all-MiniLM-L6-v2-bf16

mlx-community/all-MiniLM-L6-v2-bf16

The Model mlx-community/all-MiniLM-L6-v2-bf16 was converted to MLX format from sentence-transformers/all-MiniLM-L6-v2 using mlx-lm version 0.0.3.

Use with mlx

pip install mlx-embeddings

from mlx_embeddings import load, generate
import mlx.core as mx

model, tokenizer = load("mlx-community/all-MiniLM-L6-v2-bf16")

# For text embeddings
output = generate(model, processor, texts=["I like grapes", "I like fruits"])
embeddings = output.text_embeds  # Normalized embeddings

# Compute dot product between normalized embeddings
similarity_matrix = mx.matmul(embeddings, embeddings.T)

print("Similarity matrix between texts:")
print(similarity_matrix)