Skip to content
Date and source

Dates follow the news timeline in Beijing time and include both endpoints.

Oct 18Sun
  1. TelosCompass · note AI视频创作实践22

    Choosing an AI video generation model: generating 2 clips under the same conditions is enough to decide

    When using DomoAI to generate AI video from photos, you don't need to look at names or reputations first when choosing a model: first write down the one thing this scene absolutely must not fail at, then narrow the candidates to 2—model A as the one you've used, and model B as another that looks suitable—and generate 1 clip each using the same source image, the same prompt, and output conditions that are as similar as possible. Compare in the order of "the point that must not fail → specified action → production burden (waiting time, credits consumed, places needing rework)," record the reason for adoption based only on this round's results, and mark untried models as "not compared."

  2. note · kiryu_mizuki · 图像视频创作分享22

    Adding a post-processing step to AI-generated images: a 4-step process of cutout, blur, gradient, and overlay

    Japanese creator Mizuki shared a post-processing method for AI-generated images: first cut out the character, then blur the inside of the character separately and overlay it at 15–20% opacity, next add a gradient to the character matching the background tones (overlay mode, 10–20% opacity), and finally add color to lit areas such as armor, sword, face, and petals using overlay mode at 20% opacity. The author recommends keeping layer opacity within 15–20% when cutting out the character separately, which can make the image softer and enhance the sense of depth.

  3. ArchiGen AI · 图像生成与ComfyUI实战22

    How Veras gives every generated idea a traceable parent source

    Chaos's Veras can be launched from the V-Ray VFB History panel using a selected historical rendered frame rather than the currently visible image, thereby establishing a traceable source for generated images. Users need to record the project, camera, model version, design option, and render sequence number in the frame name, and retain the prompt, engine (currently Nano Banana 2), and number of attempts. Veras works only at the image level and does not modify the 3D scene; generated results need to be rendered again for verification.

  4. ArchiGen AI · 图像生成与ComfyUI实战30

    How to review 4K ComfyUI architectural renders at three scales

    Reviewing 4K ComfyUI architectural renders requires three scales: fit view to check composition and design drift, delivery size to judge the actual look, and 200% zoom to inspect seams, halos, and fabricated details. The upscaling path determines the risk; in the official example, UpscaleModelLoader pixel upscaling and img2img second-pass generation are different in nature, and the latter requires recording denoise, sampler, and seed. Differences should be marked as intentional, tolerable, or blocking.

Oct 17Sat
  1. Generative AI · 创作教程与项目分享34

    Tensor.Art vs. PixAI anime AI generators: which should you choose?

    A hands-on comparison of two browser-based anime AI generators, Tensor.Art and PixAI, testing text scenes, signboard text, and a crowded four-person scene with three prompts. Tensor.Art requires choosing checkpoints, stacking LoRAs, and tuning parameters yourself, and supports ComfyUI workflows and self-trained models; PixAI focuses on its in-house model Tsubaki.3, allows full-sentence descriptions, character reuse via @image1, and instruction-based image edits, and hosts over 1.8 million community models and LoRAs.

  2. Reddit comfyui · 创作与工作流分享22

    How to automatically locate a specified character's face in a multi-person image for stylized PuLID local repainting (API workflow)?

    A user built a two-stage automated API-oriented workflow in ComfyUI; when using Flux for stylized PuLID local repainting on multi-person images, because character positions differ on every generation in the API scenario, selecting by bbox index order (left to right or largest face) often causes the wrong character's face to be swapped or identity bleed.

  3. Reddit comfyui · 创作与工作流分享22

    Binding multiple real human identities to their own regions in a stylized FLUX workflow

    A ComfyUI user encountered multi-character identity crosstalk in a stylized FLUX workflow: global reference conditioning caused hairstyles and accessories to bleed between characters or onto background people, and post-generation face refinement also pasted A's face onto B or background people. The goal is about 15–20 seconds per final image on a single 24–32 GB GPU, and the user is considering PuLID-Flux with per-character attn_mask, as well as FLUX 3 Image Klein's reference-image-plus-bounding-box binding approach.

  4. Wireflow · AI短剧与视频制作教程65

    Kling Motion Control hands-on: one dance clip run through three settings

    Kling Motion Control can transfer motion from a reference video onto a character image; the author used a 5-second dance clip to drive a 3D baker character, testing three settings: 3.0 Pro with orientation video, 3.0 Standard with orientation video, and 3.0 Pro with orientation image.

    Why it matters: A hands-on comparison of the same 5-second dance clip under three settings, showing differences in time, credit consumption, and image quality, which can be used to decide between Pro and Standard.

