agentlans/LFM2.5-230M-Propella
Partial fine-tune vs LiquidAI/LFM2.5-230M. Partial change: 66% of the weights changed (attention, mlp, other), in 13 of 14 layers.
Tensors changed: Embeddings 0/2, other 14/24, MLP 39/42, Norms 0/40, Attention 24/24
Partial fine-tune vs LiquidAI/LFM2.5-230M. Partial change: 66% of the weights changed (attention, mlp, other), in 13 of 14 layers.
Tensors changed: Embeddings 0/2, other 14/24, MLP 39/42, Norms 0/40, Attention 24/24