bartowski/Kwaipilot_KAT-Coder-V2.5-Dev-GGUF

Same weights, converted vs Kwaipilot/KAT-Coder-V2.5-Dev. Same weights stored in another dtype (BF16 → BF16/F32), plus 150 new tensors.

Tensors changed: Embeddings 0/1, lm_head 0/1, Norms 0/101, MLP 0/160, Attention 0/40

On the Hub · Downloads on Model Pulse