From 1bb15b0d4d83cb55d3755a51211572e94db116d9 Mon Sep 17 00:00:00 2001 From: d8ahazard Date: Thu, 10 Nov 2022 11:52:11 -0600 Subject: [PATCH 1/2] Update README.md --- README.md | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/README.md b/README.md index b7762ae8..529cab26 100644 --- a/README.md +++ b/README.md @@ -113,3 +113,17 @@ If you wish to continue training a model, you can simply select the model name f ![image](https://user-images.githubusercontent.com/1633844/200369076-8debef69-4b95-4341-83ac-cbbb02ee02f6.png) + +## Memory and Optimization + +As this is based on ShivamShiaro's repo, it should be able to run under the same GPU constraints, but is not guaranteed. + +Please check out the [https://github.com/d8ahazard/sd_dreambooth_extension/discussions](discussions) page to find some possible tips and tricks to help you get this running on your setup - or share what you've done to get it working. + +## Issues + +Please be sure to use an issue template when asking for help. Some of the questions may be tedious, but I promise, they'll help me help you faster. + +[https://github.com/d8ahazard/sd_dreambooth_extension/issues/new?assignees=&labels=&template=bug_report.md&title=](Bug Report) +[https://github.com/d8ahazard/sd_dreambooth_extension/issues/new?assignees=&labels=&template=feature_request.md&title=](Feature Request) + From f33d4259f2e73fb3a1de6b2f8734e33e29691646 Mon Sep 17 00:00:00 2001 From: d8ahazard Date: Thu, 10 Nov 2022 11:56:10 -0600 Subject: [PATCH 2/2] Update README.md --- README.md | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/README.md b/README.md index 529cab26..83f66798 100644 --- a/README.md +++ b/README.md @@ -73,7 +73,7 @@ may help with lower-VRAM GPUs. *Sample guidance scale* - Like CFG Scale in Txt2Image/Img2Img, used for generating preview. -*Sample steps* - Same as sample guidance scale, but the number of steps to run to generate preview. According to (this guide)[https://github.com/nitrosocke/dreambooth-training-guide], you should train for appx 100 steps per sample image. So, if you have 40 instance/sample images, you would train for 4k steps. This is, of course, a rough approximation, and other values will have an effect on final output fidelity. +*Sample steps* - Same as sample guidance scale, but the number of steps to run to generate preview. According to [this guide](https://github.com/nitrosocke/dreambooth-training-guide), you should train for appx 100 steps per sample image. So, if you have 40 instance/sample images, you would train for 4k steps. This is, of course, a rough approximation, and other values will have an effect on final output fidelity. ### Advanced Settings @@ -118,12 +118,12 @@ If you wish to continue training a model, you can simply select the model name f As this is based on ShivamShiaro's repo, it should be able to run under the same GPU constraints, but is not guaranteed. -Please check out the [https://github.com/d8ahazard/sd_dreambooth_extension/discussions](discussions) page to find some possible tips and tricks to help you get this running on your setup - or share what you've done to get it working. +Please check out the [discussions](https://github.com/d8ahazard/sd_dreambooth_extension/discussions) page to find some possible tips and tricks to help you get this running on your setup - or share what you've done to get it working. ## Issues Please be sure to use an issue template when asking for help. Some of the questions may be tedious, but I promise, they'll help me help you faster. -[https://github.com/d8ahazard/sd_dreambooth_extension/issues/new?assignees=&labels=&template=bug_report.md&title=](Bug Report) -[https://github.com/d8ahazard/sd_dreambooth_extension/issues/new?assignees=&labels=&template=feature_request.md&title=](Feature Request) +[Bug Report](https://github.com/d8ahazard/sd_dreambooth_extension/issues/new?assignees=&labels=&template=bug_report.md&title=) +[Feature Request](https://github.com/d8ahazard/sd_dreambooth_extension/issues/new?assignees=&labels=&template=feature_request.md&title=)