SamLowe/roberta-base-go_emotions-onnx

Cómo usar

Usando las Clases de ONNX de la Biblioteca Optimum

sentences = ["ONNX es extremadamente rápido para lotes pequeños. Impresionante"]

from transformers import AutoTokenizer, pipeline
from optimum.onnxruntime import ORTModelForSequenceClassification

model_id = "SamLowe/roberta-base-go_emotions-onnx"
file_name = "onnx/model_quantized.onnx"

model = ORTModelForSequenceClassification.from_pretrained(model_id, file_name=file_name)
tokenizer = AutoTokenizer.from_pretrained(model_id)

onnx_classifier = pipeline(
  task="text-classification",
  model=model,
  tokenizer=tokenizer,
  top_k=None,
  function_to_apply="sigmoid",  # opcional ya que es el valor predeterminado para la tarea
)

model_outputs = onnx_classifier(sentences)

print(model_outputs)

Usando ONNXRuntime

from tokenizers import Tokenizer
import onnxruntime as ort

from os import cpu_count
import numpy as np  # solo se usa para la posprocesamiento sigmoid

sentences = ["hola mundo"]  # por ejemplo un lote de 1

labels = ['admiration', 'amusement', 'anger', 'annoyance', 'approval', 'caring', 'confusion', 'curiosity', 'desire', 'disappointment', 'disapproval', 'disgust', 'embarrassment', 'excitement', 'fear', 'gratitude', 'grief', 'joy', 'love', 'nervousness', 'optimism', 'pride', 'realization', 'relief', 'remorse', 'sadness', 'surprise', 'neutral']

tokenizer = Tokenizer.from_pretrained("SamLowe/roberta-base-go_emotions")

params = {**tokenizer.padding, "length": None}
tokenizer.enable_padding(**params)

tokens_obj = tokenizer.encode_batch(sentences)

def load_onnx_model(model_filepath):
  _options = ort.SessionOptions()
  _options.inter_op_num_threads, _options.intra_op_num_threads = cpu_count(), cpu_count()
  _providers = ["CPUExecutionProvider"]  # podría usar ort.get_available_providers()
  return ort.InferenceSession(path_or_bytes=model_filepath, sess_options=_options, providers=_providers)

model = load_onnx_model("path_to_model_dot_onnx_or_model_quantized_dot_onnx")
output_names = [model.get_outputs()[0].name]  # por ejemplo ["logits"]

input_feed_dict = {
  "input_ids": [t.ids for t in tokens_obj],
  "attention_mask": [t.attention_mask for t in tokens_obj]
}

logits = model.run(output_names=output_names, input_feed=input_feed_dict)[0]

def sigmoid(x):
  return 1.0 / (1.0 + np.exp(-x))

model_outputs = sigmoid(logits)

for probas in model_outputs:
  top_result_index = np.argmax(probas)
  print(labels[top_result_index], "con puntuación:", probas[top_result_index])