OrobasVault/Vespera-Synapse-31B

Full fine-tune vs Cyclone-Labs/Twisted-Cyclone-31B. Fine-tune: 95% of the weights changed, across 60 of 60 layers.

Tensors changed: Embeddings 1/4, Norms 3/361, other 0/62, MLP 180/180, Attention 230/230

2 derivatives checked

• mradermacher/Vespera-Synapse-31B-i1-GGUF: Quantization
• mradermacher/Vespera-Synapse-31B-GGUF: Quantization

On the Hub · Downloads on Model Pulse