ND911/Mistralv0.2-StarlingLM-FrankenMaid-10.5B

Full fine-tune (fewer layers) vs Nexusflow/Starling-LM-7B-beta. Nexusflow/Starling-LM-7B-beta matches part of ND911/Mistralv0.2-StarlingLM-FrankenMaid-10.5B (289 of its 426 tensors; the rest of ND911/Mistralv0.2-StarlingLM-FrankenMaid-10.5B has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Norms 65/65, MLP 96/96, Attention 128/128

1 derivatives checked

On the Hub · Downloads on Model Pulse