ka chun cheng (@creator_kachun) - Twitter Profili

ka chun cheng retweetledi

SuperSplat Editor 2.30.0 is here! 🚀 Your free and open source 3D Gaussian splat editor ✅ @XGRIDS_OFFICIAL LCC2 import ✅ Selection tool upgrades ✅ Save settings between sessions ...and much more! Links 👇

English

3

14

110

6.1K

ka chun cheng@creator_kachun·11h

@OriSilver We can’t use Claude here, is it possible to run with other agents?

English

1

0

43

Ori Silver@OriSilver·20h

The skill now supports 33 languages... (No translation) this is native. - Styles are kept consistent - Research is done in the native language it is prompted in - Produces locally accurate languages (not translated from English) This is Google Omni x Maxfusion AI MCP x Opus 4.8

Ori Silver@OriSilver

x.com/i/article/2074…

English

4

7

76

11.5K

ka chun cheng@creator_kachun·1d

IC-LoRA-CrossView-Prompt @ChetiArt is very impressive ! huggingface.co/Cseti/LTX2.3-2…

English

0

16

ka chun cheng retweetledi

Marco@ai_artworkgen·5 Tem

Really solid structure for bringing out strong performances in Seedance. 👇 Check out @_OAK200's breakdown below ...

OAK@_OAK200

Character sheets are great, but they only solve one part of the problem. They show you how your AI character looks. They do not show you how they perform. So I made a simple workflow and a system prompt to audition my AI characters before taking them into video generation, to test their voice, emotions, expressions, and screen presence. Here is my workflow:

English

11

14

176

27.1K

ka chun cheng retweetledi

HARUKI KAWAI｜AI美女アーティスト@ai_haruki3·5 Tem

ようやく納得の顔LoRAが作成できたので量産開始。私の設定は（添付も参照）・画像は25枚前後・Model Achitecture「Krea 2（Raw）」・optimizer「automagic3」・Timestep Type「Sigmoid」・Steps「3000」だけど結果2000～2750の間で最適・Resolutions「1024」のみ、これが重要でした optimizerはリストにないので「automagic3」に変更する方法は👇

HARUKI KAWAI｜AI美女アーティスト@ai_haruki3

ようやくKrea 2の顔LoRA作成のコツがわかった気がする…（たぶん）合ってるかわからないですが、AI Toolkitでは👇がそれなりにいい感じなので誰かのお役に立てれば幸いです。こうしたら良くなったよってあれば、アドバイスもらえると助かります。 ✅ 教師画像は25枚前後でOK（もっと少なくてもいいかも） ✅ キャプションは自然言語でもタグ形式でもあまり変わらない気がする ✅ ステップ数は1,200～1,500で十分 ✅ Modelは、Turbo用のLoRAでも「Krea 2(Raw)」選択 ✅ Optimizer「Automagic V2」へ変更 ✅ sample画像の設定は不要（あまり意味ないし、サンプル画像を作るのにやけに時間かかる）

日本語

7

11

209

25.2K

ka chun cheng retweetledi

青龍聖者@bdsqlsz·2 Tem

Someone release Krea2 and Ideogram4 edit node！

English

12

20

260

25.9K

ka chun cheng@creator_kachun·1 Tem

@ai_haruki3 I will try the optimizer next time. I found that there are still some concepts bleeding even I captioned the unwanted parts. Do you have similar experience?

English

0

1

795

HARUKI KAWAI｜AI美女アーティスト@ai_haruki3·30 Haz

ようやくKrea 2の顔LoRA作成のコツがわかった気がする…（たぶん）合ってるかわからないですが、AI Toolkitでは👇がそれなりにいい感じなので誰かのお役に立てれば幸いです。こうしたら良くなったよってあれば、アドバイスもらえると助かります。 ✅ 教師画像は25枚前後でOK（もっと少なくてもいいかも） ✅ キャプションは自然言語でもタグ形式でもあまり変わらない気がする ✅ ステップ数は1,200～1,500で十分 ✅ Modelは、Turbo用のLoRAでも「Krea 2(Raw)」選択 ✅ Optimizer「Automagic V2」へ変更 ✅ sample画像の設定は不要（あまり意味ないし、サンプル画像を作るのにやけに時間かかる）

日本語

5

6

150

32K

ka chun cheng@creator_kachun·30 Haz

@skbulous Very useful ~ any GitHub url?

English

1

0

1

70

SKB@skbzz7·27 Haz

LoRA testing is pure brain rot. I added one helper, then ten more, and now there’s a Node package generated entirely by my mental illness.

English

3

0

6

870

ka chun cheng retweetledi

⚙gear machine@AI@grmchn4ai·29 Haz

ここでツールの超お得情報なのですが、実はサンプル姉さんの動画は全てダウンロードできます。私が動いた動画をSCAIL2モデルでサンプル姉さん画像を動かした動画です。商用OK・改変OK・再配布OKなので、じゃんじゃん使ってね！

⚙gear machine@AI@grmchn4ai

マスコットに人物と同じ動きをさせる「mascot pose generator」を期間限定公開します！(7/1 23:59まで) #ローカルAIに向き合う展示会で出そうと思ったのですが家のGPUサーバーが落ちたので…。リボン猫や電気ネズミのような頭身のキャラを動かしてみて遊んでみてね！ mascot-pose-generator.tanosix.com

日本語

2

30

233

21K

ka chun cheng@creator_kachun·30 Haz

@PurzBeats We need your walkthrough!

English

1

0

1

45

Purz.ai@PurzBeats·29 Haz

MCP is live, go touch it. 👈

ComfyUI@ComfyUI

We just launched Comfy MCP in public beta, the first MCP built for production pipelines. Connect Claude, Codex, Cursor, or Hermes to the entire ComfyUI ecosystem. → Run any workflow in natural language → Search models, nodes, and templates → Share a workflow URL - your agent picks it up → Re-run saved workflows on new inputs without touching a single node → Hundreds of pre-built workflows. Always auto-updated 20 product-placement images for your brand ads. Character design for your game art. Script-to-shot ideation at scale and so on! Your agent is now a creative technologist.

English

12

7

145

12.8K

ka chun cheng@creator_kachun·29 Haz

My First @krea_ai Lora training for my favorite cheongsam.

English

0

23

ka chun cheng retweetledi

Tongyi Lab@Ali_TongyiLab·26 Haz

x.com/i/article/2070…

ZXX

3

10

105

1.9M

ka chun cheng@creator_kachun·25 Haz

Really love @krea_ai 2. Testing Krea2TextEncoder from Ethanfel.

English

0

16

ka chun cheng retweetledi

NYRE@sleenyre·23 Haz

Spend a lot of time writing this over the past week. If you want juicy details on our training stack and infra. Go read it! krea.ai/blog/krea-2-te…

English

9

30

167

21.2K

ka chun cheng@creator_kachun·24 Haz

@linoy_tsaban Can you share your prompt describing the reference sheet? I have been trying long and short prompt for reference sheet, adjust Lora weight to 1.5, disable the repeat Image batch node, the result still very bad.

English

0

257

Linoy Tsaban@linoy_tsaban·24 Haz

LTX 2.3 ingredients LoRA is so good! My take playing with this example (distilled, 8-step inference): > reference sheet doesnt have to be perfect, but adding a close up of Wednesday's face helped a lot with identity preservation > elaborate prompts make a big difference, being as specific as possible

English

3

10

76

9.5K

ka chun cheng@creator_kachun·18 Haz

Black Rainstorm Warning was issued by the Hong Kong Observatory at 12:55 pm. Staff and students are advised to stay in safety place for their own safety. @klingai_official 3.0 Turbo

English

0

37

LTX.io@ltx_io·17 Haz

Most AI video tools stop at prompting, but LTX goes a layer deeper. Introducing the new LTX Trainer. Train LoRAs and IC-LoRAs across video, audio, cross-modal, and reference-conditioned workflows from a single framework. Plus: • New agentic skill • Flexible conditioning • Free IC-LoRAs • Fully open source Most models are generic. The best ones become yours. GitHub: github.com/Lightricks/LTX… Documentation: docs.ltx.video/open-source-mo… HuggingFace: huggingface.co/collections/Li…

English

53

120

1K

6M

ka chun cheng@creator_kachun·17 Haz

@ltx_io Thanks ! Hope there will be more guides for non technical artists!

English

0

2

663

ka chun cheng@creator_kachun·12 Haz

@ostrisai @kohya_tech Please!

English

0

40

Ostris@ostrisai·11 Haz

I prototyped out an idea. Want to gauge interest. It is a live reward trainer. Images are generated and you pick you favorite variant and your preference is fed straight into GRPO. It is a very involved process, but should adjust a model to your personal preference.