Sign In
home
models
images
videos
posts
articles
bounties
tools
challenges
events
shop

coco-anime pony model 2nd

223
2.1k
3
51
Updated: Feb 5, 2025
base modelanimeartwomangirlspony
Verified:
SafeTensor
Type
Checkpoint Merge
Stats
1,594
0
Reviews
(197)
Published
Oct 12, 2024
Base Model
Pony
Usage Tips
Clip Skip: 2
Hash
AutoV2
B56AECF80C
COCO_OIOI01's Avatar
COCO_OIOI01
License:
CreativeML Open RAIL++-M

Good to see you! Try the new version!

If you would like to use my Checkpoint Online Generation, please click here. I added it because my model cannot be generated online with civitai.

LINK. TensorArt. yodayo. SeaArt.

And if you like it, please download it, write a review and press the Like icon. All of this will be my motivation.

各バージョンの日本語説明はAbout this versionにあります。

To prevent unauthorized reproduction and for the benefit of those who wish to build LoRAs, we will be publishing models on other sites in the future. However, commercial use is prohibited. (01/21/2025)

All rights reserved. 無断転載を禁止します。(02/4/2025)

  • Notes.

This model is capable of sensitive illustrations such as nudity. Therefore, caution is required.

注意事項です

このモデルはヌードのようなセンシティブなイラストも可能です。その為注意が必要となります。

  • Usage(tips). The last one is a detailed hint on how to use it.

Enter realistic or 3D in the negative prompt if you want an anime or manga look, or anime screencap, or anime color if you want PVC (figure) or 3D look, then set the source core to photorealistic and add words like PVC or figure.

Please refer to the work I posted. If you would like, please take a look at the illustration I posted.

使い方のヒントです。

アニメやマンガ風にしたい場合はネガティブプロンプトにリアリスティックや3D、アニメのスクリーンショット、PVC（フィギュア）や3D風にしたい場合はアニメカラーを入力し、ソースコアをフォトリアリスティックに設定し、PVCやフィギュアなどの単語を追加してください。

不明な場合は私が投稿した作品を参考にしてください。

  • Basically the same usage as the animation-based pony model

This model will be stable and produce beautiful illustrations even with extremely short character counts. With the use of my recommended prompts, it is possible to reproduce the model in 90 characters or so. The model is good at both fantasy and real world illustrations. Scores are required, but basically negative prompts are not necessary.

このモデルは極端に短い字数でも安定し美麗なイラストが可能です。私の推奨するプロンプトを使えば90字前後でも再現できます。ファンタジーな描写から現実世界も得意です。スコアは必須ですが、基本的にネガティブなプロンプト必要はありません。ただし、使用するLoraによっては、アニメスクリーンキャプやリアスティック、アニメカラーと入力してバランスを取って下さい。ウェイトの数値を調整してください。

例として(realistic:0.5),

  • I'm pinning an illustration I made with about 85 words. You can refer to it if you like.

85文字程度で作ったイラストをピン止めしてます。良かったら参考にしてね

  • What kind of people would prefer this model?

People who do not want real or semi-real. I would be very happy if people who like Japanese anime and manga use this site.

どんな人におすすめですか？リアルやセミリアルを求めない人。アニメや漫画が好きな人。アニメキャラのLoraに適しています。

  • What this model does best. Strengths.

I felt that the camera excels in expressions and effects such as light, water droplets, petals, and fallen leaves. I also thought that good compositions and poses were possible.

光、水滴、花びら、落ち葉などの表現やエフェクト、構図やポーズもわりと良い。

We have added v2.0 and v1.0, which were treated as separate models. The reason for this is that there was a serious problem when updating the models that caused an error and they were not updated correctly, so they needed to be separated. It is unfortunate that the illustrations and statuses you have submitted cannot be transferred.

別モデル扱いとなっていたv2.0とv1.0を追加しました。理由としてモデルを更新する際にエラーが起きて正しく更新されない深刻な問題が起きて分ける必要がありました。投稿してくださったイラストとステータスが引き継ぎ出来ないのは残念です。

Where to check illustrations submitted to v2.0 and v1.0

v2.0およびv1.0に投稿されたイラストの確認場所です。

LINK.

v2.0 anime, PVC-2D

v1.0

Recommended minimum prompts

(score_9, score_8_up, score_8, score_9, ASCII score_9), (source_anime), (masterpiece best quality), (beautiful atmosphere), (8K, UHD, high detail RAW color art, photo film maker), (sharp eyelids, sharp eyelashes, thick eyelashes, long eyelashes), (1girl, solo, eyelashes, eyelash perm), 1girl, solo

Negative prompt: realistic,

Prompt Usage Example

  • CFG Scale: 7 - 8

  • Clip skip: 2

  • Sampler : DDPM, DPM++ 2M Karras,

  • Sampling steps: 25-35

  • Resolution: mainly tested on 832x1376, Or your choice.(I think there are other options)

  • HiRes Fix: use I didn't use it, so it's your choice,

  • VAE: SDXL VAE LINK Required Resources

Optimal size

width:1152, height:896 nealy 4:3

width:1216, height:832 nealy 3:2

width:1344, height:768 nealy 16:9

width:768, height:1344 nealy 9:16

width:832, height:1216 nealy 2:3

width:896, height:1152 3:4

width:864, height1536 16:9 (Recommended)

Uploaded v3.0! Thanks for downloading my model! Have fun!

This is the final version of this model. It retains the features of the previous models, but with more detailed and delicate outlines, backgrounds, and beautiful illustrations.

The concept of the model was to create a beautiful illustration without the use of Lora. Here it is final.

Lora makes it even more effective! So try to generate it once without Lora!

This model is a hint to help you free your imagination and desires.

  • CFG Scale: 7 - 8

  • Clip skip: 2

  • Sampler : DPM++ 2M Karras

  • Sampling steps: 25-35

  • Resolution: mainly tested on 832x1216, 896x1344, Or your choice.(I think there are other options)

  • HiRes Fix: use I didn't use it, so it's your choice,

  • VAE: SDXL

Negative prompts

nose, realistic, anime screencap, 3D, (Your choice)

easy tricks,

Adding words such as nose, realistic, anime screencap, 3D, (Your choice), etc. to the negative prompt will greatly change the taste.

Illustration of old version