Summary

This has been a long chapter, but learning about the power of model training is worth the length. Once we have mastered the training skill, we can train any models based on our needs. The whole training process isn’t easy, as there are so many details and trivial things to deal with. However, writing the training code is the only way to fully understand how model training works; considering the fruitful outcome, it is worth spending time to figure it out from the bottom up.

Due to the length limitation of one chapter, I can only cover the entire LoRA training process, but once you succeed with LoRA training, you can find more training samples from Diffusers, change the code based on your specific needs, or simply write your training code, especially if you are working on a new model’s architecture.

In this chapter, we began by training one simple model; the model itself isn’t that interesting, but it helped you to understand the core steps of model...