[GH-ISSUE #43] [Feature Request] LoRA support for SD (cloud) #31

New issue

Closed

opened 2026-10-15 18:03:35 +00:00 by bsakaguchi · 1 comment

bsakaguchi commented

2026-10-15 18:03:35 +00:00

Owner

Originally created by @SoftString181 on GitHub (May 31, 2023).
Original GitHub issue: https://github.com/SillyTavern/SillyTavern-Extras/issues/43

I was wondering if it would be possible to add a LoRA feature for SD (cloud) in future updates.

I'm not sure if it's viable at this point, since Google is quite harsh with things related to SD running in Colab, but it would be nice to have some consistency to the character's look when chatting.

Thanks in advance.

Originally created by @SoftString181 on GitHub (May 31, 2023). Original GitHub issue: https://github.com/SillyTavern/SillyTavern-Extras/issues/43 I was wondering if it would be possible to add a LoRA feature for SD (cloud) in future updates. I'm not sure if it's viable at this point, since Google is quite harsh with things related to SD running in Colab, but it would be nice to have some consistency to the character's look when chatting. Thanks in advance.

bsakaguchi closed this issue

2026-10-15 18:03:36 +00:00

bsakaguchi commented

2026-10-15 18:03:36 +00:00

Author

Owner

@Cohee1207 commented on GitHub (Jun 1, 2023):

No, it won't be possible. The default SD pipeline doesn't have support for LoRAs or hi-res fixes, or any upscalers. Your only bet is running AUTOMATIC1111 locally, as I see no point in reimplementing features from SD Web UI here, it's not a goal of this project. LoRAs from remote SD already work, just specify the lora name in the prompt prefix as you would in the Web UI prompt.

@Cohee1207 commented on GitHub (Jun 1, 2023): No, it won't be possible. The default SD pipeline doesn't have support for LoRAs or hi-res fixes, or any upscalers. Your only bet is running AUTOMATIC1111 locally, as I see no point in reimplementing features from SD Web UI here, it's not a goal of this project. LoRAs from remote SD already work, just specify the lora name in the prompt prefix as you would in the Web UI prompt.