AI image generation

Hallo2 is a portrait image animation technology based on a latent diffusion generation model that generates high-resolution, long-term videos driven by audio. It expands Hallo's capabilities by introducing several design improvements, including generating long-duration videos, 4K resolution videos, and adding the ability to enhance expression control through text prompts. Hallo2's key advantages include high-resolution output, long-term stability, and enhanced control through text prompts, which make it a significant advantage in generating rich and diverse portrait animation content.

Flux AI is a platform that utilizes advanced AI algorithms to generate high-quality images. It uses deep learning models to transform users' ideas into visual masterpieces in seconds. The platform provides features such as real-time generation, customized output, multi-language support, ethical AI and seamless integration, aiming to help users quickly realize their ideas and improve work efficiency. Background information on Flux AI shows that it is committed to responsible AI development, respecting copyright, avoiding bias, and promoting positive social impact.

ComfyGen is an adaptive workflow system focused on text-to-image generation that automates and customizes efficient workflows by learning user prompts. The advent of this technology marks a shift from the use of a single model to complex workflows that combine multiple specialized components to improve the quality of image generation. The main benefit behind ComfyGen is the ability to automatically adjust the workflow based on the user's text prompts to produce higher quality images, which is important for users who need to produce images of a specific style or theme.

AnimeGen is an online tool that uses advanced AI models to convert text prompts into anime-style images. It provides users with a simple and fast way to generate high-quality animation pictures through complex algorithms and machine learning technology, which is very suitable for artists, content creators and animation enthusiasts to explore new creative possibilities. AnimeGen supports more than 80 languages, and the generated images are publicly displayed and can be crawled by search engines. It is a multi-functional creative tool.

AnyPhoto.co is an online platform that uses artificial intelligence technology to provide photo stylization and artistic effects. It achieves efficient model adaptability, fine style control, fast processing speed and excellent image quality through LoRA (Low Rank Adaptation) technology. Users can upload their own portrait photos, easily convert them into hand-drawn sketches, and try out a variety of unique painting styles to create one-of-a-kind works of art. The platform has a friendly interface, supports personalized adjustments, and provides highly complete output, making it very suitable for users who require fast, high-quality image processing.

ComfyUI-Fluxtapoz is a collection of nodes designed for Flux to edit images in ComfyUI. It allows users to edit and style images through a series of node operations, and is especially suitable for professionals who need to perform image processing and creative work. This project is currently open source and follows the GPL-3.0 license agreement, which means that users can freely use, modify and distribute the software, but they need to comply with the relevant provisions of the open source license.

Toy Box Flux is a 3D rendering model trained on AI-generated images, which combines the weights of existing 3D LoRA models and Coloring Book Flux LoRA to form a unique style. This model is particularly suitable for generating images of toy designs with a specific style. It performs best on objects and human subjects, with animal performance erratic due to insufficient data in the training images. In addition, the model can improve the realism of indoor 3D renderings. There are plans to strengthen the consistency of this style in v2 by mixing more generated and pre-existing output.

DisEnvisioner is an advanced image generation technology that isolates and enhances subject features to generate customized images without tedious adjustments or reliance on multiple reference images. This technology effectively distinguishes and enhances subject features while filtering out irrelevant attributes, achieving superior personalization quality in terms of editability and identity preservation. The research background of DisEnvisioner is based on the current need in the field of image generation for extracting subject features from visual cues. It solves the challenges of existing technologies in this field through innovative methods.

Animate-X is a universal LDM-based animation framework for various character types (collectively referred to as X), including human mimic characters. This framework enhances motion representation by introducing pose indicators, which can more comprehensively capture motion patterns from driving videos. Key benefits of Animate-X include in-depth modeling of motion, the ability to understand the motion patterns driving video and flexibly apply them to target characters. In addition, Animate-X introduces a new Animated Anthropomorphic Benchmark (A2Bench) to evaluate its performance on general and widely applicable animated images.

RealAnime - Detailed V1 is a LoRA model based on Stable Diffusion, specifically designed to generate realistic anime-style images. Through deep learning technology, this model can understand and generate high-quality animation character images to meet the needs of animation enthusiasts and professional illustrators. Its importance lies in its ability to greatly improve the efficiency and quality of animation-style image generation and provide strong technical support for the animation industry. Currently, the model is provided on the Tensor.Art platform, and users can use it online without downloading and installing, which is convenient and fast. In terms of price, users can unlock download benefits by purchasing the Buffet plan and enjoy more flexible usage.

FacePoke is an AI-powered real-time head and face transformation tool that allows users to manipulate facial features through an intuitive drag-and-drop interface, breathing life into portraits for realistic animations and expressions. FacePoke utilizes advanced AI technology to ensure that all edits maintain a natural and realistic appearance, while automatically adjusting surrounding facial areas to maintain the overall integrity of the image. This tool stands out for its user-friendly interface, real-time editing capabilities, and advanced AI-driven adjustments, making it suitable for users of all skill levels, whether they are professional content creators or beginners.

Meissonic is a non-autoregressive masked image modeling text-to-image synthesis model capable of generating high-resolution images. It is designed to run on consumer grade graphics cards. The importance of this technology lies in its ability to utilize existing hardware resources to provide users with a high-quality image generation experience while maintaining high operating efficiency. Background information on Meissonic includes its paper published on arXiv, and its model and code on Hugging Face.

The text-to-image generation model developed by the Tsinghua University team is open source, has broad application prospects in the field of image generation, and has the advantages of high-resolution output.

Flux Ghibsky Illustration is a text-based image generation model that combines the fantastical details of Hayao Miyazaki's animation studio with the serene skies of Makoto Shinkai's work to create enchanting scenes. This model is particularly suitable for creating fantastic visual effects, and users can generate images with a unique aesthetic through specific trigger words. It is an open source project based on the Hugging Face platform, allowing users to download models and run them on Replicate.

Easy Anime Maker is an artificial intelligence-based animation generator that uses deep learning techniques such as generative adversarial networks to convert user-entered text descriptions or uploaded photos into anime-style artwork. The importance of this technology is that it lowers the threshold for creating animation art, allowing users without professional painting skills to create personalized animation images. Product background information shows that it is an online platform where users can generate anime art through simple text prompts or uploading photos, making it ideal for anime enthusiasts and professionals who need to quickly generate anime-style images. The product provides a free trial, and users can get 5 free points after registration. If you need more generation needs, you can choose to purchase points without subscribing.

Image Describer is a tool that uses artificial intelligence technology to upload images and output image descriptions according to user needs. It understands image content and generates detailed descriptions or explanations to help users better understand the meaning of the image. This tool is not only suitable for ordinary users, but also helps visually impaired people understand the content of pictures through text-to-speech function. The importance of the image description generator lies in its ability to improve the accessibility of image content and enhance the efficiency of information dissemination.

FLUX.1-dev-Controlnet-Inpainting-Beta is an image inpainting model developed by Alimama's creative team. This model has significant improvements in the field of image inpainting, supporting direct processing and generation of 1024x1024 resolution without additional amplification steps, providing higher quality and more detailed output results. The model has been fine-tuned to capture and reproduce more detail of the repaired area and provide more precise control over the generated content with enhanced prompt interpretation.

FLUX.1-Turbo-Alpha is an 8-step distillation Lora based on the FLUX.1-dev model, released by AlimamaCreative Team. This model uses a multi-head discriminator to improve the quality of distillation and can be used in FLUX-related models such as text-to-image (T2I) and inpainting control networks. It is recommended to use a guide ratio of 3.5 and a Lora ratio of 1. The model is trained on 1M open source and in-house source images, uses adversarial training to improve quality, fixes the original FLUX.1-dev transformer as the discriminator backbone, and adds multi-heads on each layer of transformers.

FLUX.1-dev-LoRA-One-Click-Creative-Template is an image generation model trained on LoRA, provided by Shakker-Labs. This model focuses on creative photo generation and can transform users' text prompts into creative images. The model uses advanced text-to-image generation technology and is particularly suitable for users who need to quickly generate high-quality images. It is based on the Hugging Face platform and can be easily deployed and used. Non-commercial use of the model is free, but commercial use requires compliance with the corresponding license agreement.

Free AI Anime Generator is an online platform based on artificial intelligence technology that allows users to generate high-quality anime-style pictures with simple clicks. This platform utilizes advanced AI algorithms to make it easy for even non-professionals to create unique works of art. It not only provides animation fans with a platform to realize their creativity, but also provides artists and designers with a tool to explore new ideas. The platform is completely free and easy to use, and is an innovation in the field of animation art creation.

Flux 1.1 Pro AI is an advanced artificial intelligence-based image generation platform that leverages cutting-edge AI technology to transform users' text prompts into high-quality visuals. The platform delivers 6x faster image generation, significantly improved image quality, and enhanced compliance with prompts. Flux 1.1 Pro AI is not only suitable for artists and designers, but also for content creators, marketers and other professionals, helping them realize visual ideas in their respective fields and improve creative efficiency and quality.

OneIMG is an online image generation tool based on artificial intelligence technology. It generates corresponding images through text descriptions input by users. The application of this technology can greatly improve the work efficiency of designers and creative workers, as it can quickly transform ideas into visual images. The background information of OneIMG shows that it is an innovative product designed to simplify the image creation process through AI technology. Currently, OneIMG offers a free trial, but the specific pricing strategy has not yet been made clear.

Cooraft is an app that uses artificial intelligence technology to transform ordinary photos into works of art. It can transform selfies and everyday photos into creative and artistic animations and renderings, offering a variety of art styles from 3D cartoons to classic paintings. Cooraft can not only beautify portraits, but also convert various inputs such as sketches, paintings, and line drawings into new renderings to achieve the transformation from 2D to 3D. In addition, Cooraft also provides a subscription service through which users can obtain more advanced features.

Momo XL is an SDXL-based anime-style model that has been fine-tuned to generate high-quality, detailed, and colorful anime-style images. It is especially suitable for artists and animation enthusiasts, and supports tag-based prompts to ensure the accuracy and relevance of output results. In addition, Momo XL is also compatible with most LoRA models, allowing users to perform diverse customization and style conversion.

ACE is a versatile creator and editor based on diffusion transformation, which can achieve joint training of multiple visual generation tasks through the unified conditional format Long-context Condition Unit (LCU) input. ACE solves the problem of lack of training data through efficient data collection methods and generates accurate text instructions through multi-modal large-scale language models. ACE has significant performance advantages in the field of vision generation, making it easy to build chat systems that respond to any image creation request, avoiding the cumbersome processes typically employed by vision agents.

Viewly is a powerful AI image recognition application that can identify the content in images, compose poems and translate them into multiple languages ​​through AI technology. It represents the current cutting-edge technology of artificial intelligence in the fields of image recognition and language processing. Its main advantages include high recognition accuracy, multi-language support and creative AI poetry writing functions. Viewly’s background information shows that it is a continuously updated product dedicated to providing users with more innovative features. Currently, the product is available to users for free.

PixelHaha is an AI art image generator that allows users to create various styles of AI art works through text prompts. Users can describe desired images based on their inspiration, and AI will then transform these descriptions into images. The importance of this product lies in its ability to quickly transform ideas into visual works, greatly lowering the threshold for artistic creation, and providing a unique AI character to combine with the user's soul mate.

DressRecon is a method for reconstructing temporally consistent 4D human models from monocular videos, focusing on handling very loose clothing or handheld object interactions. The technique combines general prior knowledge of the human body (learned from large-scale training data) with specific "bag-of-skeleton" deformations for individual videos (fitted through test-time optimization). DressRecon separates body and clothing deformations by learning a neural implicit model as separate motion model layers. To capture the subtle geometry of clothing, it leverages image-based prior knowledge such as human pose, surface normals, and optical flow to adjust during the optimization process. The generated neural fields can be extracted into temporally consistent meshes or further optimized into explicit 3D Gaussians to improve rendering quality and enable interactive visualization. DressRecon delivers higher 3D reconstruction fidelity than previous techniques on datasets containing highly challenging clothing deformations and object interactions.

DreamWaltz-G is an innovative framework for text-driven generation of 3D avatars and expressive full-body animation. At its core is skeleton-guided scoring distillation and hybrid 3D Gaussian avatar representation. This framework improves the consistency of viewing angles and human poses by integrating the skeleton control of a 3D human template into a 2D diffusion model, thereby generating high-quality avatars and solving problems such as multiple faces, extra limbs, and blur. In addition, the hybrid 3D Gaussian avatar representation enables real-time rendering, stable SDS optimization and expressive animation by combining neural implicit fields and parametric 3D meshes. DreamWaltz-G is very effective in generating and animating 3D avatars, surpassing existing methods in both visual quality and animation expressiveness. Additionally, the framework supports a variety of applications, including human video reenactment and multi-subject scene composition.

PMRF (Posterior-Mean Rectified Flow) is a newly proposed image restoration algorithm designed to solve the distortion-perceptual quality trade-off problem in image restoration tasks. It proposes a novel image restoration framework by combining posterior mean and correction flow, which can reduce image distortion while ensuring the perceptual quality of the image.

BlinkShot is a real-time AI image generator based on Together AI, which uses Flux technology to generate images in milliseconds when the user inputs prompts. The product is 100% free and open source and is designed to provide creatives and developers with the ability to quickly generate images to support their design and creative work.

Inverse Painting is a diffusion model-based method that generates time-lapse videos of the painting process from a target painting. The technology learns the painting process of real artists through training, can handle multiple art styles, and generates videos similar to the painting process of human artists. It combines text and region understanding, defines a set of painting instructions, and updates the canvas using a novel diffusion-based renderer. This technique is not only capable of handling the limited acrylic painting styles in which it was trained, but also provides reasonable results for a wide range of art styles and genres.

HeadshotAI is a platform that uses artificial intelligence technology to generate realistic avatars. It uses advanced algorithms to analyze uploaded photos and generate avatars with professional photography effects. The importance of this technology is that it allows individuals to obtain high-quality avatars at a lower cost and in a more convenient way, thereby enhancing their personal brand and professional image. Key benefits of HeadshotAI include unparalleled realism, easy customization, rapid generation, affordability, and seamless integration.

Depth Pro is a research project for monocular depth estimation that can quickly generate high-precision depth maps. The model leverages multi-scale visual transformers for dense predictions and is trained on a combination of real and synthetic datasets to achieve high accuracy and detail capture. It only takes 0.3 seconds to generate a 2.25-megapixel depth map on a standard GPU. It is fast and highly accurate and is of great significance to fields such as machine vision and augmented reality.

Minionverse is an AI-based creative workflow that generates images by using different nodes and models. This workflow is inspired by an online glif application and provides a video tutorial to guide users on how to use it. It contains a variety of custom nodes that can perform text replacement, conditional loading, image saving and other operations. It is very suitable for users who need to generate and edit images.

Flex3D is a two-stage process that generates high-quality 3D assets from a single image or text prompt. This technology represents the latest advancement in the field of 3D reconstruction and can significantly improve the efficiency and quality of 3D content generation. The development of Flex3D is supported by Meta and team members with deep backgrounds in 3D reconstruction and computer vision.

Flux_Xiaohongshu real style model is an AI model focused on generating extremely realistic and natural daily photos. It uses the latest artificial intelligence technology and deep learning algorithms to generate photos with the realistic style of Xiaohongshu. This mockup is particularly suitable for users who need to post high-quality, photorealistic photos on social media, as well as professionals who work in art and design. The model provides a variety of parameter settings to adapt to different usage scenarios and needs.

FLUX1.1 [pro] is the latest image generation model released by Black Forest Labs, which has significant improvements in speed and image quality. This model delivers six times the speed of its predecessor while improving image quality, prompt compliance, and diversity. FLUX1.1 [pro] also provides more advanced customization options and better cost performance, suitable for developers and enterprises that require efficient, high-quality image generation.

PuLID-Flux ComfyUI implementation is an image processing model based on ComfyUI, which uses PuLID technology and Flux model to achieve advanced customization and processing of images. This project was inspired by cubiq/PuLID_ComfyUI and is a prototype that uses some handy model tricks to handle the encoder part. The developers wish to test the quality of the model before re-implementing it more formally. For better results, it is recommended to use the 16-bit or 8-bit version of the GGUF model.

OpenFLUX.1 is a fine-tuned version based on the FLUX.1-Schnell model, removing the distillation process, making it fine-tunable, and has the open source, permissive license Apache 2.0. This model is capable of generating stunning images and does it in just 1-4 steps. It's an attempt to remove the distillation process and create an open source licensing model that can be fine-tuned.

Stable Video Portraits is an innovative hybrid 2D/3D generation method that utilizes pre-trained text-to-image models (2D) and 3D morphological models (3D) to generate realistic dynamic face videos. This technology upgrades the general 2D stable diffusion model to a video model through person-specific fine-tuning. By providing a time-series 3D morphological model as a condition and introducing a temporal denoising process, it generates a face image with temporal smoothness that can be edited and transformed into a text-defined celebrity image without additional test-time fine-tuning. This method outperforms existing monocular head avatar methods in both quantitative and qualitative analyses.

Posterior-Mean Rectified Flow (PMRF) is a novel image restoration algorithm that minimizes the mean square error (MSE) by optimizing the posterior mean and rectified flow model while ensuring image fidelity. The PMRF algorithm is simple and efficient, and its theoretical basis is to optimize the posterior mean prediction (minimum mean square error estimate) to match the real image distribution. This algorithm performs well in image restoration tasks, can handle various degradation problems such as noise and blur, and has good perceptual quality.

PhysGen is an innovative image-to-video generation method that converts single images and input conditions (e.g., forces and torques exerted on objects in the image) into realistic, physically plausible, and temporally coherent videos. The technology enables dynamic simulation in image space by combining model-based physical simulation with a data-driven video generation process. Key benefits of PhysGen include that the generated videos appear physically and visually realistic and can be precisely controlled, demonstrating its superiority over existing data-driven image-to-video generation efforts through quantitative comparisons and comprehensive user studies.

CogView3 is a cascade diffusion-based text-to-image generation system using the relay diffusion framework. The system begins the diffusion process from these noisy images by breaking the high-resolution image generation process into multiple stages and adding Gaussian noise on the low-resolution generation results by relaying the super-resolution process. CogView3 surpasses SDXL in generating images, with faster generation speed and higher image quality.

Revisit Anything is a visual location recognition system that uses image fragment retrieval technology to identify and match locations in different images. It combines SAM (Spatial Attention Module) and DINO (Distributed Knowledge Distillation) technologies to improve the accuracy and efficiency of visual recognition. This technology has important application value in fields such as robot navigation and autonomous driving.

GGHead is a 3D generative adversarial network (GAN) based on 3D Gaussian scattering representation for learning 3D head priors from a collection of 2D images. This technique simplifies the prediction process by exploiting the regularity of the UV space of the template head mesh to predict a set of 3D Gaussian properties. The main advantages of GGHead include high efficiency, high-resolution generation, full 3D consistency, and the ability to achieve real-time rendering. It improves the geometric fidelity of the generated 3D head through a novel total variation loss, ensuring that neighboring rendered pixels come from similar Gaussians in UV space.

Omni-Zero-Couples is a zero-sample stylized couple portrait creation model using the diffusers pipeline. It uses deep learning technology to generate portraits of couples with specific artistic styles without the need for pre-defined style samples. This technology has broad application prospects in the fields of artistic creation, personalized gift making and digital entertainment.

Flux.1-dev Controlnet Upscaler is an image upscaling model based on the Hugging Face platform, which uses advanced deep learning technology to increase the resolution of images while maintaining image quality. This model is particularly suitable for scenarios that require lossless amplification of images, such as image editing, game development, virtual reality, etc.

HelloMeme is a diffusion model integrating spatial weaving attention, aiming to embed high-fidelity and rich conditions into the image generation process. This technology generates videos by extracting the features of each frame in the driving video and using them as input to HMControlModule. By further optimizing the Animatediff module, the continuity and fidelity of the generated videos are improved. In addition, HelloMeme also supports facial expressions generated through ARKit facial blend shape control, as well as Lora or Checkpoint based on SD1.5, implementing a hot-swappable adapter for the framework that will not affect the generalization ability of the T2I model.

Cog inference for flux models is an inference engine for FLUX.1 [schnell] and FLUX.1 [dev] models developed by Black Forest Labs. It supports compilation and quantization, sensitive content checking, and img2img support, aiming to improve the performance and security of image generation models.

PortraitGen is a 2D portrait video editing tool based on multi-modal generation priors. It can upgrade 2D portrait videos to 4D Gaussian fields to achieve multi-modal portrait editing. This technology can quickly generate and edit 3D portraits by tracking SMPL-X coefficients and using a neural Gaussian texture mechanism. It also proposes an iterative dataset update strategy and multi-modal face-aware editing module to improve expression quality and maintain personalized facial structure.

This is a method of creating re-illuminable radiation fields by leveraging priors extracted from 2D image diffusion models. This method is able to convert multi-view data captured under single illumination conditions into a dataset with multiple illumination effects and represent the re-illuminable radiation field through 3D Gaussian splats. This method does not rely on precise geometry and surface normals, and is therefore more suitable for handling cluttered scenes with complex geometries and reflective BRDFs.

diffusion-e2e-ft is an open source image conditional diffusion model fine-tuning tool that improves the performance of specific tasks by fine-tuning pre-trained diffusion models. The tool supports a variety of models and tasks, such as depth estimation and normal estimation, and provides detailed usage instructions and model checkpoints. It has important applications in the fields of image processing and computer vision, and can significantly improve the accuracy and efficiency of models on specific tasks.

MagicFace is a technology that enables personalized portrait synthesis without training and is able to generate high-fidelity portrait images based on multiple given concepts. This technology enables multi-concept personalization by precisely integrating reference concept features into generated regions at the pixel level. MagicFace introduces a coarse-to-fine generation process, including two stages of semantic layout construction and conceptual feature injection, implemented through the Reference-aware Self-Attention (RSA) and Region-grouped Blend Attention (RBA) mechanisms. Not only does this technology excel in portrait synthesis and multi-concept portrait customization, it can also be used for texture transfer, enhancing its versatility and practicality.

StoryMaker is an AI model focused on text-to-image generation that can generate coherent images of characters and scenes based on text descriptions. By combining advanced image generation technology with face encoding technology, it provides users with a powerful tool for creating storytelling visual content. The main advantages of this model include efficient image generation capabilities, precise control of details, and high responsiveness to user input. It has broad application prospects in the creative industries, advertising and entertainment fields.

Diffusers Image Outpaint is an image epitaxy technology based on the diffusion model, which can generate additional parts of the image based on the existing image content. This technology has broad application prospects in image editing, game development, virtual reality and other fields. It uses advanced machine learning algorithms to make image generation more natural and realistic, providing users with an innovative image processing method.

Open-MAGVIT2 is an autoregressive image generation model series open sourced by Tencent ARC Laboratory, including models of different sizes from 300M to 1.5B. This project reproduces Google's MAGVIT-v2 tokenizer and achieves advanced reconstruction performance of 1.17 rFID on the ImageNet 256×256 data set. By introducing asymmetric word segmentation technology, the large vocabulary is decomposed into sub-vocabularies of different sizes, and 'next sub-tag prediction' is introduced to enhance the interaction between sub-tags to improve the quality of generation. All models and code have been made open source, aiming to drive innovation and creativity in the field of autoregressive vision generation.

OmniGen is an innovative diffusion framework that unifies multiple image generation tasks into a single model without the need for task-specific networks or fine-tuning. This technology simplifies the image generation process, improves efficiency, and reduces development and maintenance costs.

ViewCrafter is a novel approach that exploits the generative power of video diffusion models and the coarse 3D cues provided by point-based representations to synthesize high-fidelity new views of universal scenes from single or sparse images. This method gradually expands the area covered by 3D clues and new perspectives through iterative view synthesis strategies and camera trajectory planning algorithms, thereby expanding the range of new perspective generation. ViewCrafter can facilitate various applications such as immersive experiences and real-time rendering through optimized 3D-GS representation, and more imaginative content creation through scene-level text-to-3D generation.

