Cat Tower is anime style NoobAI XL checkpoint.

VAE is baked in.

v-prediction versions are experimental models.

You need to use the webui that support v-prediction.

• ComfyUI

• reForge

• Forge (maybe)

• AUTOMATIC1111 (from v1.4, and you need to place the config file in the same folder as the checkpoint)

License

This model adheres to the NoobAI-XL License (Modified Fair AI Public License 1.0-SD) and is available for non-commercial use only. This license prohibits any form of commercialization, including but not limited to monetization or commercial use of the model, derivative models, or model-generated products.

Recommended Setting

• Steps: 25-30

• CFG scale: 5-7

• Sampler (vPred): I recommend using "Euler". Using "Euler a" sometimes generates colored skin and objects such as red, blue and so on.

• Sampler (Epsilon-pred): DPM++ 2M Karras, Euler a

• Scheduler (vPred) : Normal, Simple ("Automatic" produce noisy images.)

• Use ADetailer as needed

Positive Prompt

masterpiece, best quality

Negative Prompt

worst quality, bad quality, low quality, lowres, scan artifacts, jpeg artifacts, sketch, light particles

Merge recipes (Epsilon-pred v1.3)

1. Merging Cat Tower Epsilon-pred v1.2 and Cat Tower v-pred v1.7 by "DARE-TIES" (Cat Tower Epsilon-pred v1.3)

Merge recipes (vPred v1.7)

1. Merging the difference between CyberRealistic XL v5 and stable-diffusion-xl-base-1.0 to NoobAI-V-Pred-1.0-Version by "Add Difference" (Checkpoint A)

2. Replacing the CLIP of Checkpoint A by NoobAI-V-Pred-1.0-Version (Checkpoint B)

3. Merging Cat Carrier v3.0 to Checkpoint B by "Weighted Sum" with block (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.4,0.4,0.4,0.4,0.4,0.4,0.0,0,0,0,0) (Checkpoint C)

4. Merging Cat Tower vpred v1.5 to Checkpoint C by "Add Cosine B" with block (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint D)

5. Merging Cat Bread eps-pred v1.0 to Checkpoint D by "Weighted Sum" with block (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint E)

6. Merging the difference between Catloaf and ponyDiffusionV6XL to Checkpoint E by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint F)

7. Merging the difference between copycat-noob Vpred_v1.01 and NoobAI-V-Pred-1.0-Version to Checkpoint F by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.15,0.15,0.15,0.15,0,0,0,0,0) (Checkpoint G)

8. Merging style LoRA to Checkpoint G (Checkpoint H)

9. Add v_pred and ztsnr keys to Checkpoint H by noob_v_pencil-XL merge recipes (Cat Tower vPred v1.7)

Merge recipes (vPred v1.6)

1. Merging NoobAI-V-Pred-1.0-Version and IterComp by Noob Vpred itercomp merge's recipes. (Checkpoint A)

2. Finetuning NoobAI-V-Pred-1.0-Version with anime style dataset(0.5k) by OFT method, dim = 8, alpha = 1e-3, learning rate = 1e-4, 15000 steps

3. Merging the OFT model to Checkpoint A, ratio = 1.0 (Checkpoint B)

4. Merging Checkpoiint B to Checkpoint A by "Weighted Sum", alpha = 0.4 (Checkpoint C)

5. Merging Cat Carrier v3.0 to Checkpoint C by "Weighted Sum" with block (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.6,0.6,0.6,0.4,0.6,0.6,0.0,0,0,0,0) (Checkpoint D)

6. Merging Checkpoint D and HarmoniqMix_vPred_v1 by "Weighted Sum" with block (0,0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint E)

7. Merging the difference between Catloaf and ponyDiffusionV6XL to Checkpoint E by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint F)

8. Merging the difference between copycat-noob Vpred_v1.01 and NoobAI-V-Pred-1.0-Version to Checkpoint F by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.15,0.15,0.15,0.15,0,0,0,0,0) (Checkpoint G)

9. Merging style LoRA to Checkpoint G (Checkpoint H)

10. Add v_pred and ztsnr keys to Checkpoint H by noob_v_pencil-XL merge recipes (Cat Tower vPred v1.6)

Merge recipes (Epsilon-pred v1.2)

1. Merging the difference between cyberrealisticXL_v31 and stable-diffusion-xl-base-1.0 to NoobAI-XL Epsilon-pred v1.1 by "Add Difference", alpha=1.0. (Checkpoint A)

2. Replacing the CLIP of Checkpoint A by NoobAI-XL Epsilon-pred v1.1. (Checkpoint B)

3. Merging the difference between Cat Carrier v3.0 and Checkpoint B by "Add Difference" with blocks (0.01,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.35,0.35,0.35,0.2,0.35,0.35,0,0,0,0,0). (Checkpoint C)

