DreamLite, a 0.39 B‑parameter diffusion model from ByteDance, unifies text‑to‑image generation and text‑guided editing in a single on‑device network, delivering 1024×1024 results in about three seconds on an iPhone 17 Pro while surpassing existing mobile and even many server‑side baselines.
DreamLite is a compact 0.39 B unified diffusion model open‑sourced by ByteDance that runs on smartphones, delivering text‑to‑image generation and text‑guided editing in about three seconds for 1024×1024 pictures, with performance comparable to Flux, Z‑Image and LongCat‑Image and offering two variants to balance fidelity and latency.
PromptEcho computes a continuous reward for text‑to‑image generation by measuring how well a frozen vision‑language model can reconstruct the original prompt from the generated image, eliminating the need for annotated data or a trained reward model and outperforming prior methods across multiple benchmarks.
The paper introduces CFG‑Ctrl, a control‑theoretic redesign of classifier‑free diffusion guidance that treats the generation process as a dynamic system, achieving more stable and accurate text‑to‑image results across multiple model scales and evaluation metrics.
The guide explains why GPT Image 2 dramatically reduces low‑quality outputs, outlines five essential prompt elements, provides eight ready‑to‑use scene templates, shares advanced tricks, common pitfalls, and concrete examples to help users generate professional AI images reliably.
The article introduces three open‑source AI tools—a video editor that turns raw footage and a brief into a finished ad, Baidu's 8‑billion‑parameter text‑to‑image model that runs on 24 GB GPUs, and a weekly AI‑developer digest that auto‑generates Chinese reports—detailing their workflows, benchmarks, usage commands, and target users.
Gen-Searcher equips text-to-image models with searchable, reasoning, and web‑browsing capabilities, turning the traditional direct‑generation pipeline into an agentic system that fetches and verifies real‑world knowledge, dramatically improving accuracy and quality across multiple benchmarks.
The ACCORD framework formalizes the concept‑coupling issue in text‑to‑image diffusion models as a statistical dependency problem and resolves it with two plug‑and‑play regularization losses, dramatically improving fidelity and text control without altering model architecture.
Qwen‑Image‑2.0, the latest text‑to‑image model from Tongyi Qianwen, delivers pixel‑perfect 2K text rendering, supports 1K‑token prompts, and combines generation and editing in one model, achieving a score of 1029 and third place in the global AI Arena benchmark, positioning it as an AI‑powered designer for PPTs, posters, infographics, and comics.
This guide walks through designing and implementing an AI assistant that connects FastAPI to Alibaba Cloud large‑model services, supports streaming text Q&A, image understanding, text‑to‑image generation, network search, and MCP‑based map queries, with full front‑end and back‑end code examples.
The article reviews Black Forest Labs' FLUX.2 Klein model, highlighting its sub‑second 1024×1024 image generation, low‑VRAM requirements, four‑step inference speedups, and competitive quality versus SD3 and Midjourney V6, while also sharing Midjourney‑style prompt examples for creative design.
The 2026 AI model leaderboards across Text, Web Development, Vision, and Text-to-Image arenas reveal Google’s Gemini series leading in text and vision, Anthropic’s Claude Opus unexpectedly topping web‑dev rankings, and OpenAI’s GPT‑Image‑1.5 clinching the top spot in creative image generation, highlighting an increasingly competitive AI landscape.
GLM-Image, a Chinese text‑to‑image model trained end‑to‑end on Huawei Ascend 800T A2 NPUs, combines an autoregressive decoder with a diffusion encoder, supports resolutions up to 2048×2048, and offers open‑source code, API access, and detailed prompts that demonstrate its strong layout and typography capabilities.
This guide explains how to use advanced AI models such as Gemini, Midjourney, and Stable Diffusion to expand brief, informal user descriptions into comprehensive, high‑quality English prompts that include visual style, subject details, environment, lighting, and camera parameters for image or video generation.
This article evaluates the Nano Banana Pro text‑to‑image model, built on Gemini 3 Pro, across fourteen low‑level vision tasks and forty datasets using only prompts without fine‑tuning, revealing strong perceptual quality but weak pixel‑level metrics, and highlighting both its generative strengths and failure modes such as hallucinations and color shifts.
Black Forest Labs released Flux.2 Max, the top‑performing model in the Flux.2 series featuring real‑time context generation, superior texture handling, and strong instruction following, ranking second on the Artificial Analysis leaderboard, with detailed examples, API usage, and pricing information provided.
This case study details how the 58.com UX team used AI tools to create a 40‑second pre‑video for the China Good Employer awards, outlining the visual framework, prompt engineering, segmented generation workflow, and the balance between automation and designer insight.
This article reviews the YOLO-Count model, a fully differentiable, open‑vocabulary object counting system that guides text‑to‑image generators to produce the exact number of objects specified in prompts, achieving state‑of‑the‑art results on both generic counting and controlled image synthesis tasks.
YOLO-Count introduces a fully differentiable, open‑vocabulary object counting model that guides text‑to‑image generators to produce the exact number of objects specified in prompts, achieving state‑of‑the‑art performance on both generic counting and controlled image synthesis tasks.
