tdot604/Obsidian-70b

Full fine-tune vs BeaverLegacy/Moist-Miqu-70B-v1.1. Fine-tune: 100% of the weights changed, across 80 of 80 layers.

Tensors changed: Norms 161/161, MLP 240/240, Attention 320/320

2 derivatives checked

• mradermacher/Obsidian-70b-i1-GGUF: Quantization
• mradermacher/Obsidian-70b-GGUF: Quantization

On the Hub · Downloads on Model Pulse