Darkmere-8B-v0.1

The 8B version of Darkmere fine-tune, based on Ministral 3 8B Instruct 2512. The 14B version is available here.

The SillyTavern preset is available here.

Training Notes

• Training method: Full fine-tuning (not LoRA)
• Context Length: 16384
• Learning Rate: 5e-6
• Vision: the vision encoder was frozen during training, so the model retains its native vision capabilities.

Special Thanks

This fine-tune wouldn't be possible without the incredible work of the community:

• p-e-w for developing Heretic - an essential tool for censorship removal.
• Mistral AI for their Ministral 3 weights.
• AMD for their Instinct™ MI300X GPU.