[Andrew]

Hi, the notebook in the quick start guide is for using Stable Diffusion Webui (A1111)

You can access the training notebooks and images on the resources page.

https://stable-diffusion-art.com/members-resources/

For training SD 1.5 models, see the section: “Train LoRA for Stable Diffusion 1.5”

For training SDXL models, see the section: “Train LoRA for Stable Diffusion XL”

Please let me know if you have any questions!

[Andrew]

Hi David, you can access the Colab notebok here:

https://stable-diffusion-art.com/notebook

[Andrew]

Hi, here’s the instructions for installing an upscaler.

https://stable-diffusion-art.com/ai-upscaler/#Installing_new_upscaler

What you did sounds right. It seems that the file you downloaded may be corrupted. You can follow the links in the article to download one, put it in the ESRGAN folder, restart A1111, and try again.

[Andrew]

Hi, we can discuss this in the lessons. I have replied to you there.

[Andrew]

openpose is for human poses. You can try Canny to copy outlines instead

[Andrew]

1. Yes, any SD 1.5 checkpoint models would work.

2. ComfyUI manager may not upload all nodes correctly. In terminal, Go into custom_nodes folder and do a git pull in each folder.

I see you are not using Windows. (Mac?)

I don’t understand the error. The fashion generator doesn’t use DWPose. Was the error message from running the generator?

1. Both ckpt and safetensors are OK.

[Andrew]

Yes, it is on my list!

[Andrew]

Hi! I updated the notebook to resolve the insight face error. I also added a few IP adapter models by default.

Your error in uploading the image is not related. It could be caused by the browser. Chrome should work. You should be able to drag and drop an image. Use png or jpg format.

[Andrew]

Hi! This is a common issue. You simply don’t have a training images for small faces.

You will need to paint the face big to get the face. Try turning on Hi-Res fix. Try upscaling the picture 2x, then inpaint the face with the lora.

[Andrew]

Hi! You can use “aggressive stance”, “aggressive”, etc. If they don’t work for you, it maybe easier to use controlnet to copy a pose.

[Andrew]

Welcome! I am sure you will enjoy the flexibility of SD!

[Andrew]

Welcome!

[Andrew]

Hi, I just tested the workflow. It is working correctly. You can
– Updating your comfyui with comfyui manager: click “update all”
– Make sure you have loaded the controlnet and checkpoint model.

[Andrew]

Hi, would you be able to follow the training example and reproduce the result? The first step is to make sure your training process is correct.

The notebook is recently updated — Did you use the newly updated one, or the koyha_ss (gradio-based) one? Did you train SDXL or SD 1.5?

[Andrew]

Unfortunately I don’t have experience on this. I would change the obvious targets like learning rate and number of epochs one at at time to see if they move the needle.

[Author]

Posts