Logo Lanfrica

moanbekele/English-Amharic-Translator

Domaine:

natural language processing

Type de record:

software
Créateur:
moa
Hôte:
# English/Amharic Translator Machine Translation (MT) Web Interface with OpenNMT and FairSeq models using *CTranslate* and *Streamlit*. ## Install requirements It is recommended you first create a virtual environment, and then run: ``` pip3 install -r requirements.txt ``` ## Running the App With translate.py or translate-multi.py, you can run a fast web translator using CTranslate2. For this to work, you have to change the paths to your models; search the code for [Modify] and adjust the following lines. ## Training And Data Preparation * Dataset Preparation and Processing (notebook) * Training and Streamlit Interface (notebook)