Logging steps Save steps Warmup steps NEFTune alpha Extra arguments

Enable thinking

Trainable layers Trainable modules Extra modules (optional)

LoRA rank LoRA alpha LoRA dropout LoRA+ LR ratio Create new adapter

Use rslora Use DoRA Use PiSSA LoRA modules (optional) Additional modules (optional)

Beta value Ftx gamma Loss type sigmoid hinge ipo kto_pair orpo simpo Reward model

Score norm Whiten rewards