  5. Reddit StableDiffusion · 创作者与工作流分享34

    ComfyUI too complex? A quick start to Invoke for those who prefer canvas operations

    A creator posted a roughly four-minute video demonstrating how to get started with Stable Diffusion using Invoke: generating a rainy café scene, changing the lighting with a fixed seed, then using the canvas to add a flower pot and extend the image. The author says Invoke itself is free and open source and considers it one of the most underrated Stable Diffusion tools, suited to users who would rather spend time on the image. The author also discloses being the developer of LoRA Pilot, which includes Invoke.

  6. note · imamirai_editors · 图像视频创作分享22

    ChatGPT image generation: free distribution of prompts for 5 anime illustration scenes—morning, noon, evening, night, and late night

    Aizawa released the ChatGPT image generation prompts accompanying 5 anime illustrations posted on X from the morning of October 16 to late night on the 17th, covering five scenes: early morning by the window, noon overhead view tying a ribbon, evening nightscape, night hugging a flower pillow, and late night hugging a pillow with eyes closed.

  7. note · uuhai0625 · 图像视频创作分享22

    Record of making picture book volumes 4 and 5 with Midjourney: multiple separate characters were hard to get in one image, so single-character generation followed by collage became the approach

    When making picture book volumes 4 and 5 with Midjourney, getting one image to depict multiple separate characters at once did not succeed within what was tried, so generating characters individually and then collaging became the main method. The approach of having Midjourney redraw the collage was tried on 4 spreads, and only 2 were adopted; volume 5 used no redraws. The character reference feature Omni Reference can hold only 1 reference image.

  8. note · kasmintea · 图像视频创作分享12

    AI illustration 3 specialized in atmosphere: using particles to create a sense of air

    To create a sense of air in AI illustrations, particles can be added after atmosphere and shadows. Placing particles of different sizes in front and behind creates depth, and the particles can be replaced with elements such as light, dust, leaves, water droplets, playing cards, or casino chips to strengthen presence and complete the relationships in the image. As small, numerous elements, particles are more easily perceived as a sense of air rather than noise, but how to eliminate noise remains a challenge.

  9. Stable Diffusion Tutorials · 创作教程78

    Flux 3 Klein tutorial: image generation and editing with 10 reference images (BF16/FP8/NVFP4/GGUF)

    Flux 3 Klein is a 5.6B-parameter open-weight model that merges text-to-image and image editing into a single model, accepting text prompts and up to 10 reference images and outputting a single image, supporting bounding-box instructions and 4K resolution, with Qwen3-VL-4B-Instruct as the text encoder.

    Why it matters: Flux 3 Klein merges text-to-image and multi-image editing into one model; the tutorial gives model choices for each VRAM tier and ComfyUI workflow configuration.

  10. Reddit comfyui · 创作与工作流分享15

    Request for help with MiniMax H3 REF2VA workflow for generating long videos on an RTX3060

    A user is running MiniMax H3 on an RTX3060 (12GB VRAM, 64GB RAM) to generate long videos, using the int8 convrot model and int4 encoder, with memory usage around 50GB and peaks exceeding 60GB. They use the standard ComfyUI template plus VRAM/RAM cleanup nodes, but most REF2VA workflows on Civitai rely on old custom nodes or throw errors, so they are seeking a working workflow that can support long videos with 12GB of VRAM.

  11. Reddit comfyui · 创作与工作流分享12

    Using Flux + Qwen + Architecture LoRA for architecture image-to-image in ComfyUI, detail and image cleanliness insufficient, seeking tuning advice

    A user has built an image-to-image workflow for architectural rendering in ComfyUI using FLUX, Qwen, and Architecture LoRA. The overall direction is right, but the level of detail and image cleanliness are still not ideal. The user is asking the community for advice on adjustments to denoising settings, prompt/conditioning handling, sampling, and related areas.

  12. Monster社区 · 中文ComfyUI实测与工作流22

    Kroma V0.3.1 Turbo: hands-on test of text-to-image startup and ComfyUI entry

    Kroma V0.3.1 Turbo demonstrates the process of launching and entering the default text-to-image interface, which includes a prompt input area, size settings, iteration steps, and a generate button. When using ComfyUI, a custom port must be accessed; the demo mentions 8187, but the actual value should depend on one's own environment configuration. The original video states that this version supports 4-step iteration and is usable with 8G of VRAM; this article has not independently verified VRAM usage or speed.

  13. Reddit comfyui · 创作与工作流分享10

    ComfyUI 0.40.0 workflow crashes after dragging in output with metadata?

    A ComfyUI 0.40.0 user reports that after dragging in an output with metadata, the browser briefly freezes, the entire workflow is blown apart, all subgraphs disappear, and all nodes are flattened onto the main graph; it also swallows the current workflow name, making it easy to overwrite. Rolling back comfyui-frontend-package to 1.52.7 restores stability, and the user suspects rgthree is the main culprit.

