Nanbeige/Nanbeige4.2-3B-DSpark

Full fine-tune vs Nanbeige/Nanbeige4.2-3B-Base. Fine-tune: 100% of the weights changed, across 5 of 5 layers.

Tensors changed: Norms 11/11, MLP 15/15, Attention 20/20

2 derivatives checked

• Anbeeld/Nanbeige4.2-3B-DSpark-GGUF: Same weights, converted
• SUPEROXIDES/NANBEIGE4.2_3B_DSPARK_-_GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse