Stable Diffusion
PulseAugur coverage of Stable Diffusion — every cluster mentioning Stable Diffusion across labs, papers, and developer communities, ranked by signal.
- 2026-05-25 product_launch A user released a custom workflow and nodes for Stable Diffusion to enable local 16-bit ARRI Alexa output. source
23 day(s) with sentiment data
-
Stable Diffusion fine-tuning overcomes safety filters
A user on Reddit shared their experience fine-tuning a Stable Diffusion model with a focus on regular and anime images. They reported that within a few hundred training steps, the model's safety filters were overcome, a…
-
Stable Diffusion user seeks further model quantization for smaller file size
A user on Reddit is seeking to reduce the file size of a Stable Diffusion model by quantizing it further. They have already converted a safetensor model to GGUF format but found the f16 GGUF file too large. The user is …
-
ComfyUI adds advanced latent interpolation nodes
A developer has created new nodes for ComfyUI, a popular interface for Stable Diffusion, that implement Riemannian geodesic interpolation. These nodes offer a more sophisticated method for interpolating between latent s…
-
ComfyUI Prompt Relay bug causes image cycling; downgrade suggested
Users of ComfyUI's Prompt Relay feature are experiencing an issue where keyframe images cycle at the end of generated videos. This problem appears to stem from a regression in ComfyUI's core, specifically between versio…
-
Ideogram 4 GGUF model released for image generation
The Ideogram 4 GGUF model has been released, offering enhanced capabilities for image generation. This release is available via Hugging Face, allowing users to access and utilize the model for their projects. The announ…
-
User finds Ideogram 4 superior to other AI image generators
A user shared their experience transitioning to full-time AI image generation in October 2022, finding Ideogram 4 to be the most impressive tool available. They expressed that running Ideogram 4 on current hardware is a…
-
Stable Diffusion user seeks custom LTXSixGridDirector node
A user on Reddit is seeking a specific custom node for Stable Diffusion, known as LTXSixGridDirector. They are looking for the modded version of this node and have provided a YouTube link to a video demonstrating its us…
-
Stable Diffusion users report LoRA resolution artifacts
Users on the Stable Diffusion subreddit are encountering issues with LoRA (Low-Rank Adaptation) models, specifically concerning resolution artifacts. When LoRA weights are set to 2.0, the intended effect is visible, but…
-
Ideogram 4.0 LoRA training explored by Stable Diffusion users
Users on Reddit are exploring the capabilities of Ideogram 4.0 for training LoRAs, which are custom models used to fine-tune AI image generation. Discussions revolve around achieving accurate multi-character LoRAs and a…
-
AI users seek uncensored image and video generation workflows
Users on Reddit are seeking guidance on creating uncensored images and videos using AI models like Stable Diffusion and Gemma. They are looking for specific workflows and tutorials to achieve high-quality results, parti…
-
AI enables non-coders to build complex Stable Diffusion tools
New custom nodes are rapidly emerging in the Stable Diffusion ecosystem, created by individuals with no prior coding experience who are leveraging advanced AI capabilities. These "vibe-coded" creations, while functional…
-
AI art piece "Peacebloom Burn Slow" shared on Reddit
A user on Reddit shared a piece of AI-generated art titled "Peacebloom Burn Slow," created using Stable Diffusion. The artwork was highlighted as a "masterpiece" by the user who posted it.
-
Stable Diffusion user seeks automated LoRA cycling
A user on Reddit is seeking to automate the process of generating images with Stable Diffusion using multiple LoRAs. They want to cycle through different artist LoRAs sequentially for a single prompt, rather than manual…
-
Stable Diffusion users advised on LTX 2.3 optimization
A Reddit post discusses LTX 2.3, a tool for Stable Diffusion users, highlighting potential misuse and the effectiveness of "seed hunting." The post suggests that users might not be leveraging the tool's full capabilitie…
-
Stable Diffusion users can now generate JSON prompts with Qwen 8B in ComfyUI
A user shared a tip for Stable Diffusion users on Reddit, detailing how to integrate the Qwen 8B model into their workflow using ComfyUI. The method involves utilizing a "Generate Text" node to create JSON prompts, whic…
-
Stable Diffusion users explore advanced i2i workflows with ControlNet and LoRAs
Users on Reddit are inquiring about the existence and implementation of advanced image-to-image (i2i) workflows within Stable Diffusion. Specifically, they are seeking methods to combine ControlNet for pose consistency,…
-
UnHype framework enhances AI model unlearning with dynamic LoRA adaptation
Researchers have developed a new framework called UnHype to improve the process of machine unlearning, specifically for large diffusion models. This method uses hypernetworks to dynamically adjust Low-Rank Adaptation (L…
-
User creates music video using Stable Diffusion workflow
A Reddit user shared their experience creating a music video using Stable Diffusion, detailing the process and the hardware involved. The user employed a specific workflow and generation settings, resulting in a 17-hour…
-
Developer releases offline prompt builder for Stable Diffusion
A developer has created an open-source, offline prompt builder specifically for Stable Diffusion, designed to streamline the creation of anime and fantasy-themed images. The tool features a database of over 500 characte…
-
CyberRealistic Z Image checkpoint impresses Stable Diffusion users
A user on Reddit shared a new image generation model called CyberRealistic Z Image, highlighting its impressive capabilities. The model is available on Civitai, a platform for AI art models. This checkpoint appears to b…