smarttasks/Phi-mini-MoE-instruct-GGUF

Same weights, converted vs microsoft/Phi-mini-MoE-instruct. Same weights stored in another dtype (BF16 → F16/F32), plus 96 new tensors.

Tensors changed: Embeddings 0/1, other 0/32, Norms 0/130, Attention 0/256, lm_head 0/2

On the Hub · Downloads on Model Pulse