File size: 1,144 Bytes
191d1d2 |
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 |
import torch
from typing import Any, Dict, Union
from transformers import AutoModelForSequenceClassification, AutoTokenizer
class EndpointHandler:
def __init__(self, path=""):
# load model and tokenizer from path
self.tokenizer = AutoTokenizer.from_pretrained(path)
self.model = AutoModelForSequenceClassification.from_pretrained(
path, device_map="auto", trust_remote_code=True
)
self.device = "cuda" if torch.cuda.is_available() else "cpu"
def __call__(self, data: Dict[str, Any]) -> Dict[str, Union[str, float]]:
# process input
inputs = data.pop("inputs", data)
# preprocess
inputs = self.tokenizer(inputs, return_tensors="pt").to(self.device)
# pass inputs with all kwargs in data
logits = self.model(**inputs)[0]
# postprocess the prediction
predicted_class_id = int(torch.argmax(logits, dim=-1))
predicted_score = float(logits[0, predicted_class_id])
predicted_label = str(self.model.config.id2label[predicted_class_id])
return {'label': predicted_label, 'score': predicted_score}
|