marcuscedricridia/cursa-o1-7b
Full fine-tune vs Xiaojian9992024/Qwen2.5-Dyanka-7B-Preview. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Norms 55/57, MLP 84/84, Attention 196/196
1 derivatives checked
- mradermacher/cursa-o1-7b-GGUF: Same weights, converted