mrcuddle/Dark-Hermes3-Llama3.2-3B
LoRA-style vs NousResearch/Hermes-3-Llama-3.2-3B. 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 28 of 28 layers; norms, biases and embeddings are untouched.
Tensors changed: Norms 0/57, MLP 84/84, Attention 112/112
6 derivatives checked
- mradermacher/Dark-Hermes3-Llama3.2-3B-i1-GGUF: Quantization
- mradermacher/Dark-Hermes3-Llama3.2-3B-GGUF: Same weights, converted
- mrcuddle/DarkHermes3-3B-Humanlike: LoRA-style
- Nexesenex/mrcuddle_DarkHermes3-Llama3.2-3B-Instruct: Full fine-tune
- mrcuddle/DarkHermes3-Llama3.2-3B-Instruct: Full fine-tune
- mrcuddle/Dark-Hermes3-Llama3.2-3B-Func: Full fine-tune