README.md

metadata

language: en
tags:
  - grammar
  - text2text-generation
datasets:
  - leslyarun/c4_200m_gec_train100k_test25k

Get Grammatical corrections on your English text, trained on a subset of c4-200m dataset - ONNX Quantized Model

Use the below code for running the model

from transformers import AutoTokenizer
from optimum.onnxruntime import ORTModelForSeq2SeqLM
from optimum.pipelines import pipeline

tokenizer = AutoTokenizer.from_pretrained("leslyarun/grammatical-error-correction-quantized")
model = ORTModelForSeq2SeqLM.from_pretrained("leslyarun/grammatical-error-correction-quantized",
                                              encoder_file_name="encoder_model_quantized.onnx",
                                              decoder_file_name="decoder_model_quantized.onnx",
                                              decoder_with_past_file_name="decoder_with_past_model_quantized.onnx")
  
text2text_generator = pipeline("text2text-generation", model=model, tokenizer=tokenizer)
output = text2text_generator("grammar: " + sentence)
print(output[0]["generated_text"])