mradermacher/Superthoughts-lite-v2-MOE-Llama3.2-bf16-GGUF

Same weights, converted vs Pinkstack/Superthoughts-lite-v2-MOE-Llama3.2-bf16. Same weights stored in another dtype (BF16 → F16/F32), plus 49 new tensors.

Tensors changed: Embeddings 0/1, Norms 0/33, Attention 0/64, lm_head 0/1, other 0/16

On the Hub · Downloads on Model Pulse