Nexesenex/Llama_3.x_70b_Athos_v1.0

Full fine-tune vs MaziyarPanahi/calme-2.3-llama3.1-70b. 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/Llama_3.x_70b_Athos_v1.0-i1-GGUF: Quantization
• mradermacher/Llama_3.x_70b_Athos_v1.0-GGUF: Quantization

On the Hub · Downloads on Model Pulse