|
--- |
|
pipeline_tag: text2text-generation |
|
|
|
widget: |
|
- text: "Hellooooo" |
|
example_title: "Ex 0" |
|
- text: "believ" |
|
example_title: "Ex 1" |
|
--- |
|
|
|
# Model Card for Model ID |
|
|
|
This is a model for word-based spell correction tasks. This model is generated by fine-tuning bart base model. |
|
|
|
This model works best for ''WORD-BASED'' spell correction(`not so good with the sequence of words`). |
|
|
|
|
|
|
|
## How to Get Started with the Model |
|
```python |
|
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM |
|
tokenizer = AutoTokenizer.from_pretrained("veghar/spell_correct_bart_base") |
|
model = AutoModelForSeq2SeqLM.from_pretrained("veghar/spell_correct_bart_base") |
|
|
|
text='believ' |
|
text_tok=tokenizer(text,padding=True, return_tensors='tf') |
|
input_ids = text_tok['input_ids'] |
|
outputs = model.generate(input_ids=input_ids, max_length=10,num_return_sequences=3) |
|
corrected_sentences = tokenizer.batch_decode(outputs, skip_special_tokens=True) |
|
|
|
print('Misspelled word:', text) |
|
print('Corrected word:', corrected_sentences) |
|
|
|
|
|
>>Misspelled word: believ |
|
>>Corrected word: ['believe', 'belief', 'believer'] |
|
``` |
