Canstralian
commited on
Update app.py
Browse files
app.py
CHANGED
@@ -1,80 +1,49 @@
|
|
1 |
import torch
|
2 |
from fastapi import FastAPI, HTTPException
|
3 |
from pydantic import BaseModel
|
4 |
-
import
|
5 |
-
import json
|
6 |
-
import os
|
7 |
-
|
8 |
-
# Set up logging configuration
|
9 |
-
logging.basicConfig(level=logging.INFO)
|
10 |
|
11 |
# Initialize the FastAPI app
|
12 |
app = FastAPI()
|
13 |
|
14 |
-
# Load the
|
15 |
-
|
|
|
|
|
16 |
model.eval()
|
17 |
|
18 |
# Define the input and output format for prediction requests
|
19 |
class PredictionRequest(BaseModel):
|
20 |
-
"""
|
21 |
-
Data model for the prediction request.
|
22 |
-
|
23 |
-
Attributes:
|
24 |
-
text (str): Input text for model inference.
|
25 |
-
"""
|
26 |
text: str
|
27 |
|
28 |
class PredictionResponse(BaseModel):
|
29 |
-
"""
|
30 |
-
Data model for the prediction response.
|
31 |
-
|
32 |
-
Attributes:
|
33 |
-
text (str): The original input text.
|
34 |
-
prediction (str): The predicted result from the model.
|
35 |
-
"""
|
36 |
text: str
|
37 |
prediction: str
|
38 |
|
39 |
# Define prediction endpoint
|
40 |
@app.post("/predict", response_model=PredictionResponse)
|
41 |
async def predict(request: PredictionRequest):
|
42 |
-
"""
|
43 |
-
Endpoint for generating a prediction based on input text.
|
44 |
-
|
45 |
-
Args:
|
46 |
-
request (PredictionRequest): The request body containing the input text.
|
47 |
-
|
48 |
-
Returns:
|
49 |
-
PredictionResponse: The response body containing the original text and prediction.
|
50 |
-
|
51 |
-
Raises:
|
52 |
-
HTTPException: If any error occurs during the prediction process.
|
53 |
-
"""
|
54 |
try:
|
55 |
-
# Tokenize the input text
|
56 |
-
inputs = tokenizer(request.text, return_tensors="pt")
|
57 |
|
58 |
# Perform inference with the model
|
59 |
-
|
|
|
|
|
|
|
|
|
60 |
|
61 |
-
#
|
62 |
-
|
|
|
63 |
|
64 |
# Return the prediction response
|
65 |
-
return PredictionResponse(text=request.text, prediction=
|
66 |
except Exception as e:
|
67 |
-
logging.error("Error during prediction", exc_info=True)
|
68 |
raise HTTPException(status_code=500, detail="Prediction failed")
|
69 |
|
70 |
# Define health check endpoint
|
71 |
@app.get("/health")
|
72 |
async def health_check():
|
73 |
-
"""
|
74 |
-
Health check endpoint to verify if the service is up and running.
|
75 |
-
|
76 |
-
Returns:
|
77 |
-
dict: A dictionary containing the status of the service.
|
78 |
-
"""
|
79 |
-
logging.info("Health check requested.")
|
80 |
return {"status": "healthy"}
|
|
|
1 |
import torch
|
2 |
from fastapi import FastAPI, HTTPException
|
3 |
from pydantic import BaseModel
|
4 |
+
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
|
|
|
|
|
|
|
|
|
|
5 |
|
6 |
# Initialize the FastAPI app
|
7 |
app = FastAPI()
|
8 |
|
9 |
+
# Load the model and tokenizer from Hugging Face
|
10 |
+
model_name = "Canstralian/RabbitRedux" # Replace with your model's name
|
11 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
12 |
+
model = AutoModelForSequenceClassification.from_pretrained(model_name)
|
13 |
model.eval()
|
14 |
|
15 |
# Define the input and output format for prediction requests
|
16 |
class PredictionRequest(BaseModel):
|
|
|
|
|
|
|
|
|
|
|
|
|
17 |
text: str
|
18 |
|
19 |
class PredictionResponse(BaseModel):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
20 |
text: str
|
21 |
prediction: str
|
22 |
|
23 |
# Define prediction endpoint
|
24 |
@app.post("/predict", response_model=PredictionResponse)
|
25 |
async def predict(request: PredictionRequest):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
26 |
try:
|
27 |
+
# Tokenize the input text
|
28 |
+
inputs = tokenizer(request.text, return_tensors="pt", truncation=True, padding=True)
|
29 |
|
30 |
# Perform inference with the model
|
31 |
+
with torch.no_grad():
|
32 |
+
outputs = model(**inputs)
|
33 |
+
|
34 |
+
# Get the predicted class
|
35 |
+
prediction = torch.argmax(outputs.logits, dim=-1).item()
|
36 |
|
37 |
+
# Map the prediction to a label (adjust as per your model's labels)
|
38 |
+
labels = ["Label 1", "Label 2", "Label 3"] # Replace with your actual labels
|
39 |
+
predicted_label = labels[prediction]
|
40 |
|
41 |
# Return the prediction response
|
42 |
+
return PredictionResponse(text=request.text, prediction=predicted_label)
|
43 |
except Exception as e:
|
|
|
44 |
raise HTTPException(status_code=500, detail="Prediction failed")
|
45 |
|
46 |
# Define health check endpoint
|
47 |
@app.get("/health")
|
48 |
async def health_check():
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
49 |
return {"status": "healthy"}
|