finetune_from_model

The finetune_from_model field does not contain any additional parameters. You only need to provide the path to a .ckpt, .pt, or .pth file to perform finetuning. If finetuning is not required, this field can be omitted or set to null.

Additionally, if a finetune_config.yaml file exists in the current directory, the model will default to LoRA finetuning. If this file does not exist, full-model finetuning will be applied.