Update README.md
Browse files
README.md
CHANGED
|
@@ -28,6 +28,13 @@ licenses:
|
|
| 28 |
|
| 29 |
This repository contains the [IndicBARTSS](https://huggingface.co/ai4bharat/IndicBARTSS) checkpoint finetuned on the 11 languages of [IndicParaphrase](https://huggingface.co/datasets/ai4bharat/IndicParaphrase) dataset. For finetuning details,
|
| 30 |
see the [paper](https://arxiv.org/abs/2203.05437).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
|
| 32 |
|
| 33 |
## Using this model in `transformers`
|
|
|
|
| 28 |
|
| 29 |
This repository contains the [IndicBARTSS](https://huggingface.co/ai4bharat/IndicBARTSS) checkpoint finetuned on the 11 languages of [IndicParaphrase](https://huggingface.co/datasets/ai4bharat/IndicParaphrase) dataset. For finetuning details,
|
| 30 |
see the [paper](https://arxiv.org/abs/2203.05437).
|
| 31 |
+
<ul>
|
| 32 |
+
<li >Supported languages: Assamese, Bengali, Gujarati, Hindi, Marathi, Odiya, Punjabi, Kannada, Malayalam, Tamil, and Telugu. Not all of these languages are supported by mBART50 and mT5. </li>
|
| 33 |
+
<li >The model is much smaller than the mBART and mT5(-base) models, so less computationally expensive for decoding. </li>
|
| 34 |
+
<li> Trained on large Indic language corpora (5.53 million sentences). </li>
|
| 35 |
+
<li> Unlike [MultiIndicParaphraseGeneration](https://huggingface.co/ai4bharat/MultiIndicParaphraseGeneration) each language is written in its own script, so you do not need to perform any script mapping to/from Devanagari. </li>
|
| 36 |
+
</ul>
|
| 37 |
+
|
| 38 |
|
| 39 |
|
| 40 |
## Using this model in `transformers`
|