louisbrulenaudet/DevPearl-2x7B
MoE built from dense models vs deepseek-ai/deepseek-coder-6.7b-instruct. louisbrulenaudet/DevPearl-2x7B is a mixture of 2 experts assembled from dense models (mergekit-moe style): expert 0 is an exact copy of deepseek-ai/deepseek-coder-6.7b-instruct's MLP; the other experts come from other models. Shared tensors (attention, norms, embeddings): 0 of 193 identical.
1 derivatives checked
- mradermacher/DevPearl-2x7B-GGUF: Quantization