Stable Diffusion

4304 readers

2 users here now

Discuss matters related to our favourite AI Art generation technology

Also see

Other communities

founded 1 year ago

MODERATORS

db0@lemmy.dbzer0.com

Is it possible to train Textual Inversion embeddings using custom checkpoints? (sh.itjust.works)

submitted 1 year ago by whitecapstromgard@sh.itjust.works to c/stable_diffusion@lemmy.dbzer0.com

3 comments fedilink hide all child comments

I tried doing this using Automatic1111 and some of my favourite custom checkpoints from Civitai. The result was garbage. I can train beautiful embeddings using the base SD 1.5 model, but not using any of my favourite checkpoints.

top 3 comments

sorted by: hot top controversial new old

[–] DrakeRichards@lemmy.world 1 points 1 year ago (1 children)

Embeddings should generally be trained on base models to improve compatibility with models derived from the base. For SD 1.5, that means using either regular SD 1.5 or the NovelAI leak. You can sometimes get away with using more “basic” models that don’t have many merges, but that can be tough to gauge.

[–] whitecapstromgard@sh.itjust.works 1 points 1 year ago (1 children)

Thanks! Isn't it better to train the embeddng with the model I expect to use it with?

[–] DrakeRichards@lemmy.world 1 points 1 year ago

I don’t really understand the science behind it, but in my experience I’ve had much more success using basic models for training.

Also, I’ve found that LoRAs are generally much easier and faster to train than embeddings. Is there a reason you’re going for an embedding over a LoRA?