mergekit-community/Deepseek-R1-Distill-NSFW-RPv1
Adapter vs Azazelle/ANJIR-ADAPTER-128. Azazelle/ANJIR-ADAPTER-128 is a LoRA adapter, not full weights: rank 128, on down_proj, gate_proj, k_proj, lm_head, o_proj, q_proj, up_proj, v_proj in 32 layers (225 adapted matrices). It modifies mergekit-community/Deepseek-R1-Distill-NSFW-RPv1 at load time.
7 derivatives checked
- mradermacher/Deepseek-R1-Distill-NSFW-RPv1-GGUF: Same weights, converted
- mradermacher/Deepseek-R1-Distill-NSFW-RPv1-i1-GGUF: Quantization
- hsefz-ChenJunJie/Deepseek-R1-Distill-NSFW-RPv1-mlx-4Bit: Quantization
- hsefz-ChenJunJie/Deepseek-R1-Distill-NSFW-RPv1-mlx-8Bit: Quantization
- hsefz-ChenJunJie/Deepseek-R1-Distill-NSFW-RPv1-mlx-fp16: Same weights, converted
- mergekit-community/NSFW-FFS-w-hidden-Deepseek-Distill-NSFW: Full fine-tune
- mergekit-community/Deepseek-Distill-NSFW-visible-w-NSFW-FFS: Full fine-tune