4. Merging the difference between copycat-noob v0.3test and NoobAI-XL Epsilon-pred v0.75 to Checkpoint C by "Train Difference" with blocks (0,0,0,0,0,0,0,0,0.2,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0). (Checkpoint D)

5. Merging style LoRA to Checkpoint D (Cat Tower Epsilon-pred v1.2)

Merge recipes (vPred v1.5)

1. Merging NoobAI-V-Pred-1.0-Version and IterComp by Noob Vpred itercomp merge's recipes. (Checkpoint A)

2. Finetuning NoobAI-V-Pred-1.0-Version with anime style dataset(0.5k) by OFT method, dim = 8, alpha = 1e-3, learning rate = 1e-4, 15000 steps

3. Merging the OFT model to Checkpoint A, ratio = 1.0 (Checkpoint B)

4. Merging Checkpoint B and Cat Tower v1.1 Epsilon-pred by "Weighted Sum" with block (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.4,0.4,0.4,0.4,0.0,0,0,0,0) (Checkpoint C)

5. Merging Checkpoint C and HarmoniqMix_vPred_v1 by "Weighted Sum" with block (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.0,0,0,0,0) (Checkpoint D)

6. Merging the difference between Catloaf and ponyDiffusionV6XL to Checkpoint D by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint E)

7. Merging the difference between copycat-noob Vpred_v0.4 and Noobai-XL-V-pred-0.75S-Version to Checkpoint E by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.15,0.15,0.15,0.15,0,0,0,0,0) (Checkpoint F)

8. Add v_pred and ztsnr keys to Checkpoint F by noob_v_pencil-XL merge recipes (Cat Tower vPred v1.5)

Merge recipes (vPred v1.4)

1. Merging NoobAI-V-Pred-0.75S-Version and IterComp by Noob Vpred itercomp merge's recipes. (Checkpoint A)

2. Finetuning NoobAI-V-Pred-0.75S-Version with anime style dataset(0.5k) by OFT method, dim = 8, alpha = 1e-3, learning rate = 1e-4, 15000 steps

3. Merging the OFT model to Checkpoint A, ratio = 1.0 (Checkpoint B)

4. Merging Checkpoint B and Cat Tower v1.1 Epsilon-pred by "Weighted Sum" with block (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.4,0.4,0.4,0.4,0.0,0,0,0,0) (Checkpoint C)

5. Merging Checkpoint C and HarmoniqMix_vPred_v1 by "Weighted Sum" with block (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.0,0,0,0,0) (Checkpoint D)

6. Merging the difference between Catloaf and ponyDiffusionV6XL to Checkpoint D by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint E)

7. Merging the difference between copycat-noob v0.3test and NoobAI-XL Epsilon-pred 0.75-Version to Checkpoint E by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.15,0.15,0.15,0.15,0,0,0,0,0) (Checkpoint F)

8. Merging NoobAI-V-Pred-0.75S-Version and Checkpoint F by "Weighted Sum" with block (0,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1) (Cat Tower vPred v1.4)

Merge recipes (Epsilon-pred v1.1)

1. Merging NoobAI-XL Epsilon-pred 1.1-Version and cyberrealisticXL_v31, and replace CLIP by NoobAI-XL Epsilon-pred 1.1-Version. It's a same method of NoobaiCyberFix. (Checkpoint A)

2. Merging the difference between Cat Carrier v1.0 and Raehoshi illust XL v1.0 to Checkpoint A by "trainDifference" with blocks (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.4,0.2,0.2,0.4,0.6,0.6,0.6,0.2,0.2,0.2,0.2,0.2) (Checkpoint B)

3. Merging the difference between copycat-noob v0.3test and NoobAI-XL Epsilon-pred 0.75-Version to Checkpoint B by "trainDifference" with blocks (0,0,0,0,0,0,0,0,0.2,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint C)

4. Merging style LoRA to Checkpoint C (Cat Tower Epsilon-pred v1.1)

Merge recipes (vPred v1.3)

1. Merging NoobAI-V-Pred-0.65S-Version and IterComp by Noob Vpred itercomp merge's recipes. (Checkpoint A)

2. Finetuning NoobAI-V-Pred-0.65S-Version with anime style dataset(0.5k) by OFT method, dim = 8, alpha = 1e-3, learning rate = 1e-4, 15000 steps

3. Merging the OFT model to Checkpoint A, ratio = 1.0 (Checkpoint B)

4. Merging Checkpoint B and HarmoniqMix_vPred_v1 by "Weighted Sum" with block (0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.5,0.5,0.5,0.5,0.0,0,0,0,0) (Checkpoint C)

5. Merging the difference between Catloaf and ponyDiffusionV6XL to Checkpoint C by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint D)

