
class PredictionRequest(BaseModel):
|
27 |
+
text: str
|
28 |
+
|
29 |
+
class PredictionResponse(BaseModel):
|
30 |
+
text: str
|
31 |
+
prediction: str
|
32 |
+
|
33 |
+
# Endpoint for inference
|
34 |
+
@app.post("/predict", response_model=PredictionResponse)
|
35 |
+
async def predict(request: PredictionRequest):
|
36 |
+
try:
|
37 |
+
# Tokenize input text
|
38 |
+
inputs = tokenizer(request.text, return_tensors="pt")
|
39 |
+
# Perform inference
|
40 |
+
outputs = model(**inputs)
|
41 |
+
# Generate predicted text or classification (customize as needed)
|
42 |
+
prediction = tokenizer.decode(outputs.logits.argmax(-1)[0], skip_special_tokens=True)
|
43 |
+
|
44 |
+
return PredictionResponse(text=request.text, prediction=prediction)
|
45 |
+
except Exception as e:
|
46 |
+
logging.error("Error during prediction:", exc_info=True)
|
47 |
+
raise HTTPException(status_code=500, detail="Prediction failed")
|
48 |
+
|
49 |
+
# Health check endpoint
|
50 |
+
@app.get("/health")
|
51 |
+
async def health_check():
|
52 |
+
return {"status": "healthy"}
|
53 |
+
|