To use this model, make sure to enable trust_remote_code. To tweak the layer order, go into config.json and change the layer_exec_plan. Normally (for llama 8b) it would be [0...31] but you can change this to be any combination of layers you want (including repetitions).
