djuna-test-lab/Q3-IIJAN-4B

Full fine-tune vs Intelligent-Internet/II-Search-CIR-4B. Fine-tune: 100% of the weights changed, across 36 of 36 layers.

Tensors changed: Embeddings 1/1, Norms 143/145, MLP 108/108, Attention 144/144

1 derivatives checked

• mradermacher/Q3-IIJAN-3B-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse