SL-AI/GRaPE-Nano

Full fine-tune vs LiquidAI/LFM2-700M. Fine-tune: 100% of the weights changed, across 16 of 16 layers.

Tensors changed: Embeddings 1/2, other 30/30, MLP 48/48, Norms 19/44, Attention 24/24

1 derivatives checked

• mradermacher/GRaPE-Nano-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse