[Sammelthread] AI - Bildgenerierung (Stable Diffusion, Midjourney & Co)

[Keuleman]

Dann hatte ich 30 Sekunden schieres Glück und die Sequenz ist perfekt geworden, hah.

 

[MechanimaL]

Zwischenzeitlich ist auch Seedance 2.5 rausgekommen, für alle, die Geld in die Hand nehmen wollen, um etwas zu realisieren, statt lokal zu arbeiten, wohl die aktuell beste Qualität (inkl. der Möglichkeit mit Referenzen zu arbeiten):

YouTube

An dieser Stelle steht ein externer Inhalt von YouTube, der den Forumbeitrag ergänzt. Er kann mit einem Klick geladen und auch wieder ausgeblendet werden.

YouTube-Embeds laden

Ich bin damit einverstanden, dass YouTube-Embeds geladen werden. Dabei können personen­bezogene Daten an YouTube übermittelt werden. Mehr dazu in der Datenschutzerklärung.

YouTube-Embeds laden

Datenschutzerklärung

 

[WoisthierdasKlo]

Habt ihr Glücklichen mit der 5090 schon mal TurboDiffusion ausprobiert? Damit könnt ihr faster-than-realtime KI-Videos produzieren... https://github.com/thu-ml/TurboDiffusion?utm_source=chatgpt.com
Da könnte man bestimmt spannende lustige Projekte mit bauen. Irgendein lokales LLM, das von Whisper o.ä. getragenen Transkriptionsprogramm live gesprochenes Wort in einen Prompt umwandelt und daraus direkt in Echtzeit ein Video macht.
Auch interessant: https://arxiv.org/html/2606.09150v1
Dort haben Forscher 10-30 fps Video in Echtzeit produzieren können. Okay, sie hatten auch eine B200.

 

[samuelclemens]

Hmm... im Prinzip könnte man damit also jetzt schon Videoübertragungen in Echtzeit manipulieren!🤔

Wenn die Forschung nur genauso schnell vorangetrieben würde an der KI Echtzeiterkennung für den lokalen gebrauch.

 

[blubberbirne]

Bin heute abend auf einen guten Systemprompt für Gemini/Claude/ChatGPT gestoßen. Habe mir daraus direkt in geminis Canvas eine kleine App gepromptet. Erste Tests laufen damit gerade.

 

[WoisthierdasKlo]

Hmm... im Prinzip könnte man damit also jetzt schon Videoübertragungen in Echtzeit manipulieren!🤔

Woher willst du wissen, ob das nicht schon längst passiert ist? Und woher wollen wir wissen, dass die groß angekündigten GPUs technologisch in irgendwelchen Hinterzimmern nicht schon längst genutzt werden? Okay, sorry, ich bremse mich lieber selbst, bevor ich einen strike kriege.

Wenn die Forschung nur genauso schnell vorangetrieben würde an der KI Echtzeiterkennung für den lokalen gebrauch.

Ja, das wird aber bestimmt bald kommen. Da denke ich irgendwie sehr optimistisch. Wenn die Vorhersagen (z.T. geradezu Prophezeiungen) mit "2029" und der "Singularität" zutreffen, werden wir in 3 Jahren eh alle mit offline-ChatGPT in der Hosentasche rumlaufen. (Okay, stimmt nicht ganz, das mit der Singularität bezieht sich ja nicht auf Consumer-Hardware....) Auf YouTube bauen sich Leute schon für "wenig" Geld 96 GB VRAM-Racks:

YouTube

An dieser Stelle steht ein externer Inhalt von YouTube, der den Forumbeitrag ergänzt. Er kann mit einem Klick geladen und auch wieder ausgeblendet werden.

YouTube-Embeds laden

Ich bin damit einverstanden, dass YouTube-Embeds geladen werden. Dabei können personen­bezogene Daten an YouTube übermittelt werden. Mehr dazu in der Datenschutzerklärung.

YouTube-Embeds laden

Datenschutzerklärung

 

[samuelclemens]

Wird wohl eher darauf hinauslaufen das man das bei großen Anbieter KI Erkennung als Zusatzdienst so wie VPN, Malware oder Werbefilter usw als Abodienst buchen kann.
Vielleicht auch schon bei den hostern vorgeschaltet als freiwillige Option.
Es wäre ja auch in deren Interesse. KI Slop verhählt sich inzwischen ja wie unerwünschter Spammüll.


Gibt es eigentlich in ComfyUI ne möglichkeit das letzte frame eines Videos automatisch als referenz für das neu zu generierende einzubinden?
Oder sogar mehrere:
Also quasi iwie die letzte Sekunde eines Videos als Referenz um einen nahtlosen Anschluss zu ermöglichen?
Egal was ich anstelle. Wenn ich das manuell mache gibt es immer Anschlussfehler. Und wenn es nur ein kurzer ruckler ist.

 

[Keuleman]

Jungejunge, H3 feiert einen ab... Southpark, Futurama, Simpsons... paar Sachen getestet, alles kam "richtig" raus. Aber jetzt wieder hauptsächlich für meinen Film weiter machen ;-D

 

[samuelclemens]

Kann mir mal einer für doofe erklären wie ich genau einen weiteren Eingang in den Subgraph Node lege.
Wenn ich zb von außen einen Load Image Node einbauen will.
Ich komm einfach nicht dahinter wie ich die leiste links im Subgraph editiert bekomme.
Ich soll laut ChatGPT diese zwei Eingänge einfügen aber sagt mir nicht genau wie und andere Tutorials reden immer nur vom Edit Mode.

reference_image
use_reference

 

[Keuleman]

Hot news: LTX 2.5 ist grad raus gekommen :-D

Vor 2 Tagen wollte ein Comfy LTX Node nicht mehr... musste anderen VAE Decoder nehmen. War wohl ein bekanntes Problem. Da schon vermutet: "Pass auf Keuleman es kommt ne neue LTX Version" :-D

 

[WoisthierdasKlo]

LTX 2.5

Ich lad's mir gleich mal runter, bin aber etwas desillusioniert, dass ich mit meiner 2060er Generation 8 GB VRAM GraKa viel reißen kann. Bei mir reichte es nach zahlreichen Tests in H3 immer nur für maximal 5 Sekunden Clips bei 0.4 MP. Auflösungen darunter zerreißen die Optik und Längen darüber geben OOM (egal welche Auflösung). Zumal das, wenn es frisch rauskommt, bestimmt noch das Rohmodell ist?

 

[blubberbirne]

War doch klar das von LTX auch was neues kommt. Flux 3 kommt auch bald als Open Weights Model.
Da kommen spannende Wochen auf einen zu.

Ich teste aber erstmal von Meta das Muse-Glimmer Sprachmodell mit Claude Code. Bin gespannt ob es mir ähnlich wie gemini auch eine minimax-H3 app bauen kann.

 

[Keuleman]

Wird die App portabel?

 

[blubberbirne]

Bin dabei eine Gradio app zu vibe coden. Allerdings spinnt mein Claude gerade rum. Es hängt im Plan Modus fest

 

[Ozmog]

Kann mir mal einer für doofe erklären wie ich genau einen weiteren Eingang in den Subgraph Node lege.
Wenn ich zb von außen einen Load Image Node einbauen will.

Nicht einfach von der nächsten Node auf den unteren grauen Anschluss legen? Wenn der belegt ist, dürfte auch ein neuer leerer Anschluss erscheinen, wenn ich mich richtig erinnere

 

[samuelclemens]

@Ozmog Aber wie wenn der Load Image node außerhalb des Subgraph liegt?
Ich arbeite eine ChatGPT Anleitung ab wie ich ein Referenzbild in ein Text to Image Workflow als zusätzliche Option einbauen kann und da kommen die Anpassungen im Subgraph eben danach vor nachdem der Input Eingang gelegt ist.

 

[Ozmog]

Probiere es doch mal vom ersten Node im Subgraph aus. Wäre nicht das erste mal, das KI falsch liegt, hatte ich heute bei der Arbeit schon genug Fehler von KI, die auch nicht ganz richtig lagen.

Mit Subhraph habe ich selbst noch nicht gearbeitet, daher kann ich nicht sagen, ob es wirklich funktioniert, erinnert mich aber an diverse andere Funktionen.

 

[samuelclemens]

@Ozmog Klar, wie man den eingang von innen her verlegt hab ich auch schon begriffen.

Leider schnall ich auch den gesammten Rest der Anleitung immer weniger.
Sind halt Grundlagen die man braucht wenn man mit Comfy weiterkommen will.
Aber wenn es schon an der Bedienoberfläche scheitert! Ganz zu schweigen das ich kein plan habe was die meisten der Nodes überhaupt genau tun. 🤔

Im Prinzip versuche ich was ganz simples das offizielle Krea2 Template "Text to Image" ein zusätzliches Referenzbild Node hinzuzufügen.
Vielleicht komm ich eher dahinter wenn mir das mal einer ordentlich so einbaut damit ich es durch vergleich studieren kann. Den Optionalen Schalter von außen um das Referenzbild zu aktivieren kann von mir aus auch erstmal weggelasssen werden aus übersicht gründen.

