Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/shhossain/banglatranslationkit
BanglaTranslationKit is a open-source translation package for offline conversion between both Bengali and English languages (English to Bangla and Bangla to English)
https://github.com/shhossain/banglatranslationkit
bangla-ma bangla-machine-translation bangla-translator machine-translation nlp python transformers translation
Last synced: 1 day ago
JSON representation
BanglaTranslationKit is a open-source translation package for offline conversion between both Bengali and English languages (English to Bangla and Bangla to English)
- Host: GitHub
- URL: https://github.com/shhossain/banglatranslationkit
- Owner: shhossain
- License: apache-2.0
- Created: 2023-08-27T03:56:16.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2023-08-27T11:48:09.000Z (over 1 year ago)
- Last Synced: 2024-12-24T01:43:06.784Z (10 days ago)
- Topics: bangla-ma, bangla-machine-translation, bangla-translator, machine-translation, nlp, python, transformers, translation
- Language: Python
- Homepage:
- Size: 63.5 KB
- Stars: 1
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# BanglaTranslationKit
BanglaTranslationKit is an open-source translation package for offline conversion between both Bengali and English languages (English to Bangla and Bangla to English)
## Installation
```bash
pip install bntrans
```
## Usage
You can use this package to translate any Bangla to English or English to Bangla.
```python
from bntrans import Translator
translator = Translator(src="en", dest="bn")
translation = translator.translate("Hello world!") # ্যালো বিশ্ব!
print(translation)
```
### Some Useful Methods
```python
from bntrans import Translator
# You can tell the model to use cpu
translator = Translator(src="en", dest="bn", use_gpu=False)
# Specify a custom model (from huggingface.co/models or local path)
translator = Translator(model="mymodel")
# If the custom model doesn't have a maximum length, you can specify it. It is better to specify it as some models have wrong maximum length in their config.
translator = Translator(model="mymodel", max_length=512)
# You can also use the `generate` method to get raw output from the model
translator.generate("Hello world!")
```
## Use Cloud Translation API (Free)
If you want to test the model without installing it, you can use the cloud translation API. Get a free Token from [here](https://huggingface.co/settings/tokens) (Read Token is enough) and use it like this:
```python
from bntrans import Translator
translator = Translator(src="en", dest="bn", use_cloud=True, huggingface_token="YOUR_TOKEN")
translation = translator.translate("Hello world!") # ্যালো বিশ্ব!
print(translation)
```
## Use in Command Line
You can also use this package in the command line. Just install it and run the following command:
```bash
bntrans "Hello world!"
```
```bash
usage: cli.py [-h] [-s SRC] [-d DEST] [-m MODEL] [-g GPU] [-l MAX_LENGTH] [-c USE_CLOUD] [-t HUGGINGFACE_TOKEN] text
Translate text from one language to another.
positional arguments:
text Text to translate
options:
-h, --help show this help message and exit
-s SRC, --src SRC Source language (default: en)
-d DEST, --dest DEST Destination language (default: bn)
-m MODEL, --model MODEL
Model to use for translation
-g GPU, --gpu GPU Whether to use GPU for translation (default: True)
-l MAX_LENGTH, --max-length MAX_LENGTH
The maximum length of the model (default: None)
-c USE_CLOUD, --use-cloud USE_CLOUD
Whether to use huggingface inference API (default: False)
-t HUGGINGFACE_TOKEN, --huggingface-token HUGGINGFACE_TOKEN
Huggingface inference API token (default: None)
```