Anbeeld/MiniMax-M3-DSpark-GGUF
Same weights, converted vs nvidia/MiniMax-M3-DSpark. Same weights stored in another dtype (F32/BF16 → BF16/F32), plus 1 new tensors.
Tensors changed: other 0/3, Norms 0/25, MLP 0/18, Attention 0/24
Same weights, converted vs nvidia/MiniMax-M3-DSpark. Same weights stored in another dtype (F32/BF16 → BF16/F32), plus 1 new tensors.
Tensors changed: other 0/3, Norms 0/25, MLP 0/18, Attention 0/24