kakaocorp/kanana-2-1.3b-instruct
Full fine-tune vs kakaocorp/kanana-2-1.3b-base. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/1, Norms 81/129, MLP 96/96, Attention 128/128
2 derivatives checked
- winwinplatform/Kanana-2-1.3b-instruct-GGUF: Same weights, converted
- ssolunar/Kanana-2-1.3b-instruct-Q4_K_M-GGUF: Quantization
1 other repos hold byte-for-byte identical weight files.