CivitAIArchive
    Cat Tower (NoobAI XL checkpoint)
    NSFW
    Checkpoint
    NoobAI
    by nuko_masshigura
    anime
    base model
    Preview 43303060Preview 43303131Preview 43303173Preview 43304427Preview 43303291Preview 43303368Preview 43303461Preview 43303541Preview 43303608Preview 43303694Preview 43303866Preview 43303910Preview 43303952Preview 43303994Preview 43559804Preview 43304032Preview 43304466

    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.

    • 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)

    Version: v1.3 (vPred)
    NoobAI Standard

    • NoobAI-V-Pred-0.65S-Version based

    • I changed the merge recipes because the previous merge recipes made too high brightness model.

    1,001Downloads
    AvailableCivitAI Status
    -Deleted
    12/2/2024Created

    Files

    catTowerNoobaiXL_v13Vpred.safetensors

    6.46 GB

    SHA256: 147d73e087631707d93ae5e3d041fcae71f32fc2f4476dbe4bb05ac6d7d33779
    Loading...

    catTowerNoobaiXL_v13Vpred.yaml

    0.00 GB

    SHA256: e1bffa820ca38f9f61fd9d881ccfb76df9c9ca762483101ed2212151032d3f70
    Loading...