BenevolenceMessiah/QwQ-32B-Coder-Fusion-9010-1x-TIES-v1.0

Full fine-tune vs huihui-ai/QwQ-32B-Coder-Fusion-9010. Fine-tune: 100% of the weights changed, across 64 of 64 layers.

Tensors changed: Norms 129/129, MLP 192/192, Attention 448/448

1 derivatives checked