Delta-Vector/Rei-V3-KTO-12B
LoRA-style vs Delta-Vector/Rei-12B-V3-Base. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 40 of 40 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/81, MLP 120/120, Attention 160/160, lm_head 0/1
8 derivatives checked
- mradermacher/Rei-V3-KTO-12B-i1-GGUF: Quantization
- mradermacher/Rei-V3-KTO-12B-GGUF: Quantization
- grimjim/Equatorium-v3-12B: Full fine-tune
- Vortex5/Mythic-Fabulist-12B: Full fine-tune
- lordalbior/TheBard-12B: Full fine-tune
- Vortex5/NoctyxCosma-12B: Full fine-tune
- Vortex5/Stellar-Umbra-12B: Full fine-tune
- pot99rta/GoldFusionReiV3-12B: Full fine-tune