Amharic BERT
Collection
BERT transformer encoder models pretrained on 290 million tokens of Amharic text
•
6 items
•
Updated
•
3
This model is a fine-tuned version of bert-medium-amharic on the amharic-sentiment dataset for sentiment classification. It achieves the following results on the evaluation set:
Accuracy:
0.83Precision:
0.84Recall:
0.83F1:
0.83You can use this model directly with a pipeline for text classification:
>>> from transformers import pipeline
>>> bert_sentiment = pipeline("text-classification", model="rasyosef/bert-medium-amharic-finetuned-sentiment")
>>> bert_sentiment(["አሪፍ ፊልም ነው።", "ዩክሬን እና ሩስያ ከባድ ውግያ ላይ ናቸው።"])
[{'label': 'positive', 'score': 0.9863048791885376},
{'label': 'negative', 'score': 0.9570127129554749}]
https://github.com/rasyosef/amharic-sentiment-classification