LiquidAI/LFM2-24B-A2B-MLX-bf16

Same weights, repacked vs LiquidAI/LFM2-24B-A2B. Same weights in different files (re-saved, re-sharded or new metadata), plus 120 new tensors.

Tensors changed: MLP 0/54, Embeddings 0/2, other 0/60, Norms 0/100, Attention 0/40

On the Hub · Downloads on Model Pulse