{"id":"584e26a1-36b7-40fe-a47f-6728b4edb0ff","revision":0,"last_node_id":50,"last_link_id":86,"nodes":[{"id":29,"type":"SaveImage","pos":[480,-110],"size":[880,1020],"flags":{},"order":5,"mode":0,"inputs":[{"localized_name":"images","name":"images","type":"IMAGE","link":44},{"localized_name":"filename_prefix","name":"filename_prefix","type":"STRING","widget":{"name":"filename_prefix"},"link":null}],"outputs":[{"localized_name":"images","name":"images","type":"IMAGE","links":null}],"properties":{"Node name for S&R":"SaveImage","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["Krea2_turbo"]},{"id":30,"type":"b0e5ca93-2731-42b9-8e0a-d28ea851ff81","pos":[-140,-110],"size":[560,1010],"flags":{},"order":4,"mode":0,"inputs":[{"name":"value","type":"STRING","widget":{"name":"value"},"link":null},{"label":"prompt_enhance","name":"value_1","type":"BOOLEAN","widget":{"name":"value_1"},"link":null},{"label":"LLM_thinking_mode","name":"thinking","shape":7,"type":"BOOLEAN","widget":{"name":"thinking"},"link":null},{"label":"LLM_max_token","name":"max_length_1","type":"INT","widget":{"name":"max_length_1"},"link":null},{"label":"width","name":"width_1","type":"INT","widget":{"name":"width_1"},"link":85},{"label":"height","name":"height_1","type":"INT","widget":{"name":"height_1"},"link":86},{"label":"seed","name":"seed_1","type":"INT","widget":{"name":"seed_1"},"link":null},{"label":"enable_lora","name":"value_3","type":"BOOLEAN","widget":{"name":"value_3"},"link":null},{"label":"lora_name","name":"lora_name_1","type":"COMBO","widget":{"name":"lora_name_1"},"link":null},{"label":"strength_model","name":"strength_model_1","type":"FLOAT","widget":{"name":"strength_model_1"},"link":null},{"label":"lora_trigger_word","name":"string_b_1","type":"STRING","widget":{"name":"string_b_1"},"link":null},{"name":"unet_name","type":"COMBO","widget":{"name":"unet_name"},"link":null},{"name":"clip_name","type":"COMBO","widget":{"name":"clip_name"},"link":null},{"name":"vae_name","type":"COMBO","widget":{"name":"vae_name"},"link":null}],"outputs":[{"localized_name":"IMAGE","name":"IMAGE","type":"IMAGE","links":[44]}],"properties":{"cnr_id":"comfy-core","ver":"0.25.0","previewExposures":[{"name":"$$canvas-image-preview","sourceNodeId":"3","sourcePreviewName":"$$canvas-image-preview"}],"ue_properties":{"widget_ue_connectable":{"value_1":true,"thinking":true,"max_length_1":true,"width_1":true,"height_1":true,"seed_1":true,"value_3":true,"lora_name_1":true,"strength_model_1":true,"string_b_1":true},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["A high-resolution, surreal digital illustration showing a human hand holding a martini glass. The image is overlaid with whimsical, expressive ink-style doodles, including a cartoon figure inside the glass, a drawn citrus wedge on the rim, and various abstract sketches and faces surrounding the glass against a clean, white background. The style seamlessly blends a realistic, lit photograph with loose, hand-drawn marker artistry, creating a playful and artistic juxtaposition.",true,false,512,1024,1024,594361197674106,false,"krea2_darkbrush.safetensors",0.8,"muted minimalist sketch style","krea2_turbo_fp8_scaled.safetensors","qwen3vl_4b_fp8_scaled.safetensors","qwen_image_vae.safetensors"]},{"id":47,"type":"MarkdownNote","pos":[-1200,970],"size":[690,510],"flags":{},"order":0,"mode":0,"inputs":[],"outputs":[],"title":"LoRA Trigger Words and Settings","properties":{"ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["Download the style LoRAs used in this workflow from [Comfy-Org/Krea-2/loras](https://huggingface.co/Comfy-Org/Krea-2/tree/main/loras). Place the `.safetensors` files in `ComfyUI/models/loras/`.\n\n| LoRA                   | Trigger Words                        | Recommended Strength |\n|------------------------|--------------------------------------|-------|\n| `krea2_darkbrush`      | `monochrome ink wash style`          | `1.0` |\n| `krea2_dotmatrix`      | `monochrome stippling style`         | `1.0` |\n| `krea2_kidsdrawing`    | `naive expressive sketch style`      | `1.0` |\n| `krea2_neondrip`       | `textured abstract style`            | `1.0` |\n| `krea2_rainywindow`    | `rainy window style`                 | `1.0` |\n| `krea2_retroanime`     | `purple retro anime style`           | `1.0` |\n| `krea2_softwatercolor` | `art deco watercolor style`          | `1.0` |\n| `krea2_sunsetblur`     | `ethereal motion blur style`         | `1.0` |\n| `krea2_vintagetarot`   | `vintage tarot style`                | `1.0` |"],"color":"#222","bgcolor":"#000"},{"id":48,"type":"MarkdownNote","pos":[-1940,-90],"size":[700,990],"flags":{},"order":1,"mode":0,"inputs":[],"outputs":[],"title":"Model Links","properties":{"ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["Guide: [Subgraph](https://docs.comfy.org/interface/features/subgraph)\n\n## Model Links\n\nFor BF16 and NVFP4 variants, see [Comfy-Org/Krea-2](https://huggingface.co/Comfy-Org/Krea-2)\n\n**diffusion_models**\n\n- [krea2_turbo_fp8_scaled.safetensors](https://huggingface.co/Comfy-Org/Krea-2/resolve/main/diffusion_models/krea2_turbo_fp8_scaled.safetensors)\n\n**text_encoders**\n\n- [qwen3vl_4b_fp8_scaled.safetensors](https://huggingface.co/Comfy-Org/Krea-2/resolve/main/text_encoders/qwen3vl_4b_fp8_scaled.safetensors)\n\n**vae**\n\n- [qwen_image_vae.safetensors](https://huggingface.co/Comfy-Org/Krea-2/resolve/main/vae/qwen_image_vae.safetensors)\n\n**loras**\n\n- [krea2_darkbrush.safetensors](https://huggingface.co/Comfy-Org/Krea-2/resolve/main/loras/krea2_darkbrush.safetensors)\n\n\n## Model Storage Location\n\n```\n📂 ComfyUI/\n├── 📂 models/\n│   ├── 📂 diffusion_models/\n│   │   └── krea2_turbo_fp8_scaled.safetensors\n│   ├── 📂 text_encoders/\n│   │   └── qwen3vl_4b_fp8_scaled.safetensors\n│   ├── 📂 vae/\n│   │   └── qwen_image_vae.safetensors\n│   └── 📂 loras/\n│       └── krea2_darkbrush.safetensors\n```\n\n## Report Issue\n\nNote: Please update ComfyUI first ([guide](https://docs.comfy.org/installation/update_comfyui)) and prepare required models. Desktop/Cloud updates follow stable releases, so some nightly-supported models may not be available yet.\n\n- Cannot run / runtime errors: [ComfyUI/issues](https://github.com/comfyanonymous/ComfyUI/issues)\n- UI / frontend issues: [ComfyUI_frontend/issues](https://github.com/Comfy-Org/ComfyUI_frontend/issues)\n- Workflow issues: [workflow_templates/issues](https://github.com/Comfy-Org/workflow_templates/issues)\n"],"color":"#222","bgcolor":"#000"},{"id":50,"type":"MarkdownNote","pos":[-1200,-90],"size":[700,990],"flags":{},"order":2,"mode":0,"inputs":[],"outputs":[],"title":"Note: Krea-2","properties":{"ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["Guide: [krea-2](https://docs.comfy.org/tutorials/image/krea/krea-2)\n\nThis workflow generates images with **Krea-2 Turbo**, a fast distilled text-to-image model. The workflow is organized into a few parts: each handles a different step so you can generate quickly without digging through every node.\n\n## Text to Image (Krea-2 Turbo) Subgraph\n\nDouble-click this node to open the full pipeline. Inside, the canvas is grouped into several areas:\n\n**Models**: Loads the diffusion model, text encoder, VAE, and optional style LoRA (`LoraLoaderModelOnly`).\n\n**Prompt**: Enter your image description in **Text String (User Prompt)**.\n\n**Prompt Enhancement**: Expands your prompt before generation. Enabled by default (`prompt_enhance`). You can turn it off to use your prompt as-is, or replace the built-in module with **OpenAI** or **Gemini** API nodes.\n\n**Feature Switch**: Quick toggles on the left side of the subgraph:\n- `prompt_enhance`: enable or disable prompt expansion\n- `enable_lora?`: enable or disable the style LoRA\n\n**Image Generation**: Handles latent sizing, sampling (8 steps), and VAE decode to produce the final image.\n\nWhen using a LoRA, set `enable_lora?` to true, select the matching file in **LoraLoaderModelOnly**, and adjust `lora_strength` as needed.\n\n## LoRA & Trigger Word (CustomCombo)\n\nOn the main canvas, the **CustomCombo** node is where you choose which style LoRA you are using (e.g. `krea2_coolblue`, `krea2_darkbrush`). The workflow uses your selection to automatically append the correct **trigger word** to your prompt.\n\n- Download LoRA files from [Comfy-Org/Krea-2/loras](https://huggingface.co/Comfy-Org/Krea-2/tree/main/loras) and place them in `ComfyUI/models/loras/`\n- Select the LoRA name that matches the file loaded inside the subgraph\n- The trigger word is applied for you: no need to type it manually\n- For trigger words and recommended strength values, see the **LoRA Trigger Words and Settings** note beside this workflow\n"],"color":"#222","bgcolor":"#000"},{"id":49,"type":"ResolutionSelector","pos":[-470,270],"size":[270,190],"flags":{},"order":3,"mode":0,"inputs":[{"localized_name":"aspect_ratio","name":"aspect_ratio","type":"COMBO","widget":{"name":"aspect_ratio"},"link":null},{"localized_name":"megapixels","name":"megapixels","type":"FLOAT","widget":{"name":"megapixels"},"link":null},{"localized_name":"multiple","name":"multiple","type":"INT","widget":{"name":"multiple"},"link":null}],"outputs":[{"localized_name":"width","name":"width","type":"INT","links":[85]},{"localized_name":"height","name":"height","type":"INT","links":[86]}],"properties":{"Node name for S&R":"ResolutionSelector","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["1:1 (Square)",1,8]}],"links":[[44,30,0,29,0,"IMAGE"],[85,49,0,30,4,"INT"],[86,49,1,30,5,"INT"]],"groups":[],"definitions":{"subgraphs":[{"id":"b0e5ca93-2731-42b9-8e0a-d28ea851ff81","version":1,"state":{"lastGroupId":10,"lastNodeId":50,"lastLinkId":86,"lastRerouteId":0},"revision":0,"config":{},"name":"Text to Image (Krea-2 Turbo)","inputNode":{"id":-10,"bounding":[-1663.0467311105908,442.6713226822096,160.76667022705078,328]},"outputNode":{"id":-20,"bounding":[1560,496,128,68]},"inputs":[{"id":"b0f5c754-5322-440c-8c1c-ba732e910d33","name":"value","type":"STRING","linkIds":[45],"pos":[-1526.28006088354,466.6713226822096]},{"id":"47b0e74d-ef80-4423-a7a6-9bd1673092bb","name":"value_1","type":"BOOLEAN","linkIds":[46],"label":"prompt_enhance","pos":[-1526.28006088354,486.6713226822096]},{"id":"bece097f-5c57-4aeb-9113-16499634a69c","name":"thinking","type":"BOOLEAN","linkIds":[73],"label":"LLM_thinking_mode","pos":[-1526.28006088354,506.6713226822096]},{"id":"9b4edd63-9421-47d9-8c41-5a44510fcba9","name":"max_length_1","type":"INT","linkIds":[74],"label":"LLM_max_token","pos":[-1526.28006088354,526.6713226822096]},{"id":"2a2c225e-95ca-4c45-ab19-5072a890b064","name":"width_1","type":"INT","linkIds":[75],"label":"width","pos":[-1526.28006088354,546.6713226822096]},{"id":"acdd37ed-8b29-4b6d-8d6a-51aacaaefc73","name":"height_1","type":"INT","linkIds":[76],"label":"height","pos":[-1526.28006088354,566.6713226822096]},{"id":"08bcb3eb-ae87-41cc-b695-d1f934f05e35","name":"seed_1","type":"INT","linkIds":[77],"label":"seed","pos":[-1526.28006088354,586.6713226822096]},{"id":"c7f9b617-8ef0-4b9e-944d-b85f01c7a9be","name":"value_3","type":"BOOLEAN","linkIds":[78],"label":"enable_lora","pos":[-1526.28006088354,606.6713226822096]},{"id":"04ec0239-3e74-4a97-90b8-74e881e22e7e","name":"lora_name_1","type":"COMBO","linkIds":[79],"label":"lora_name","pos":[-1526.28006088354,626.6713226822096]},{"id":"61e23017-849c-4879-bbef-3801e2b75198","name":"strength_model_1","type":"FLOAT","linkIds":[80],"label":"strength_model","pos":[-1526.28006088354,646.6713226822096]},{"id":"9396dc65-64f2-4fb2-84d7-4a8fa853bd28","name":"string_b_1","type":"STRING","linkIds":[81],"label":"lora_trigger_word","pos":[-1526.28006088354,666.6713226822096]},{"id":"f4e9f716-afa5-48ca-a314-64f4659b0c99","name":"unet_name","type":"COMBO","linkIds":[82],"pos":[-1526.28006088354,686.6713226822096]},{"id":"add6cb96-21da-41c2-9ef9-60dfe1cb7535","name":"clip_name","type":"COMBO","linkIds":[83],"pos":[-1526.28006088354,706.6713226822096]},{"id":"81640110-5a4d-45c0-9d7c-59966875bf0f","name":"vae_name","type":"COMBO","linkIds":[84],"pos":[-1526.28006088354,726.6713226822096]}],"outputs":[{"id":"e4073813-752d-4584-96b5-6e42eac355dd","name":"IMAGE","type":"IMAGE","linkIds":[43],"localized_name":"IMAGE","pos":[1584,520]}],"widgets":[],"nodes":[{"id":6,"type":"CLIPTextEncode","pos":[100,-250],"size":[480,420],"flags":{},"order":3,"mode":0,"inputs":[{"localized_name":"clip","name":"clip","type":"CLIP","link":10},{"localized_name":"text","name":"text","type":"STRING","widget":{"name":"text"},"link":72}],"outputs":[{"localized_name":"CONDITIONING","name":"CONDITIONING","type":"CONDITIONING","slot_index":0,"links":[4,13]}],"properties":{"Node name for S&R":"CLIPTextEncode","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["A high-resolution, surreal digital illustration showing a human hand holding a martini glass. The image is overlaid with whimsical, expressive ink-style doodles, including a cartoon figure inside the glass, a drawn citrus wedge on the rim, and various abstract sketches and faces surrounding the glass against a clean, white background. The style seamlessly blends a realistic, lit photograph with loose, hand-drawn marker artistry, creating a playful and artistic juxtaposition."],"color":"#232","bgcolor":"#353"},{"id":5,"type":"EmptyLatentImage","pos":[170,350],"size":[320,170],"flags":{},"order":2,"mode":0,"inputs":[{"localized_name":"width","name":"width","type":"INT","widget":{"name":"width"},"link":75},{"localized_name":"height","name":"height","type":"INT","widget":{"name":"height"},"link":76},{"localized_name":"batch_size","name":"batch_size","type":"INT","widget":{"name":"batch_size"},"link":null}],"outputs":[{"localized_name":"LATENT","name":"LATENT","type":"LATENT","slot_index":0,"links":[2]}],"properties":{"Node name for S&R":"EmptyLatentImage","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":[1024,1024,1]},{"id":3,"type":"KSampler","pos":[640,-100],"size":[320,590],"flags":{},"order":1,"mode":0,"inputs":[{"localized_name":"model","name":"model","type":"MODEL","link":30},{"localized_name":"positive","name":"positive","type":"CONDITIONING","link":4},{"localized_name":"negative","name":"negative","type":"CONDITIONING","link":14},{"localized_name":"latent_image","name":"latent_image","type":"LATENT","link":2},{"localized_name":"seed","name":"seed","type":"INT","widget":{"name":"seed"},"link":77},{"localized_name":"steps","name":"steps","type":"INT","widget":{"name":"steps"},"link":null},{"localized_name":"cfg","name":"cfg","type":"FLOAT","widget":{"name":"cfg"},"link":null},{"localized_name":"sampler_name","name":"sampler_name","type":"COMBO","widget":{"name":"sampler_name"},"link":null},{"localized_name":"scheduler","name":"scheduler","type":"COMBO","widget":{"name":"scheduler"},"link":null},{"localized_name":"denoise","name":"denoise","type":"FLOAT","widget":{"name":"denoise"},"link":null}],"outputs":[{"localized_name":"LATENT","name":"LATENT","type":"LATENT","slot_index":0,"links":[7]}],"properties":{"Node name for S&R":"KSampler","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":[735915477938686,"randomize",8,1,"euler","simple",1]},{"id":8,"type":"VAEDecode","pos":[1010,-270],"size":[230,100],"flags":{"collapsed":false},"order":4,"mode":0,"inputs":[{"localized_name":"samples","name":"samples","type":"LATENT","link":7},{"localized_name":"vae","name":"vae","type":"VAE","link":12}],"outputs":[{"localized_name":"IMAGE","name":"IMAGE","type":"IMAGE","slot_index":0,"links":[43]}],"properties":{"Node name for S&R":"VAEDecode","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":[]},{"id":10,"type":"UNETLoader","pos":[-410,-240],"size":[450,140],"flags":{},"order":5,"mode":0,"inputs":[{"localized_name":"unet_name","name":"unet_name","type":"COMBO","widget":{"name":"unet_name"},"link":82},{"localized_name":"weight_dtype","name":"weight_dtype","type":"COMBO","widget":{"name":"weight_dtype"},"link":null}],"outputs":[{"localized_name":"MODEL","name":"MODEL","type":"MODEL","links":[16,28]}],"properties":{"Node name for S&R":"UNETLoader","cnr_id":"comfy-core","ver":"0.25.0","models":[{"name":"krea2_turbo_fp8_scaled.safetensors","url":"https://huggingface.co/Comfy-Org/Krea-2/resolve/main/diffusion_models/krea2_turbo_fp8_scaled.safetensors","directory":"diffusion_models"}],"ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["krea2_turbo_fp8_scaled.safetensors","default"]},{"id":11,"type":"CLIPLoader","pos":[-410,150],"size":[450,170],"flags":{},"order":6,"mode":0,"inputs":[{"localized_name":"clip_name","name":"clip_name","type":"COMBO","widget":{"name":"clip_name"},"link":83},{"localized_name":"type","name":"type","type":"COMBO","widget":{"name":"type"},"link":null},{"localized_name":"device","name":"device","shape":7,"type":"COMBO","widget":{"name":"device"},"link":null}],"outputs":[{"localized_name":"CLIP","name":"CLIP","type":"CLIP","links":[10,18]}],"properties":{"Node name for S&R":"CLIPLoader","cnr_id":"comfy-core","ver":"0.25.0","models":[{"name":"qwen3vl_4b_fp8_scaled.safetensors","url":"https://huggingface.co/Comfy-Org/Krea-2/resolve/main/text_encoders/qwen3vl_4b_fp8_scaled.safetensors","directory":"text_encoders"}],"ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["qwen3vl_4b_fp8_scaled.safetensors","krea2","default"]},{"id":12,"type":"VAELoader","pos":[-410,390],"size":[450,120],"flags":{},"order":7,"mode":0,"inputs":[{"localized_name":"vae_name","name":"vae_name","type":"COMBO","widget":{"name":"vae_name"},"link":84}],"outputs":[{"localized_name":"VAE","name":"VAE","type":"VAE","links":[12]}],"properties":{"Node name for S&R":"VAELoader","cnr_id":"comfy-core","ver":"0.25.0","models":[{"name":"qwen_image_vae.safetensors","url":"https://huggingface.co/Comfy-Org/Krea-2/resolve/main/vae/qwen_image_vae.safetensors","directory":"vae"}],"ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["qwen_image_vae.safetensors"]},{"id":13,"type":"ConditioningZeroOut","pos":[350,220],"size":[225,26],"flags":{"collapsed":true},"order":8,"mode":0,"inputs":[{"localized_name":"conditioning","name":"conditioning","type":"CONDITIONING","link":13}],"outputs":[{"localized_name":"CONDITIONING","name":"CONDITIONING","type":"CONDITIONING","links":[14]}],"properties":{"Node name for S&R":"ConditioningZeroOut","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":[]},{"id":15,"type":"LoraLoaderModelOnly","pos":[-410,-50],"size":[450,170],"flags":{},"order":9,"mode":0,"inputs":[{"localized_name":"model","name":"model","type":"MODEL","link":16},{"localized_name":"lora_name","name":"lora_name","type":"COMBO","widget":{"name":"lora_name"},"link":79},{"localized_name":"strength_model","name":"strength_model","type":"FLOAT","widget":{"name":"strength_model"},"link":80}],"outputs":[{"localized_name":"MODEL","name":"MODEL","type":"MODEL","links":[29]}],"properties":{"Node name for S&R":"LoraLoaderModelOnly","cnr_id":"comfy-core","ver":"0.25.0","models":[{"name":"krea2_darkbrush.safetensors","url":"https://huggingface.co/Comfy-Org/Krea-2/resolve/main/loras/krea2_darkbrush.safetensors","directory":"loras"}],"ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["krea2_darkbrush.safetensors",0.8]},{"id":16,"type":"TextGenerate","pos":[60,650],"size":[400,412],"flags":{},"order":10,"mode":0,"inputs":[{"localized_name":"clip","name":"clip","type":"CLIP","link":18},{"localized_name":"image","name":"image","shape":7,"type":"IMAGE","link":null},{"localized_name":"video","name":"video","shape":7,"type":"IMAGE","link":null},{"localized_name":"audio","name":"audio","shape":7,"type":"AUDIO","link":null},{"localized_name":"prompt","name":"prompt","type":"STRING","widget":{"name":"prompt"},"link":19},{"localized_name":"max_length","name":"max_length","type":"INT","widget":{"name":"max_length"},"link":74},{"localized_name":"sampling_mode","name":"sampling_mode","type":"COMFY_DYNAMICCOMBO_V3","widget":{"name":"sampling_mode"},"link":null},{"localized_name":"temperature","name":"sampling_mode.temperature","type":"FLOAT","widget":{"name":"sampling_mode.temperature"},"link":null},{"localized_name":"top_k","name":"sampling_mode.top_k","type":"INT","widget":{"name":"sampling_mode.top_k"},"link":null},{"localized_name":"top_p","name":"sampling_mode.top_p","type":"FLOAT","widget":{"name":"sampling_mode.top_p"},"link":null},{"localized_name":"min_p","name":"sampling_mode.min_p","type":"FLOAT","widget":{"name":"sampling_mode.min_p"},"link":null},{"localized_name":"repetition_penalty","name":"sampling_mode.repetition_penalty","type":"FLOAT","widget":{"name":"sampling_mode.repetition_penalty"},"link":null},{"localized_name":"seed","name":"sampling_mode.seed","type":"INT","widget":{"name":"sampling_mode.seed"},"link":null},{"localized_name":"presence_penalty","name":"sampling_mode.presence_penalty","shape":7,"type":"FLOAT","widget":{"name":"sampling_mode.presence_penalty"},"link":null},{"localized_name":"thinking","name":"thinking","shape":7,"type":"BOOLEAN","widget":{"name":"thinking"},"link":73},{"localized_name":"use_default_template","name":"use_default_template","shape":7,"type":"BOOLEAN","widget":{"name":"use_default_template"},"link":null}],"outputs":[{"localized_name":"generated_text","name":"generated_text","type":"STRING","links":[33]}],"properties":{"Node