JohnP1/d1a-e2b
Adapter vs google/gemma-4-E2B. JohnP1/d1a-e2b is a LoRA adapter, not full weights: rank 16, on down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj in 35 layers (205 adapted matrices). It modifies google/gemma-4-E2B at load time.
1 derivatives checked
- JohnP1/d1a-e2b-mlx-q8: Adapter