inferencerlabs/Bonsai-4B-MLX-Q2

Partial fine-tune vs prism-ml/Bonsai-4B-mlx-1bit. Partial change: 50% of the weights changed (attention, embeddings, mlp), in 36 of 36 layers.

Tensors changed: Embeddings 1/2, Norms 0/145, MLP 108/216, Attention 144/288

On the Hub · Downloads on Model Pulse