metadata
language: en
tags:
- sentence correction
- text2text-generation
license: cc-by-nc-sa-4.0
datasets:
- jfleg
Model
This model utilises T5-base sentence correction pre-trained model. It was fine tuned using a modified version of the JFLEG dataset and Happy Transformer framework. This model was pre-trained for educational purposes only for correction on local Caribbean dialect. For more on Caribbean dialect checkout the library Caribe. .
Re-training/Fine Tuning
The results of fine-tuning resulted in a final accuracy of 90%
Usage
from happytransformer import HappyTextToText, TTSettings
pre_trained_model="T5"
model = HappyTextToText(pre_trained_model, "KES/T5-KES")
arguments = TTSettings(num_beams=4, min_length=1)
sentence = "Wat iz your nam"
correction = model.generate_text("grammar: "+sentence, args=arguments)
if(correction.text.find(" .")):
correction.text=correction.text.replace(" .", ".")
print(correction.text) # Correction: "What is your name?".
_
Usage with Transformers
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("KES/T5-KES")
model = AutoModelForSeq2SeqLM.from_pretrained("KES/T5-KES")
text = "I am lived with my parenmts "
inputs = tokenizer("grammar:"+text, truncation=True, return_tensors='pt')
output = model.generate(inputs['input_ids'], num_beams=4, max_length=512, early_stopping=True)
correction=tokenizer.batch_decode(output, skip_special_tokens=True)
print("".join(correction)) #Correction: I am living with my parents.