OrionLLM/Terminus-LFM2.5-350m

Full fine-tune vs LiquidAI/LFM2.5-350M. Fine-tune: 100% of the weights changed, across 16 of 16 layers.

Tensors changed: Embeddings 1/2, other 30/30, MLP 48/48, Norms 8/44, Attention 24/24

2 derivatives checked

• mradermacher/Terminus-LFM2.5-350m-i1-GGUF: Quantization
• mradermacher/Terminus-LFM2.5-350m-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse