Commit
·
021f5d1
1
Parent(s):
58d7d58
Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
tags:
|
| 5 |
+
- grammar-correction
|
| 6 |
+
license: mit
|
| 7 |
+
datasets:
|
| 8 |
+
- c4_200m
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# T5-Efficient-TINY for grammar correction
|
| 12 |
+
|
| 13 |
+
This is a [T5-Efficient-TINY](https://huggingface.co/google/t5-efficient-tiny) model that was trained on a subset of [C4_200M](https://ai.googleblog.com/2021/08/the-c4200m-synthetic-dataset-for.html) dataset to solve the grammar correction task in English. To bring additional errors, random typos were introduced to the input sentences using the [nlpaug](https://github.com/makcedward/nlpaug) library. Since the model was trained on only one task, there are no prefixes needed.
|
| 14 |
+
|
| 15 |
+
The model was trained as a part of the project during the [Full Stack Depp Learning](https://fullstackdeeplearning.com/course/2022/) course. ONNX version of the model is deployed on the [site](https://edge-ai.vercel.app/models/grammar-check) and can be run directly in the browser.
|