Nexesenex/Llama_3.x_70b_SmarTricks_v1.30_flat

Full fine-tune vs TheDrummer/Fallen-Llama-3.3-R1-70B-v1. Fine-tune: 100% of the weights changed, across 80 of 80 layers.

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

1 derivatives checked

• mradermacher/Llama_3.x_70b_SmarTricks_v1.30_flat-i1-GGUF: Quantization

On the Hub · Downloads on Model Pulse