  14. Reddit comfyui · 创作与工作流分享10

    How to undervolt a 4080 Super GPU for ComfyUI use

    A user on the Reddit ComfyUI subreddit is asking for help undervolting their 4080 Super with MSI Afterburner, mainly for ComfyUI image generation. After following YouTube tutorials, ComfyUI either crashes outright or outputs all-black images, and nearly all undervolting tutorials they can find online are aimed at gaming, with none related to ComfyUI, so they want to ask whether there are more suitable steps.

  15. Reddit StableDiffusion · 创作者与工作流分享8

    Workflow Page Blank After ComfyUI Update, MOSS Unusable

    After updating ComfyUI, a user encountered a completely blank workflow page, with only "idle" shown at the top; dragging in a workflow only produced JSON-format content, and the MOSS they originally wanted to use could not run either. The issue was posted for help in the Reddit StableDiffusion section, and there is currently no solution.

  16. Reddit comfyui · 创作与工作流分享12

    minimax H3 Voice Prompts with Audio Reference Fail, Cannot Change Volume or Timbre

    A user reported that in the minimax H3 reference-image-to-video workflow, after inputting about 20 seconds of normal speech as an audio reference, voice prompts produced no change at all, and all outputs had the same volume and timbre. Attempts to use annotations such as says and whispers quietly were ineffective, and changes such as shouting, whispering, or an excited tone could not be achieved.

  17. あいきみ · 图像编辑与LoRA实测22

    Hands-On with the FLUX 3 LoRA "PastelLora": Pastel Animation Style and ON/OFF Comparison

    The animation-style LoRA "PastelLora" for FLUX 3 has been released, with the trigger word pastelstyle; the author used Aikimi Forge Neo on the FLUX 3 Image Klein distilled version (INT8 ConvRot quantization) with strength 1.0, Steps 16, CFG 1, and 832×1216 portrait orientation to generate 5 sample images.

  18. TelosCompass · note AI视频创作实践22

    Corrected AI Video Should Be Compared with the Pre-Correction Version at the Same Time Point Before Use

    When correcting AI video, keep the pre-correction video and generation conditions, change only one prompt and generate only once, then pause both the pre- and post-correction versions at the same playback moment and compare the same position. The basis for judgment is whether the target problem is fixed, whether new problems appear, and whether the required duration and expression are preserved; then decide whether to use it directly, use only the usable segment, change the approach, or hold off. The article uses a 5-second bike shop video generated by DomoAI as an example, comparing the front wheel shape at 0:03; because the saddle color changes from 0:04 onward, only the 0:00 to 0:03 segment is ultimately used.

  19. Reddit StableDiffusion · 创作者与工作流分享66

    Hands-On with MOSS-Image 5B INT8 Quantization and 8-Step DMD on a 12 GB RTX 3060

    The author tested MOSS-Image (5B single-stream DiT, Qwen3.5-9B text encoder, FLUX.2 VAE, Apache 2.0) on a 12 GB RTX 3060 using 35 difficult prompts and a single seed (42), without cherry-picking images and including failure cases.

    Why it matters: A hands-on test of MOSS-Image's INT8 quantization and 8-step DMD acceleration on 12 GB of VRAM, with a full comparison of text rendering, layout and failure cases.

Oct 16Fri
  1. note · aicreatorpathnet · 图像视频创作分享62

    Using Midjourney 8.2's Edit Model to unify 10 AI character designs

    The author created 10 characters (9 animals plus 1 human) with Midjourney 8.2 for a YouTube AI comedy show, and summarized the core approach to unifying their appearance: use the Edit Model with the exact same two reference images attached each time (a full-body three-view and a face close-up), and state in the prompt what role each reference image plays.

  2. note · promptdesign · 图像视频创作分享18

    A collection of image prompts suited to Instagram ㉘: 100 JSON prompts you can copy directly into ChatGPT, Gemini and more

    This collection compiles 100 JSON-format image generation prompts that can be copied and used directly, aimed at tools such as ChatGPT and Gemini, covering everyday scenes like a gas station at night, a taxi after rain, a station clock, a fishing harbor and a hotel corridor, using cinematic composition and light-and-reflection to create a sense of story. For each subject, besides the base version, it also provides a 9:16 vertical phone version, a 1280×670 note header version, and a variation that changes the background composition and lighting atmosphere.

  3. note · singtothesky79 · 图像视频创作分享15

    Self-taught AI video making ④: the unexpected unfolding from writing a novel with Gemini to manga, animation and a theme song

    A self-taught AI video creator used Gemini to generate a Japanese sword-themed novel in 6 parts, then tried to adapt it into manga and animation. The manga stage yielded only storyboards and images that couldn't form a complete work, but the image quality was very high; after switching to animation, they succeeded in making a "moving manga," and after adding color and voiceover, a theme song was unexpectedly generated automatically.