6. Merging the difference between copycat-noob v0.3test and NoobAI-XL Epsilon-pred 0.75-Version to Checkpoint D by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0.15,0.15,0.15,0.15,0,0,0,0,0) (Checkpoint E)

7. Merging the difference between ExilluSPO Anime v1 and illustriousXL_smoothftSPO to Checkpoint D by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0.15,0.15,0.15,0.15,0,0,0,0,0) (Checkpoint F)

8. Merging NoobAI-V-Pred-0.65S-Version + (Checkpoint F - NoobAI-V-Pred-0.65S-Version) (Cat Tower vPred v1.3)

Merge recipes (vPred v1.2)

1. Merging NoobAI-V-Pred-0.65S-Version and cyberrealisticXL_v31, and replace CLIP by NoobAI-V-Pred-0.65S-Version. It's a same method of NoobaiCyberFix. (Checkpoint A)

2. Merging NoobAI-V-Pred-0.65S-Version and IterComp by Noob Vpred itercomp merge's recipes. (Checkpoint B)

3. Merging Checkpoint A * (1 - 0.5) + Checkpoint B * 0.5 (Checkpoint C)

4. Merging the difference between Cat Carrier v1.0 and Raehoshi illust XL v1.0 to Checkpoint C by "Train Difference" with block (0,0,0,0,0,0,0,0,0.25,0,0,0.25,0.45,0.45,0.45,0,0,0,0,0) (Checkpoint D)

5. Merging the difference between copycat-noob v0.3test and NoobAI-XL Epsilon-pred 0.75-Version to Checkpoint D by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint E)

6. Merging style LoRA to Checkpoint E (Checkpoint F)

7. Merging NoobAI-V-Pred-0.65S-Version + (Checkpoint F - NoobAI-V-Pred-0.65S-Version) with adjust (0,0,0,0,0,2.34,0,-0.95). (Cat Tower vPred v1.2)

Merge recipes (vPred v1.1)

1. Merging NoobAI-V-Pred-0.6-Version and cyberrealisticXL_v31, and replace CLIP by NoobAI-V-Pred-0.6-Version. It's a same method of NoobaiCyberFix. (Checkpoint A)

2. Merging the difference between Cat Carrier v1.0 and Raehoshi illust XL v1.0 to Checkpoint A by "Train Difference" with block (0,0,0,0,0,0,0,0,0.3,0,0,0.3,0.5,0.5,0.5,0,0,0,0,0) (Checkpoint B)

3. Merging the difference between copycat-noob v0.3test and NoobAI-XL Epsilon-pred 0.75-Version to Checkpoint B by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint C)

4. Merging style LoRA to Checkpoint C (Checkpoint D)

5. Merging NoobAI-V-Pred-0.6-Version and Checkpoint D by the following calc, NoobAI-V-Pred-0.6-Version + (Checkpoint D - NoobAI-V-Pred-0.6-Version). (Cat Tower vPred v1.1)

Merge recipes (vPred v1)

1. Merging NoobAI-V-Pred-0.5-Version and cyberrealisticXL_v31, and replace CLIP by NoobAI-V-Pred-0.5-Version. It's a same method of NoobaiCyberFix. (Checkpoint A)

2. Merging the subtract (HarmoniqMix_vPred_v1 - NoobAI-V-Pred-0.5-Version) to Checkpoint A by "Add Difference" and "Clamp" (Checkpoint B)

3. Merging the difference between Cat Carrier v1.0 and Raehoshi illust XL v1.0 to Checkpoint B by "Train Difference" with block (0.1,0.1,0.1,0.1,0.1,0.1,0.1,0.2,0.1,0.1,0.2,0.3,0.3,0.3,0.1,0.1,0.1,0.1,0.1) (Checkpoint C)

4. Merging NoobAI-V-Pred-0.5-Version and Checkpoint C by the following calc, NoobAI-V-Pred-0.5-Version + (Checkpoint C - NoobAI-V-Pred-0.5-Version). (Cat Tower vPred v1)

Merge recipes (Epsilon-pred v1)

1. Merging the difference between Cat Carrier v1.0 and Raehoshi illust XL v1.0 to NoobaiCyberFix v1.0, NoobaiCyberFix + (Cat Carrier - Raehoshi illust XL) x alpha (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.4,0.2,0.2,0.4,0.6,0.6,0.6,0.2,0.2,0.2,0.2,0.2), mode = Add difference, calcmode = trainDifference (Checkpoint A)

2. Merging the difference between copycat-noob v0.3test and NoobAI-XL Epsilon-pred 0.75-Version to Checkpoint A, Checkpoint A + (copycat-noob - NoobAI-XL) x alpha (0,0,0,0,0,0,0,0,0.2,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0), mode = Add difference, calcmode = trainDifference (Checkpoint B)

3. Merging style LoRA to Checkpoint B (Cat Tower Epsilon-pred v1)