poolside/Laguna-XS-2.1-DFlash

Full fine-tune vs poolside/Laguna-XS-2.1. Fine-tune: 100% of the weights changed, across 5 of 5 layers.

Tensors changed: Norms 21/21, MLP 3/3, Attention 6/6

2 derivatives checked

• Lucebox/Laguna-XS-2.1-DFlash-GGUF: Quantization
• dev7a/Laguna-XS-2.1-DFlash-GGUF: Quantization

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse