cnmoro/bert-tiny-embeddings-english-portuguese

Model Trained Using AutoTrain

• Problem type: Sentence Transformers

Validation Metrics

loss: 0.056979671120643616

Info

This is the bert-tiny model finetuned on 15B tokens for embedding/feature extraction, for English and Brazillian Portuguese languages.

The output vector size is 128.

This model only has 4.4M params but the quality of the embeddings punch way above its size after tuning.

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the Hugging Face Hub
model = SentenceTransformer("cnmoro/bert-tiny-embeddings-english-portuguese")
# Run inference
sentences = [
    'first passage',
    'second passage'
]
embeddings = model.encode(sentences)
print(embeddings.shape)