KTransformers

DPO Status

DPO is not documented as current KT fine-tuning support yet.

The current public fine-tuning section should stay focused on MoE LoRA SFT through LLaMA-Factory. DPO can be added only after the exact KT path is confirmed:

ItemRequired evidence
Training entryCurrent LLaMA-Factory command, not an old patching path.
KT backendExplicit kt_config backend that maps to current KT SFT code.
ModelExact checkpoint and prepared expert weights if needed.
RuntimeSmoke result on a named machine and environment.
OutputAdapter files and at least one minimal post-training sanity check.

Until then, older DPO pages should be treated as historical references rather than user-facing instructions.