Hosstia/albert-tiny-chinese-ner-onnx

ALBERT-tiny Chinese NER (ONNX INT8)

INT8 quantized ONNX export of ckiplab/albert-tiny-chinese-ner for token-classification (NER) tasks.

Files

• model_int8.onnx — INT8 quantized ONNX model
• vocab.txt — WordPiece vocabulary
• labels.txt — NER label set (line N = label ID N)

Usage

import onnxruntime as ort
import numpy as np

session = ort.InferenceSession("model_int8.onnx")
inputs = {
    "input_ids": np.array([[...]], dtype=np.int64),
    "attention_mask": np.array([[...]], dtype=np.int64),
    "token_type_ids": np.array([[...]], dtype=np.int64),
}
logits = session.run(None, inputs)[0]  # (batch, seq_len, num_labels)