Nexesenex/Llama_3.x_70b_L3.3_UltraMedical_128K_v1.02

Full fine-tune vs TsinghuaC3I/Llama-3-70B-UltraMedical. 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

On the Hub · Downloads on Model Pulse