Cyclone-Labs/Twilight-Embrace-31B

Full fine-tune vs Cyclone-Labs/Eon-Blossom-V2-31B. Fine-tune: 93% of the weights changed, across 60 of 60 layers.

Tensors changed: Embeddings 1/4, Norms 0/523, other 0/62, MLP 179/261, Attention 230/338

2 derivatives checked

• mradermacher/Twilight-Embrace-31B-i1-GGUF: Quantization (fewer layers)
• mradermacher/Twilight-Embrace-31B-GGUF: Quantization (fewer layers)

On the Hub · Downloads on Model Pulse