Vortex5/Helios-Flare-31B

Full fine-tune vs Cyclone-Labs/Lucid-Shadow-31B. Fine-tune: 98% of the weights changed, across 60 of 60 layers.

Tensors changed: Embeddings 2/4, Norms 0/523, other 0/62, MLP 181/261, Attention 231/338

2 derivatives checked

• mradermacher/Helios-Flare-31B-i1-GGUF: Quantization (fewer layers)
• mradermacher/Helios-Flare-31B-GGUF: Quantization (fewer layers)

On the Hub · Downloads on Model Pulse