File size: 1,383 Bytes
3d19edf 1de5e91 3d19edf 253f1f9 1de5e91 253f1f9 176c89c 253f1f9 2e2a941 253f1f9 781701c 1de5e91 781701c 2e2a941 253f1f9 402c710 253f1f9 2e2a941 |
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 |
import torch
import io
from typing import Any, Dict
from PIL import Image
from transformers import ViltProcessor, ViltForQuestionAnswering
class EndpointHandler:
def __init__(self, path=""):
# load model and processor from path
self.processor = ViltProcessor.from_pretrained(path)
self.model = ViltForQuestionAnswering.from_pretrained(path)
self.device = "cuda" if torch.cuda.is_available() else "cpu"
def __call__(self, data: Dict[str, Any]) -> Dict[str, Any]:
# process input
inputs = data.pop("inputs", data)
image = inputs["image"]
image = Image.open(io.BytesIO(eval(image)))
text = inputs["text"]
# preprocess
encoding = self.processor(image, text, return_tensors="pt")
outputs = self.model(**encoding)
# postprocess the prediction
logits = outputs.logits
best_idx = logits.argmax(-1).item()
best_answer = self.model.config.id2label[best_idx]
probabilities = torch.softmax(logits, dim=-1)[0]
id2label = self.model.config.id2label
answers = []
for idx, prob in enumerate(probabilities):
answer = id2label[idx]
answer_score = float(prob)
answers.append({"answer": answer, "answer_score": answer_score})
return {"best_answer": best_answer, "answers": answers}
|