name for S&R":"TextGenerate","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["",512,"on",0.7,64,0.95,0.05,1.05,0,0,true,true]},{"id":17,"type":"StringConcatenate","pos":[-210,1310],"size":[225,34],"flags":{"collapsed":true},"order":11,"mode":0,"inputs":[{"localized_name":"string_a","name":"string_a","type":"STRING","widget":{"name":"string_a"},"link":21},{"localized_name":"string_b","name":"string_b","type":"STRING","widget":{"name":"string_b"},"link":22},{"localized_name":"delimiter","name":"delimiter","type":"STRING","widget":{"name":"delimiter"},"link":null}],"outputs":[{"localized_name":"STRING","name":"STRING","type":"STRING","links":[19]}],"properties":{"Node name for S&R":"StringConcatenate","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["","",""]},{"id":18,"type":"PrimitiveStringMultiline","pos":[-430,650],"size":[460,610],"flags":{},"order":0,"mode":0,"inputs":[{"localized_name":"value","name":"value","type":"STRING","widget":{"name":"value"},"link":null}],"outputs":[{"localized_name":"STRING","name":"STRING","type":"STRING","links":[21]}],"title":"Text String (System Prompt)","properties":{"Node name for S&R":"PrimitiveStringMultiline","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["You are an expert prompt engineer for text-to-image models. Your task is to expand the user's prompt into a highly effective image-generation prompt.\n\nThink step by step about the request before writing the answer:\n- What is the subject and mood?\n- What visual styles, mediums, and lighting options would fit? Consider two or three alternatives and pick the one that best serves the caption.\n- What composition, framing, and grounded details will help the text-to-image model?\n\nThen output a single expanded prompt paragraph.\n\nFollow these rules strictly:\n1. **Faithfulness First:** Preserve all original subjects, actions, colors, and spatial relationships. Do not add new objects, props, characters, or animals unless the user clearly implies them.\n2. **Practical T2I Structure:** Write a prompt that a text-to-image model can parse cleanly. Group subjects with their own attributes and actions. Use grounded phrasing for poses, interactions, and spatial layout.\n3. **Style Planning Stays Internal:** Use your internal reasoning to choose style, medium, framing, and lighting. Do not emit planning tags or wrappers in the visible answer body.\n4. **Text Rendering:** If the user requests visible text, quotes, labels, or typography, specify the exact text clearly and wrap requested words in quotes.\n5. **Avoid Over-Specification:** Do not invent highly specific clothing, colors, materials, or scene details unless the input supports them.\n6. **Structure:** Write one cohesive paragraph after the thinking block. No bullets, JSON, or markdown.\n7. **Respect Existing Detail:** If the user's prompt is already detailed, lightly polish and finalize rather than heavily expanding — preserve their phrasing and direction.\n8. **Respect the Human Form:** Treat depictions of people with dignity. Assume clothing covers genitals and intimate anatomy.\n9. **Preserve User Medium:** When the user explicitly requests a medium (e.g. \"photo of\", \"photograph of\", \"illustration of\", \"painting of\", \"sketch of\", \"3D render of\"), honor it. Do not pivot to a different medium to avoid difficulty — match the user's stated intent.\n\nUser's Input:\n\n"]},{"id":19,"type":"PrimitiveStringMultiline","pos":[-1070,-290],"size":[570,800],"flags":{},"order":12,"mode":0,"inputs":[{"localized_name":"value","name":"value","type":"STRING","widget":{"name":"value"},"link":45}],"outputs":[{"localized_name":"STRING","name":"STRING","type":"STRING","links":[22,34]}],"title":"Text String (User Prompt)","properties":{"Node name for S&R":"PrimitiveStringMultiline","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["A high-resolution, surreal digital illustration showing a human hand holding a martini glass. The image is overlaid with whimsical, expressive ink-style doodles, including a cartoon figure inside the glass, a drawn citrus wedge on the rim, and various abstract sketches and faces surrounding the glass against a clean, white background. The style seamlessly blends a realistic, lit photograph with loose, hand-drawn marker artistry, creating a playful and artistic juxtaposition."],"color":"#322","bgcolor":"#533"},{"id":20,"type":"PreviewAny","pos":[500,660],"size":[470,550],"flags":{"collapsed":false},"order":13,"mode":0,"inputs":[{"localized_name":"source","name":"source","type":"*","link":27}],"outputs":[{"localized_name":"STRING","name":"STRING","type":"STRING","links":[37,41]}],"properties":{"Node name for S&R":"PreviewAny","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":[]},{"id":21,"type":"ComfySwitchNode","pos":[170,1270],"size":[270,130],"flags":{},"order":14,"mode":0,"inputs":[{"localized_name":"on_false","name":"on_false","type":"*","link":34},{"localized_name":"on_true","name":"on_true","type":"*","link":33},{"localized_name":"switch","name":"switch","type":"BOOLEAN","widget":{"name":"switch"},"link":32}],"outputs":[{"localized_name":"output","name":"output","type":"*","links":[27]}],"properties":{"Node name for S&R":"ComfySwitchNode","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":[false]},{"id":22,"type":"ComfySwitchNode","pos":[640,-270],"size":[320,130],"flags":{},"order":15,"mode":0,"inputs":[{"localized_name":"on_false","name":"on_false","type":"*","link":28},{"localized_name":"on_true","name":"on_true","type":"*","link":29},{"localized_name":"switch","name":"switch","type":"BOOLEAN","widget":{"name":"switch"},"link":31}],"outputs":[{"localized_name":"output","name":"output","type":"*","links":[30]}],"title":"Switch (Model)","properties":{"Node name for