<h1 id="v.4">V.4</h1>I’ve simplified the "One for All" mode and added a refiner with upscaling that can be toggled on or off. I’ve also included a "skimmed" CFG, which allows for a much higher CFG value without causing the image to become overexposed or "burnt."He simplificado el modo One for all y he puesto un refiner con upscale que se puede activar o desactivar. También le he puesto un skimmed cfg para poder subir mucho el cfg sin que queme la imagen.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/253044f5-bb75-4dc9-92ee-8239771647d0/original=true/253044f5-bb75-4dc9-92ee-8239771647d0.jpeg" /><hr /><h2 id="v.3">V.3</h2>In this new version now saves the image prompt and can also be used without needing to enable Nodes 2.0. I’ve added some mini-tutorials to the workflow, along with useful links and tips.En esta nueva versión ya guarda el prompt de las imágenes y también puede usarse sin tener que activar el Nodes 2.0. He añadido unos mini tutoriales en el workflow y también enlaces y consejos utiles.<hr /><h3 id="v.2">V.2</h3><h2 id="hello!">Hello!</h2>I’ve created a version that doesn’t use subgraphs or the LC node for BBOXes, as the previous setup was causing a lot of strange errors.The only annoying part is having to constantly toggle the "Nodes 2.0" setting on and off, since I had to switch to the native ComfyUI BBOX node to prevent the workflow from breaking.I ran into this error:<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/b831e064-e8db-436e-a8de-3bc5f8693dc3/original=true/b831e064-e8db-436e-a8de-3bc5f8693dc3.jpeg" />It was a total disaster... I hope you didn’t experience the same thing.<div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist src="https://www.youtube.com/embed/CsliQD7TvhI?si=mBCok0FDIcl7XMux" start="0"></iframe></div>I still haven’t figured out why the metadata isn’t being saved to the generated image... If I manage to fix that, I’ll release another update to the workflow.<h1 id="hola!">¡Hola!</h1>He hecho una versión sin subgrafos y sin el nodo de LC para las BBOXes porque me daba muchos errores extraños.Lo único fastidioso es tener que activar y desactivar todo el rato lo de Nodes 2.0 porque he tenido que usar el nodo nativo de comfyui para BBOXes que no causa que se rompa el workflow.A mí me ha salido este error:<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/986c3c50-0367-481b-be8e-2f134b59e701/original=true/986c3c50-0367-481b-be8e-2f134b59e701.jpeg" />Ha sido una catastrofe... espero que no os haya pasado a vosotros.<div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist src="https://www.youtube.com/embed/63m5GXLQq08?si=01EHMQzqboiK7HBW" start="0"></iframe></div>Aún sigo sin saber por qué no le guarda los metadatos a la imagen generada... Si consigo hacerlo haré otra actualización del Workflow...<hr />Hello!This workflow is designed for creating story sequences. I personally use the prompts from my pack available here: <a target="_blank" rel="ugc" href="https://civarchive.com/models/2545356/wildcards-for-81189-loras">https://civarchive.com/models/2545356/wildcards-for-81189-loras</a>I simply remove the parts of the text that Moss-image doesn't recognize, and that’s it!Example:Text from my pack: &lt;lora:groping:0.3&gt;, gropingV1, yuri, kiss, groping, 2girls, grabbing another's breast, fingering under clothes, heavy breathing, open mouth, heavy breathing, steam, tongue out, kiss, hug from behindText to copy into the input box: yuri, kiss, groping, 2girls, grabbing another's breast, fingering under clothes, heavy breathing, open mouth, heavy breathing, steam, tongue out, kiss, hug from behindI could look into creating a version of the pack that excludes the LoRA tags and trigger words—it just depends on what people would like to see.Since LoRAs cannot currently be used in the input bboxes, you’ll need to describe the characters without including their specific LoRA.¡Hola!Este workflow es para hacer secuencias de una historia. Yo lo que hago es usar las de mi pack de <a target="_blank" rel="ugc" href="https://civarchive.com/models/2545356/wildcards-for-81189-loras">https://civarchive.com/models/2545356/wildcards-for-81189-loras</a>Quito la parte de texto que Moss-image no entiende y ya está.Ejemplo:Texto en mi pack: &lt;lora:groping:0.3&gt;, gropingV1, yuri, kiss, groping, 2girls, grabbing another's breast, fingering under clothes, heavy breathing, open mouth, heavy breathing, steam, tongue out, kiss, hug from behindTexto para copiar en el recuadro: yuri, kiss, groping, 2girls, grabbing another's breast, fingering under clothes, heavy breathing, open mouth, heavy breathing, steam, tongue out, kiss, hug from behindPodría intentar hacer un pack ya sin la lora y la palabra clave, depende de lo que diga la gente.Como aún no se pueden poner loras en las cajas, hay que describir a los personajes sin poner su Lora.

Eto_txt2imgSecuence_Moss-Image

Files with SHA256: 60c781d7238a15f76771aa10ea519025938263e87999065e22a6934e35c95c95