Uploaded November 2026 | Updated November 2026, 4 days ago
Why train your own LoRA when big image models can do so much? From Week 1 of Training Image + Video LoRAs: styles base models still can't make, distilling an expensive model into a cheaper one, training on your best results for a consistent look, extending a finished project, and writing fewer prompts.
Full Week 1 lecture and lab are in the course playlist.
Slides: artificial-images.com/lora-week1
0:00 Why train a LoRA?
0:23 What a LoRA is
0:41 Why still train LoRAs in 2026
1:44 Styles base models can't make
3:23 Distilling big models into cheaper ones
4:41 Training on your best results for a consistent look
6:21 Extending a finished project
7:16 Less prompt engineering
Why train your own LoRA when big image models can do so much? From Week 1 of Training Image + Video LoRAs: styles base models still can't make, distilling an expensive model into a cheaper one, training on your best results for a consistent look, extending a finished project, and writing fewer prompts.
Full Week 1 lecture and lab are in the course playlist.
Slides: artificial-images.com/lora-week1
0:00 Why train a LoRA?
0:23 What a LoRA is
0:41 Why still train LoRAs in 2026
1:44 Styles base models can't make
3:23 Distilling big models into cheaper ones
4:41 Training on your best results for a consistent look
6:21 Extending a finished project
7:16 Less prompt engineering