FLUX.1-dev-LoRA-Dark-Fantasy is a LoRA model trained by Shakker AI's GUIZANG, focusing on generating fantasy creatures and characters. Influenced by artists such as Klee, Odilon Redon, and Eyvind Earle, the model is capable of producing images with a cinematic texture, complex light and shadow effects, and fine detail. Model follows flux-1-dev-non-commercial-license and is suitable for non-commercial use.

Dark fantasy FLUX is an AI model focused on generating fantasy creatures and characters. It is good at creating clothing with fluid metallic textures and images with magical or technological light effects. It is capable of producing images with a dark-toned atmosphere without compromising responsiveness to realistic content. This model is licensed from Black Forest Labs, Inc. and is suitable for non-commercial use.

Pixtral-12b-240910 is a multi-modal large-scale language model released by the Mistral AI team, which is capable of processing and understanding image as well as text information. The model uses an advanced neural network architecture that can provide richer and more accurate output results through a combination of images and text input. It shows excellent performance in image recognition, natural language processing and multi-modal interaction, and is of great significance for application scenarios that require simultaneous processing of images and text.

LongLLaVA is a multi-modal large-scale language model that efficiently scales to 1000 images through a hybrid architecture, aiming to improve image processing and understanding capabilities. Through innovative architectural design, this model achieves effective learning and reasoning on large-scale image data, which is of great significance to fields such as image recognition, classification and analysis.

RECE is a concept erasure technique for text-to-image diffusion models, which achieves reliable and efficient erasure of specific concepts by introducing regularization terms during the model training process. This technology is important for improving the security and control of image generation models, especially in scenarios where the generation of inappropriate content needs to be avoided. The main advantages of RECE technology include high efficiency, high reliability and easy integration into existing models.

M&M VTO is a mix-and-match virtual try-on method that accepts multiple images of clothing, a text description of the clothing layout, and a picture of a person as input, and the output is a visualization of these clothes worn on a given person in a specified layout. The main advantages of this technology include: a single-stage diffusion model, without the need for super-resolution cascades, capable of mixing and matching multiple garments at 1024x512 resolution, while retaining and distorting complex garment details; the architectural design (VTO UNet Diffusion Transformer) can separate denoising and character-specific features, achieving an efficient identity-preserving fine-tuning strategy; controlling the layout of multiple garments through text input, specifically fine-tuning for virtual try-on tasks. The M&M VTO achieves state-of-the-art performance both qualitatively and quantitatively and opens up new possibilities for verbal guidance and multi-garment fitting.

FlexClip AI Image to Image Generator is an online image conversion tool that uses advanced AI technology to convert user-uploaded images into different artistic styles. This product ensures high-quality image style conversion through continuously updated AI models, and is suitable for professional and personal use. It also provides rich AI features such as AI text to image, AI text to video, and AI background remover to speed up the photo and video creation process.

VectorJourney is a model that uses AI technology to generate travel-style pictures. Users can generate cartoon-style pictures with travel elements through simple text descriptions. This model is especially suitable for users who want to share their travel experiences on social media without showing their faces. It offers a novel virtual travel experience through an artistic style that blends realistic and illustrative elements.

OmniRe is a comprehensive approach for efficient reconstruction of high-fidelity dynamic urban scenes through device logs. This technology achieves comprehensive reconstruction of different objects in the scene by constructing a dynamic neural scene graph based on Gaussian representation and constructing multiple local specification spaces to simulate various dynamic actors including vehicles, pedestrians and cyclists. OmniRe allows us to fully reconstruct the different objects present in the scene and subsequently implement a simulation of the reconstructed scene with the participation of all participants in real time. Extensive evaluation on the Waymo dataset shows that OmniRe significantly outperforms previous state-of-the-art methods both quantitatively and qualitatively.

CSGO is a text-to-image generation model based on content style synthesis. It generates and automatically cleans stylized data triples through a data construction pipeline, and builds the first large-scale style transfer data set IMAGStyle, which contains 210k image triples. The CSGO model adopts end-to-end training, clearly decouples content and style features, and achieves it through independent feature injection. It implements image-driven style transfer, text-driven style synthesis, and text editing-driven style synthesis. It has the advantages of inference without fine-tuning, maintaining the generation ability of the original text-to-image model, and unifying style transfer and style synthesis.

AWPortrait-FL is an advanced portrait generation model fine-tuned on the basis of FLUX.1-dev. It is trained using the AWPortrait-XL training set and nearly 2000 high-quality fashion photography photos. The model offers significant improvements in composition and detail, producing portraits with more detailed, realistic skin and textures. It is trained by DynamicWang on AWPlanet.

FLUX.1-dev-LoRA-blended-realistic-illustration is an AI image generation model based on LoRA technology, trained by Muertu, focusing on combining cartoon-style characters with realistic backgrounds to create unique mixed reality artistic effects. This model is innovative in the field of image generation and can provide new creative tools for artists and designers, while providing new perspectives for image processing and artistic creation. Model follows flux-1-dev-non-commercial-license and is suitable for non-commercial use.

