oddadmix/Katib-Qwen3.5-0.8B-0.1
Adapter vs unsloth/Qwen3.5-0.8B. oddadmix/Katib-Qwen3.5-0.8B-0.1 is a LoRA adapter, not full weights: rank 32, on down_proj, gate_proj, in_proj_a, in_proj_b, in_proj_qkv, in_proj_z, k_proj, linear_fc1… in 24 layers (234 adapted matrices). It modifies unsloth/Qwen3.5-0.8B at load time.