CivBot: CivitAI Bot Mk 2 [Illustrious & Pony & SD1.5]

v1.0
CitronLegacy
6 months ago

Help fuel my passion! Even $1 makes a difference. β˜• https://ko-fi.com/citronlegacy

Happy April Fools Day!

Civbot is upgraded to Mk 2 and is highly compatible with CivChan.

Read more...
Download (35.3 MB) Download available on desktop only

Popularity

390 ~10

Info

Latest version (v1.0): 1 File

CivBot.safetensors (35.3 MB)
To download these files, please visit this page from a desktop computer.

About this version: v1.0

The purpose of this Lora is to make more images which can be used as training data for CivitAI Bot Mk I

The CivitAI logo isn't stable but I think the overall shape and facial features are ok.

Is he supposed to be human-shaped or short? A few images show him as a short boy but some other images show his torso as if he was human-shape.

Right after making this Lora I found a really good highres image of the Civbot. I'll include it in the next version's dataset.

As you can see in my images the CivitAI logo can easily be photoshopped on to the image which gives me hope for making a nice set of high quality training data.

Trained on: 20 images (There are a lot of repeats)

Training Model: Animefull-final-pruned-fp16

flip_aug: True (This was a mistake because it ruined the CivitAI logo)

Num of Repeats: 10

Unit is Epochs or Steps: Epochs

Number of Epochs or Steps: 10

Training Batch Size: 2

Total Steps: 1000

Resolution: 512

Network Dim: Dim 32 and Alpha 16

Lora Creation Process took: 00:12:12

V1

This is trained on only 5 images of the CivitAI bot that I found on articles and ads.

Training Data: https://civitai.com/posts/2029774 (The 5ht image is the popup that asks you to subscribe to the newsletter.)

The dataset also has some other "round" humanoid robot pictures that I thought looked similar to the CivBot design.

4 Versions

πŸ˜₯ There are no CivBot: CivitAI Bot Mk 2 [Illustrious & Pony & SD1.5] v1.0 prompts yet!

Go ahead and upload yours!

No results

Your query returned no results – please try removing some filters or trying a different term.