OmnicromsBrain/NeuralStar_FusionWriter_4x7b
MoE built from dense models vs OmnicromsBrain/Eros_Scribe-7b. OmnicromsBrain/NeuralStar_FusionWriter_4x7b is a mixture of 4 experts assembled from dense models (mergekit-moe style): expert 1 is an exact copy of OmnicromsBrain/Eros_Scribe-7b's MLP; the other experts come from other models. Shared tensors (attention, norms, embeddings): 0 of 195 identical.
2 derivatives checked
- mradermacher/NeuralStar_FusionWriter_4x7b-i1-GGUF: Quantization
- mradermacher/NeuralStar_FusionWriter_4x7b-GGUF: Quantization