From 327162d1705665782d57ba0b048d844fcf02bc01 Mon Sep 17 00:00:00 2001 From: Redwan Date: Thu, 18 Apr 2024 23:45:18 +0200 Subject: [PATCH] Delete _posts/2023-06-09-Stable-Diffusion-Journey1.md --- .../2023-06-09-Stable-Diffusion-Journey1.md | 29 ------------------- 1 file changed, 29 deletions(-) delete mode 100644 _posts/2023-06-09-Stable-Diffusion-Journey1.md diff --git a/_posts/2023-06-09-Stable-Diffusion-Journey1.md b/_posts/2023-06-09-Stable-Diffusion-Journey1.md deleted file mode 100644 index ca129b632cc4..000000000000 --- a/_posts/2023-06-09-Stable-Diffusion-Journey1.md +++ /dev/null @@ -1,29 +0,0 @@ ---- -published: true ---- -## Journey to train my own Stable Diffusion - -My goal is to train my own AI model using photos of myself. -Stable Diffuion is the open source text to image model that I found. There is a lot of guide an tutorials. -The process requires a lot of ressources and I dont want to spin up a server for this. -I decided to use the service [RunDiffusion.](https://app.rundiffusion.com/) -- 1$ per hours -- 35$/months to have access to creator mode and advanced extension - -It's a no brainer. - -For the rest I am going to use -- [automatic1111](https://github.com/AUTOMATIC1111/stable-diffusion-webui) the web UI of SD. -- [Dreambooth](https://dreambooth.github.io/) to train my model -20 photos of my in 512 by 512. -- [Euler A sampler to train my model](https://stable-diffusion-ui.github.io/docs/samplers/) - - - -To do this I watched a lot of tutorials but 2 stood out. -- [Reddit post about custom embedding](https://www.reddit.com/r/StableDiffusion/comments/zxkukk/detailed_guide_on_training_embeddings_on_a/?utm_source=embedv2&utm_medium=post_embed&utm_content=action_bar&embed_host_url=https%3A%2F%2Flearn.rundiffusion.com%2Ftraining-embeddings%2F) -- [This great 3hours tutorial from Revolved](https://www.youtube.com/watch?v=IPI7QmMOUU8&) - -Test on going stay put. - -![Screenshot 2023-06-09 at 18.17.40.png](/_posts/Screenshot 2023-06-09 at 18.17.40.png)