kilian-group/KBevo-Qwen3-1.7B-SFT
Full fine-tune vs Qwen/Qwen3-1.7B. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Norms 108/113, MLP 84/84, Attention 112/112
2 derivatives checked
- kilian-group/KBevo-Qwen3-1.7B-GRPO: Full fine-tune
- mradermacher/KBevo-Qwen3-1.7B-SFT-GGUF: Same weights, converted