Add Qwen3-Embedding-8B converted to MLX fp16 via mlx-embeddings 668c3a6 verified bsisduck commited on 6 days ago