Dark Gray Photography is an image generation model focused on generating images of dark gray tones and East Asian women. This model is based on LoRA technology and is trained through deep learning to generate images with consistent style and bright colors. It is particularly suitable for users who need to use dark gray tones in portrait, product, architectural and nature landscape photography.

HivisionIDPhotos is a lightweight AI ID photo production tool that uses advanced image processing algorithms to intelligently identify and cut out images to generate ID photos that meet a variety of specifications. The development background of this tool is to quickly respond to users' needs for ID photos on different occasions, and to improve the efficiency and quality of ID photo production through automated image processing technology. The main advantages of the product include lightweight, high efficiency, ease of use and support for multiple ID photo formats.

GenWarp is a model for generating new perspective images from a single image through a semantically preserving generative warping framework that enables text-to-image generative models to learn where to warp and where to warp. This model solves the limitations of existing methods by enhancing cross-view attention and self-attention. It generates a model on the source view image through conditionalization and incorporates geometric deformation signals to improve performance in different field scenarios.

Qwen2-VL is the latest generation visual language model based on Qwen2. It has multi-language support and powerful visual understanding capabilities. It can process pictures of different resolutions and aspect ratios, understand long videos, and can be integrated into mobile phones, robots and other devices for automatic operations. It has achieved world-leading performance in multiple visual understanding benchmarks, especially in document understanding.

DiPIR is a physics-based method jointly developed by the Toronto AI Lab and NVIDIA Research that enables virtual objects to be realistically inserted into indoor and outdoor scenes by recovering scene lighting from a single image. The technology not only optimizes materials and tone mapping, but also automatically adjusts to different environments to improve the realism of images.

Deforum-x-flux is a Deforum implementation based on flux-dev, developed by XLabs-AI. It is an open source image generation model capable of generating highly realistic images through text prompts. This model utilizes the latest artificial intelligence technology, has the ability to generate high-quality images, and can be applied to a variety of scenarios, such as art creation, game design, etc.

Show-o is a single transformer model for multimodal understanding and generation that is capable of handling image captioning, visual question answering, text-to-image generation, text-guided repair and expansion, and mixed-modality generation. This model was jointly developed by Show Lab of the National University of Singapore and ByteDance. It uses the latest deep learning technology and can understand and generate data in multiple modalities. It is a major breakthrough in the field of artificial intelligence.

Kolors Virtual Try-On is a virtual try-on application that combines artificial intelligence and augmented reality technology to generate natural and beautiful try-on effects based on given model images and selected clothes. This product supports the entire process generation from model material pictures to model short videos, meeting the needs of e-commerce model material generation.

dark-fantasy-illustration-flux is a LoRa adapter based on the FLUX1.-dev model, specifically designed to generate images inspired by dark fantasy retro illustrations. It does not require specific trigger words, only natural language prompts to generate images, and is compatible with other LoRa models and suitable for generating images with unique artistic styles.

mPLUG-Owl3 is a multi-modal large-scale language model focused on the understanding of long image sequences. It can learn knowledge from the retrieval system, engage in alternating text and picture conversations with users, and watch long videos to remember their details. The source code and weights of the model have been released on HuggingFace, and are suitable for scenarios such as visual question answering, multi-modal benchmarking, and video benchmarking.

Ideogram 2.0 is a cutting-edge text-to-image model with the ability to generate realistic images, graphic design, typesetting, and more. It is trained from scratch and significantly outperforms other text-to-image models, outperforming multiple quality metrics such as image-text alignment, overall subjective preference, and text rendering accuracy. Ideogram 2.0 also launches an iOS app, bringing the high-end platform into the hands of mobile users and providing developers with technology at a competitive price via API to enhance their apps and workflows.

flux-ip-adapter is an image generation adapter based on the FLUX.1-dev model, developed by Black Forest Labs. The model is trained to support image generation at 512x512 and 1024x1024 resolutions, and new checkpoints are released regularly. It is primarily designed for use with ComfyUI, a user interface design tool that can be integrated via custom nodes. This product is currently in beta testing and may require several attempts to achieve ideal results.

DiffusionKit is an open source project designed to provide native inference capabilities for diffusion models in Apple silicon devices. It achieves efficient image processing capabilities by converting PyTorch models into Core ML format and using MLX for image generation. The project supports Stable Diffusion 3 and FLUX models, capable of image generation and image-to-image conversion.

TurboEdit is a technology developed based on Adobe Research to solve the challenges of precise image inversion and decoupled image editing. It achieves the ability to precisely edit images in a few steps through iterative inversion technology and conditional control based on text prompts. This technique is not only fast, but also outperforms existing multi-step diffusion model editing techniques.

