DevQuasar/YOYO-AI.Qwen3-30B-A3B-Mixture-2507-GGUF

Same weights, converted vs YOYO-AI/Qwen3-30B-A3B-Mixture-2507. Same weights stored in another dtype (BF16 → F16/F32), plus 144 new tensors.

Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/193, MLP 0/48, Attention 0/192

On the Hub · Downloads on Model Pulse