beyoru/Luna-7B-A4B
MoE built from dense models vs beyoru/EvolLLM. beyoru/Luna-7B-A4B is a mixture of 2 experts with the same shape as beyoru/EvolLLM's MLP, but none of them is a copy of it: they come from other models or were trained further. Shared tensors (attention, norms, embeddings): 0 of 290 identical.
4 derivatives checked
- mradermacher/Luna-7B-A4B-i1-GGUF: Quantization
- mradermacher/Luna-7B-A4B-GGUF: Same weights, converted
- MuXodious/Luna-7B-A4B-PaperWitch-heresy: Abliterated
- MuXodious/Luna-7B-A4B-absolute-heresy: Abliterated