File size: 1,506 Bytes
dcb81ce f3e9e5d dcb81ce f3e9e5d dcb81ce f3e9e5d 09e84ad f3e9e5d 09e84ad f3e9e5d 09e84ad |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 |
---
language:
- ur
- en
license: apache-2.0
datasets:
- iwslt14
metrics:
- bleu
library_name: fairseq
pipeline_tag: translation
---
### Urdu to English Translation
Urdu to English translation model is a Transformer model trained on IWSLT back-translated data using Faireq.
This model is produced during the experimentation related to building Context-Aware NMT models for low-resourced languages such as Urdu, Hindi, Sindhi, Pashtu and Punjabi. This particular model does not contains any contextual information and it is baseline sentence-level transformer model.
The evaluation is done on WMT2017 standard test set.
* source group: English
* target group: Urdu
* model: transformer
* Contextual
* Test Set: WMT2017
* pre-processing: Moses + Indic Tokenizer
* Dataset + Libray Details: [DLNMT](https://github.com/sami-haq99/nrpu-dlnmt)
## Benchmarks
| testset | BLEU |
|-----------------------|-------|
| Wmt2017 | 50.03 |
## How to use model?
* This model can be accessed via git clone:
```
git clone https://huggingface.co/samiulhaq/iwslt-bt-en-ur
```
* You can use Fairseq library to access the model for translations:
```
from fairseq.models.transformer import TransformerModel
```
### Load the model
```
model = TransformerModel.from_pretrained('path/to/model')
```
#### Set the model to evaluation mode
```
model.eval()
```
#### Perform inference
```
input_text = 'Hello, how are you?'
output_text = model.translate(input_text)
print(output_text)
```
|