Spaces:
Sleeping
Sleeping
File size: 4,228 Bytes
e705023 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 | # app.py - Deploy this to Hugging Face Spaces
# Install: pip install fastapi uvicorn torch transformers huggingface_hub
import json
import os
from pathlib import Path
import torch
import torch.nn as nn
from fastapi import FastAPI, HTTPException
from huggingface_hub import hf_hub_download
from pydantic import BaseModel
from transformers import AutoModel, AutoTokenizer
app = FastAPI(title="Sentiment Analysis API")
# Global variables for lazy loading
model = None
tokenizer = None
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# Model definition (must match training code)
class SentimentClassifier(nn.Module):
def __init__(self):
super().__init__()
self.bert = AutoModel.from_pretrained("distilbert-base-uncased")
self.dropout = nn.Dropout(0.3)
self.classifier = nn.Linear(768, 2)
def forward(self, input_ids, attention_mask, **kwargs):
outputs = self.bert(input_ids=input_ids, attention_mask=attention_mask)
pooled = outputs.last_hidden_state[:, 0]
x = self.dropout(pooled)
return self.classifier(x)
# Request/Response models
class PredictionRequest(BaseModel):
text: str
class PredictionResponse(BaseModel):
sentiment: str
confidence: float
def load_model_from_hf(repo_id: str):
"""Load model from Hugging Face on-demand"""
global model, tokenizer
if model is not None:
return # Already loaded
print(f"📥 Loading model from {repo_id}...")
# Download model files
cache_dir = "./model_cache"
Path(cache_dir).mkdir(exist_ok=True)
model_path = hf_hub_download(
repo_id=repo_id, filename="model.pt", cache_dir=cache_dir
)
config_path = hf_hub_download(
repo_id=repo_id, filename="config.json", cache_dir=cache_dir
)
# Load tokenizer
tokenizer = AutoTokenizer.from_pretrained(repo_id, cache_dir=cache_dir)
# Load model
model = SentimentClassifier()
model.load_state_dict(torch.load(model_path, map_location=device))
model.to(device)
model.eval()
print(f"✅ Model loaded successfully on {device}")
@app.on_event("startup")
async def startup_event():
"""Load model when server starts"""
# Read from environment variable or use default
REPO_ID = os.environ.get("MODEL_REPO_ID", "angeruiizz/sentiment-model")
load_model_from_hf(REPO_ID)
@app.get("/")
def root():
return {
"message": "Sentiment Analysis API",
"status": "running",
"endpoints": {
"/predict": "POST - Analyze sentiment of text",
"/health": "GET - Check if model is loaded",
"/docs": "GET - Interactive API documentation",
},
}
@app.get("/health")
def health_check():
return {
"status": "healthy",
"model_loaded": model is not None,
"device": str(device),
}
@app.post("/predict", response_model=PredictionResponse)
def predict(request: PredictionRequest):
if model is None or tokenizer is None:
raise HTTPException(status_code=503, detail="Model not loaded")
try:
# Tokenize input
inputs = tokenizer(
request.text,
return_tensors="pt",
padding=True,
truncation=True,
max_length=512,
)
inputs = {k: v.to(device) for k, v in inputs.items()}
# Get prediction
with torch.no_grad():
outputs = model(**inputs)
probs = torch.softmax(outputs, dim=1)
prediction = torch.argmax(probs, dim=1).item()
confidence = probs[0][prediction].item()
sentiment = "positive" if prediction == 1 else "negative"
return PredictionResponse(sentiment=sentiment, confidence=round(confidence, 4))
except Exception as e:
raise HTTPException(status_code=500, detail=str(e))
if __name__ == "__main__":
import uvicorn
port = int(os.environ.get("PORT", 7860)) # HF Spaces uses port 7860
print("🚀 Starting API server...")
uvicorn.run(app, host="0.0.0.0", port=port)
|