File size: 1,407 Bytes
69847a0
c5a0bf8
20688a8
c5a0bf8
c3ffcdd
4814cd0
29e22ca
c5a0bf8
20688a8
4814cd0
 
 
 
 
 
 
 
69847a0
4814cd0
69847a0
4814cd0
c5a0bf8
 
 
4814cd0
c3ffcdd
4814cd0
 
 
 
 
 
 
 
 
 
 
 
 
c3ffcdd
 
4814cd0
 
c3ffcdd
69847a0
4814cd0
 
 
 
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
from fastapi import FastAPI, Request
from pydantic import BaseModel
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
import torch
import time
import logging

app = FastAPI()

# Logging setup
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger("summarizer")

# Model & tokenizer
MODEL_NAME = "VietAI/vit5-base-vietnews-summarization"
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
model = AutoModelForSeq2SeqLM.from_pretrained(MODEL_NAME)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

class InputText(BaseModel):
    text: str

@app.post("/summarize")
async def summarize(req: Request, input: InputText):
    start_time = time.time()
    logger.info(f"\U0001F535 Received request from {req.client.host}")

    text = input.text.strip()
    inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=512).to(device)

    outputs = model.generate(
        **inputs,
        max_length=128,
        num_beams=2,
        no_repeat_ngram_size=2,
        early_stopping=True
    )
    summary = tokenizer.decode(outputs[0], skip_special_tokens=True)

    end_time = time.time()
    duration = end_time - start_time
    logger.info(f"\u2705 Response sent — total time: {duration:.2f}s")

    return {"summary": summary}

@app.get("/")
def root():
    return {"message": "Vietnamese Summarization API is up and running!"}