AuraFlow v0.3 is a completely open source flow-based text-to-image generation model. Compared with the previous version AuraFlow-v0.2, the model has been trained with more calculations and fine-tuned on the aesthetic dataset to support various aspect ratios with width and height up to 1536 pixels. This model achieved state-of-the-art results on GenEval and is currently in the beta testing stage. It is being continuously improved and community feedback is very important.

half_illustration is a text-to-image generation model based on the Flux Dev 1 model that combines photography and illustration elements to create artistic images. This model uses LoRA technology, which can maintain style consistency through specific trigger words, and is suitable for use in the fields of art creation and design.

Freepik AI image generator is an online tool that uses artificial intelligence technology to automatically generate images based on text prompts entered by users. It simplifies the image creation process, allowing users to quickly create personalized and creative images even without professional design skills. The application of this technology not only improves design efficiency, but also broadens the boundaries of image creation, providing users with unlimited possibilities.

IPAdapter-Instruct is an image generation model developed by Unity Technologies. It adds additional text embedding conditions to the transformer model, allowing a single model to efficiently perform multiple image generation tasks. The main advantage of this model is the ability to flexibly switch between different conditional interpretations, such as style transfer, object extraction, etc., in the same workflow via 'Instruct' prompts, while maintaining minimal quality loss compared to task-specific models.

Object Images is an innovative 3D model generation technology that simplifies the generation and processing of 3D shapes by encapsulating complex 3D shapes in a 64x64 pixel image, so-called 'Object Images' or 'omages'. This technology solves the challenges of geometric and semantic irregularities in traditional polygonal meshes by using image generation models, such as Diffusion Transformers, directly for 3D shape generation.

FLUX.1-dev-Controlnet-Union-alpha is a text-to-image generative model that belongs to the Diffusers series and uses ControlNet technology for control. The currently released alpha version has not yet been fully trained, but it has demonstrated the effectiveness of its code. This model aims to promote the development of the Flux ecosystem through the rapid growth of the open source community. Although a fully trained Union model may not be as good as a professional model in specific areas such as posture control, its performance will continue to improve as training progresses.

HeadGAP is an advanced 3D avatar creation model that can create realistic and animatable 3D avatars from a small number or even a single picture of a target person. The model learns 3D head prior knowledge by utilizing large-scale multi-view dynamic data sets, and implements dynamic modeling through a self-decoding network based on Gaussian Splatting. HeadGAP learns the properties of Gaussian primitives through identity sharing encoding and personalized latent codes, achieving rapid avatar personalization.

UniPortrait is an innovative portrait personalization framework that enables high-fidelity single-ID and multi-ID portrait customization through two plug-in modules: ID embedding module and ID routing module. The model extracts editable facial features through a decoupling strategy and embeds them into the context space of the diffusion model. The ID routing module adaptively combines and distributes these embedded features to corresponding areas in the synthetic image to achieve single-ID and multi-ID customization. UniPortrait achieves excellent performance in single-ID and multi-ID customization through a carefully designed two-stage training scheme.

LLaVA-OneVision is a multi-modal large-scale model (LMMs) developed by ByteDance in collaboration with multiple universities that pushes the performance boundaries of open large-scale multi-modal models in single image, multi-image and video scenarios. The design of the model allows for powerful transfer learning between different modalities/scenarios, exhibiting new comprehensive capabilities, especially in video understanding and cross-scenario capabilities, demonstrated through image-to-video task conversion.

Flux1.dev-AsianFemale is a LoRA (Low-Rank Adaptation) experimental model based on the Flux.1 D model. It aims to explore how to make the default female image of the Flux model more Asian in appearance through training. This model has not been trained on facial beautification or Internet celebrity faces, and is experimental in nature, and may have some training issues and challenges.

ImageFX is an online image generation tool that uses advanced AI technology to allow users to easily create images with artistic effects. It uses a simple operation interface to allow users to input descriptions or seed values ​​to quickly generate images with a specific style. It is very suitable for designers and artists who need quick creativity and artistic effects.

ai-toolkit is a research GitHub repository created by Ostris and is mainly used for experiments and training of Stable Diffusion models. It contains various AI scripts to support model training, image generation, LoRA extractor, etc. The toolkit is still under development and may be unstable, but offers rich functionality and a high degree of customization.

flux-lora-collection is a series of LoRAs training checkpoints released by the XLabs AI team for the FLUX.1-dev model. This model collection supports image generation in a variety of styles and themes, such as animal anthropomorphism, animation, Disney style, etc., and is highly customizable and innovative.

VFusion3D is a scalable 3D generative model built on pre-trained video diffusion models. It solves the problem of difficulty and limited quantity of 3D data acquisition, generates large-scale synthetic multi-view data sets by fine-tuning the video diffusion model, and trains a feed-forward 3D generation model that can quickly generate 3D assets from a single image. The model performed well in user studies, with users preferring results generated by VFusion3D more than 90% of the time.

ml-mdm is a Python package for efficient training of high-quality text-to-image diffusion models. This model uses Matryoshka diffusion model technology to train a single pixel space model at a resolution of 1024x1024 pixels, demonstrating strong zero-sample generalization capabilities.