S&R":"ComfySwitchNode","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":[false]},{"id":23,"type":"PrimitiveBoolean","pos":[-910,670],"size":[270,110],"flags":{},"order":16,"mode":0,"inputs":[{"localized_name":"value","name":"value","type":"BOOLEAN","widget":{"name":"value"},"link":78}],"outputs":[{"localized_name":"BOOLEAN","name":"BOOLEAN","type":"BOOLEAN","links":[31,42]}],"title":"Boolean (Enable LoRA?)","properties":{"Node name for S&R":"PrimitiveBoolean","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":[false]},{"id":24,"type":"PrimitiveBoolean","pos":[-910,840],"size":[270,110],"flags":{},"order":17,"mode":0,"inputs":[{"localized_name":"value","name":"value","type":"BOOLEAN","widget":{"name":"value"},"link":46}],"outputs":[{"localized_name":"BOOLEAN","name":"BOOLEAN","type":"BOOLEAN","links":[32]}],"title":"Boolean (Refine Prompt?)","properties":{"Node name for S&R":"PrimitiveBoolean","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":[true]},{"id":27,"type":"StringConcatenate","pos":[1030,960],"size":[350,250],"flags":{"collapsed":false},"order":18,"mode":0,"inputs":[{"localized_name":"string_a","name":"string_a","type":"STRING","widget":{"name":"string_a"},"link":41},{"localized_name":"string_b","name":"string_b","type":"STRING","widget":{"name":"string_b"},"link":81},{"localized_name":"delimiter","name":"delimiter","type":"STRING","widget":{"name":"delimiter"},"link":null}],"outputs":[{"localized_name":"STRING","name":"STRING","type":"STRING","links":[38]}],"title":"Concatenate Text (LoRA Trigger Word)","properties":{"Node name for S&R":"StringConcatenate","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":["","muted minimalist sketch style",", "]},{"id":28,"type":"ComfySwitchNode","pos":[1050,670],"size":[340,130],"flags":{},"order":19,"mode":0,"inputs":[{"localized_name":"on_false","name":"on_false","type":"*","link":37},{"localized_name":"on_true","name":"on_true","type":"*","link":38},{"localized_name":"switch","name":"switch","type":"BOOLEAN","widget":{"name":"switch"},"link":42}],"outputs":[{"localized_name":"output","name":"output","type":"*","links":[72]}],"properties":{"Node name for S&R":"ComfySwitchNode","cnr_id":"comfy-core","ver":"0.25.0","ue_properties":{"widget_ue_connectable":{},"version":"7.8","input_ue_unconnectable":{}}},"widgets_values":[false]}],"groups":[{"id":1,"title":"Models","bounding":[-430,-320,490,850],"color":"#3f789e","flags":{}},{"id":4,"title":"Prompt Enhancement","bounding":[-440,580,1940,840],"color":"#3f789e","flags":{}},{"id":5,"title":"Group","bounding":[-1080,-360,610,910],"color":"#3f789e","flags":{}},{"id":6,"title":"Image Generation","bounding":[-440,-360,1940,910],"color":"#3f789e","flags":{}},{"id":7,"title":"Prompt","bounding":[90,-320,500,570],"color":"#3f789e","flags":{}},{"id":8,"title":"Size","bounding":[90,270,500,260],"color":"#3f789e","flags":{}},{"id":9,"title":"Feature Switch","bounding":[-1080,580,610,400],"color":"#3f789e","flags":{}}],"links":[{"id":10,"origin_id":11,"origin_slot":0,"target_id":6,"target_slot":0,"type":"CLIP"},{"id":30,"origin_id":22,"origin_slot":0,"target_id":3,"target_slot":0,"type":"MODEL"},{"id":4,"origin_id":6,"origin_slot":0,"target_id":3,"target_slot":1,"type":"CONDITIONING"},{"id":14,"origin_id":13,"origin_slot":0,"target_id":3,"target_slot":2,"type":"CONDITIONING"},{"id":2,"origin_id":5,"origin_slot":0,"target_id":3,"target_slot":3,"type":"LATENT"},{"id":7,"origin_id":3,"origin_slot":0,"target_id":8,"target_slot":0,"type":"LATENT"},{"id":12,"origin_id":12,"origin_slot":0,"target_id":8,"target_slot":1,"type":"VAE"},{"id":13,"origin_id":6,"origin_slot":0,"target_id":13,"target_slot":0,"type":"CONDITIONING"},{"id":16,"origin_id":10,"origin_slot":0,"target_id":15,"target_slot":0,"type":"MODEL"},{"id":18,"origin_id":11,"origin_slot":0,"target_id":16,"target_slot":0,"type":"CLIP"},{"id":19,"origin_id":17,"origin_slot":0,"target_id":16,"target_slot":4,"type":"STRING"},{"id":21,"origin_id":18,"origin_slot":0,"target_id":17,"target_slot":0,"type":"STRING"},{"id":22,"origin_id":19,"origin_slot":0,"target_id":17,"target_slot":1,"type":"STRING"},{"id":27,"origin_id":21,"origin_slot":0,"target_id":20,"target_slot":0,"type":"STRING"},{"id":34,"origin_id":19,"origin_slot":0,"target_id":21,"target_slot":0,"type":"STRING"},{"id":33,"origin_id":16,"origin_slot":0,"target_id":21,"target_slot":1,"type":"STRING"},{"id":32,"origin_id":24,"origin_slot":0,"target_id":21,"target_slot":2,"type":"BOOLEAN"},{"id":28,"origin_id":10,"origin_slot":0,"target_id":22,"target_slot":0,"type":"MODEL"},{"id":29,"origin_id":15,"origin_slot":0,"target_id":22,"target_slot":1,"type":"MODEL"},{"id":31,"origin_id":23,"origin_slot":0,"target_id":22,"target_slot":2,"type":"BOOLEAN"},{"id":41,"origin_id":20,"origin_slot":0,"target_id":27,"target_slot":0,"type":"STRING"},{"id":37,"origin_id":20,"origin_slot":0,"target_id":28,"target_slot":0,"type":"STRING"},{"id":38,"origin_id":27,"origin_slot":0,"target_id":28,"target_slot":1,"type":"STRING"},{"id":42,"origin_id":23,"origin_slot":0,"target_id":28,"target_slot":2,"type":"BOOLEAN"},{"id":43,"origin_id":8,"origin_slot":0,"target_id":-20,"target_slot":0,"type":"IMAGE"},{"id":45,"origin_id":-10,"origin_slot":0,"target_id":19,"target_slot":0,"type":"STRING"},{"id":46,"origin_id":-10,"origin_slot":1,"target_id":24,"target_slot":0,"type":"BOOLEAN"},{"id":72,"origin_id":28,"origin_slot":0,"target_id":6,"target_slot":1,"type":"STRING"},{"id":73,"origin_id":-10,"origin_slot":2,"target_id":16,"target_slot":14,"type":"BOOLEAN"},{"id":74,"origin_id":-10,"origin_slot":3,"target_id":16,"target_slot":5,"type":"INT"},{"id":75,"origin_id":-10,"origin_slot":4,"target_id":5,"target_slot":0,"type":"INT"},{"id":76,"origin_id":-10,"origin_slot":5,"target_id":5,"target_slot":1,"type":"INT"},{"id":77,"origin_id":-10,"origin_slot":6,"target_id":3,"target_slot":4,"type":"INT"},{"id":78,"origin_id":-10,"origin_slot":7,"target_id":23,"target_slot":0,"type":"BOOLEAN"},{"id":79,"origin_id":-10,"origin_slot":8,"target_id":15,"target_slot":1,"type":"COMBO"},{"id":80,"origin_id":-10,"origin_slot":9,"target_id":15,"target_slot":2,"type":"FLOAT"},{"id":81,"origin_id":-10,"origin_slot":10,"target_id":27,"target_slot":1,"type":"STRING"},{"id":82,"origin_id":-10,"origin_slot":11,"target_id":10,"target_slot":0,"type":"COMBO"},{"id":83,"origin_id":-10,"origin_slot":12,"target_id":11,"target_slot":0,"type":"COMBO"},{"id":84,"origin_id":-10,"origin_slot":13,"target_id":12,"target_slot":0,"type":"COMBO"}],"extra":{"ue_links":[]}}]},"config":{},"extra":{"frontendVersion":"1.48.5","VHS_latentpreview":false,"VHS_latentpreviewrate":0,"VHS_MetadataImage":true,"VHS_KeepIntermediate":true,"ue_links":[],"ds":{"scale":0.3905487804878049,"offset":[2748.07962529274,423.33333333333326]}},"version":0.4}

