DOEJGI/genomeocean-moe-8x2.4b-upcycled-from-4b

Full fine-tune vs DOEJGI/GenomeOcean-4B. Fine-tune: 100% of the weights changed, across 24 of 24 layers.

Tensors changed: Norms 49/49, Attention 96/96

On the Hub · Downloads on Model Pulse