KaraKaraWarehouse/Llama-EveningMirai-Moonwalker-MS-3.3-70B

Full fine-tune vs Delta-Vector/Austral-70B-Preview. Fine-tune: 100% of the weights changed, across 80 of 80 layers.

Tensors changed: lm_head 1/1, Norms 161/161, Embeddings 1/1, MLP 240/240, Attention 320/320

2 derivatives checked

On the Hub · Downloads on Model Pulse