Info, Guides and Lessons Learned

Here you can find various pages for guides, info and lessons learned

Misc Info

• For SD3 and Flux (gated models in general) training you need a read token from HuggingFace (In HF go in your settings and create a read token) and set it in as an environment variable named "HF_TOKEN". Other solutions exist like a Command line instruction (Discord discussion) or using a local copy (Discord discussion) of the HF folder.
• Caption tokens limit: actually, most of them are limited to 77 tokens (75 without start and end tokens) because of clip. The only exceptions are: pixart (alpha/sigma): 120, sana: 300. May change in the future.
• Common Mistakes Coming From Kohya If you're failing to get similar results that you could get in Kohya.

Diffusion Models

• Diffusion Models Overview
• Info on Flux

Guides

• OneTrainer March 2024 Guide (WIP) A more exhaustive guide on OneTrainer built as a checklist.
• Run One trainer on Runpod, How to run OT on Runpod with CLI, still relevant but consider the Cloud Training instead. Running on Runpod or any remote Linux server is now a built-in feature and offers many advantages.

Other tools - Helpful Links

Other Tools - Helpful Links

Lessons Learned

• Frequently Asked Questions
• Lessons Learnt and Tutorials