|
from typing import Dict, List, Any |
|
from optimum.onnxruntime import ORTModelForSequenceClassification |
|
from transformers import pipeline, AutoTokenizer |
|
|
|
|
|
class PreTrainedPipeline(): |
|
def __init__(self, path=""): |
|
|
|
model = ORTModelForSequenceClassification.from_pretrained(path) |
|
tokenizer = AutoTokenizer.from_pretrained(path, model_max_length=128) |
|
|
|
self.pipeline = pipeline("feature-extraction", model=model, tokenizer=tokenizer) |
|
|
|
|
|
def __call__(self, inputs: Any) -> List[List[Dict[str, float]]]: |
|
""" |
|
Args: |
|
data (:obj:`str`): |
|
a string containing some text |
|
Return: |
|
A :obj:`list`:. The object returned should be a list of one list like [[{"label": 0.9939950108528137}]] containing : |
|
- "label": A string representing what the label/class is. There can be multiple labels. |
|
- "score": A score between 0 and 1 describing how confident the model is for this label/class. |
|
""" |
|
|
|
def cls_pooling(pipeline_output): |
|
""" |
|
Return the [CLS] token embedding |
|
""" |
|
return [_h[0] for _h in pipeline_output] |
|
|
|
embeddings = cls_pooling(self.pipeline(inputs)) |
|
return embeddings |