 

[Ozmog]

das offizielle Krea2 Template "Text to Image"

Die Templates habe ich dann gar nicht erst angerührt und von Scratch neu gebaut, hatte da das eine oder andere Youtube-Video gefunden, die gut erklärt haben. Statt auswendig zu lernen, will ich eben auch verstehen, für was ich das mache. So behalte ich es auch dauerhaft im Kopf, als kurz- bis mittelfristig auswendig zu können, und wenn man mal eine Zeit lang nicht dran war, vergisst man es dann doch wieder.

zusätzliches Referenzbild Node hinzuzufügen

Bei Krea2 ist meineswissens nicht ein einfacher Referenzbild-Workflow möglich. Es benötigt da die Krea2edit Nodes. Hast du die?
Ich habe es jedenfalls so laufen. Mit einem und zwei Referenzbilder. Es läuft, aber im Vergleich zum einfachen Generieren auch mal träge und ach mal mit Hängern oder sogar Fehler wegen VRAM Mangel.

 

[Keuleman]

Puh, Zwickmühle. H3 Turbo Lora oder LTX 2.5 fummeln, bis die bestehenden Workflows damit laufen (LTX Director muss wohl damit laufen, auch bestehende Loras von 2.3). LTX hat immer noch Probleme, Prompts zu befolgen... H3 ist tatsächlich meist eine "1 Prompt, 1 Result" Ding. Dafür halt langsamer und ich habe noch nichts "image to image" und "LTX Director" mässiges.

Ich hab einfach zu wenig Zeit neben Arbeit und Familie ;-D

Ich sollte was in dem Bereich arbeiten. Dann hätte ich Zeit (während das eine läuft was anderes anstossen etc) X-D