Model Description

This is a fine-tuned version of Sentence-BERT (SBERT) specifically designed for the Amharic language. It was trained on a Natural Language Inference (NLI) dataset written in the Amharic language. The model outputs sentence embeddings, which are numerical representations of sentences in the form of 768-dimensional vectors.

Usage

This model can be used as input for downstream tasks such as sentiment analysis, recommendation systems, question answering, text summarization, named entity recognition, etc.

from sentence_transformers import SentenceTransformer

SentenceModel = SentenceTransformer('heran/SBERT-am-finetune')
textEncoding = SentenceModel.encode("ዛሬ አየሩ በጣም ጥሩ ነው።")

Limitations and Known Issues

It is important to note that the model was trained on a limited dataset, which means it may have inherent biases and may not perform optimally for sentences that contain infrequently used words. It is recommended to carefully evaluate the model's output and consider supplementing it with additional training data or methods to mitigate these limitations.