Google's Gemini 2.5 Flash Image, nicknamed Nano‑Banana, tops community leaderboards with a 0.855 score, offers high‑fidelity likeness preservation for editing and generation at about $0.04 per 1024×1024 image, and is demonstrated through scene‑swap, virtual‑try‑on, and text‑to‑image examples.
This guide explains how to write precise text prompts for Google Gemini’s image‑generation model, covering six essential prompt elements, feature overviews, and concrete examples that demonstrate character consistency, targeted edits, creative composition, style transfer, and logical reasoning, while also noting current limitations.
This tutorial walks you through logging into Coze, building a text‑to‑image workflow with start, image generation, and end nodes, testing and publishing it, then generating a personal access token and invoking the workflow via the Coze API to retrieve the image URL.
Orthus, a new unified multimodal model presented at ICML 2025, leverages an autoregressive Transformer backbone with separate language and diffusion heads to enable lossless image‑text interleaved generation, outperforming existing models on both understanding and generation benchmarks while remaining computationally efficient.
Recent large language models have shown strong reasoning abilities, and this work extends chain‑of‑thought reasoning to autoregressive image generation by introducing T2I‑R1, a dual‑level (Semantic‑CoT and Token‑CoT) framework trained with reinforcement learning that unifies high‑level planning and low‑level token generation, achieving state‑of‑the‑art results.
Learn step‑by‑step how to set up a stable, high‑efficiency local text‑to‑image generation service using the Flux model series on Alibaba Cloud’s Bailen platform, integrate it with Cursor IDE’s MCP tool, configure environments, manage API keys, and run the service with sample code and results.
ImageRAG introduces a retrieval‑augmented generation framework that dynamically fetches relevant images to guide diffusion models, dramatically improving the synthesis of rare and fine‑grained concepts across multiple text‑to‑image systems, as demonstrated by extensive quantitative and user studies.
SANA 1.5 introduces an efficient model‑growth pipeline, depth‑pruning, and inference‑time scaling that reuse a 1.6 B‑parameter foundation to train a 4.8 B model with 8× lower memory, 60 % less training time, and GenEval scores that rival or surpass much larger diffusion models.
This guide shows how to use Cloudflare Workers AI's free daily quota to quickly create a custom Flux‑1‑Schnell text‑to‑image API, covering project initialization, AI binding configuration, request validation, error handling, authentication, deployment, and testing with curl.
This article presents a 1.58‑bit quantization of the FLUX.1‑dev text‑to‑image model that reduces 99.5% of its 11.9 B parameters, introduces a custom low‑bit kernel, and achieves storage, memory, and latency improvements while preserving generation quality on standard benchmarks.
The CVPR 2025 NTIRE workshop launches an AI-generated image quality assessment competition featuring two tracks—fine‑grained text‑image matching and structural issue detection—supported by the large‑scale EvalMuse dataset, detailed evaluation metrics, baseline code, and a prize pool of up to $10,000.
SnapGen is a 379 M‑parameter text‑to‑image diffusion model that produces 1024 px images on mobile devices in about 1.4 seconds, using a compact U‑Net design, multi‑stage knowledge distillation, step distillation, and optimized training tricks to outperform much larger models on standard benchmarks.
ArtCrafter introduces a novel text‑image style transfer framework that leverages attention‑based style extraction, text‑image alignment enhancement, and explicit modulation to achieve controllable, diverse, and high‑fidelity visual results, outperforming existing methods in both qualitative and quantitative evaluations.
SnapGen introduces a compact 379M‑parameter diffusion model that produces 1024‑pixel text‑to‑image results in about 1.4 seconds on a mobile device, achieving competitive FID scores and outperforming much larger models through a series of architecture refinements, advanced training tricks, and multi‑level knowledge distillation.
This survey acts as a comprehensive portal that organizes AIGC research across seven domains—text, image, and audio generation, cross‑modal association, text‑guided image and audio synthesis, and supporting resources—detailing seminal models such as GPT, Diffusion, CLIP, DALL·E, Stable Diffusion, MusicLM, and key papers that shaped each field.
This article introduces Ideogram, a free AI image tool that can instantly create high‑quality graphics with integrated text, walks through its simple two‑step workflow, showcases font and poster design examples, compares results with Midjourney, and discusses current limitations and pricing.
Recent ACL 2024 papers from Alibaba Cloud’s PAI platform showcase open‑source Chinese diffusion models, an interactive multi‑turn prompt generator, a long‑tail knowledge‑aware retrieval‑augmented LLM approach, and a dynamic fusion network for sequential model editing, all integrated into cloud services.
The article presents a comprehensive overview of Kuaishou’s Kolors (formerly 可图) multimodal generative model, detailing its data collection strategy, diffusion‑based architecture, evaluation metrics, derived capabilities such as prompt refinement and interactive generation, and a range of practical applications from AI‑powered live‑stream gifts to virtual try‑on, while also offering strategic advice for the domestic visual‑generation community.
This article introduces the Multidimensional Preference Model (MPS), the first multi‑dimensional scoring system for evaluating text‑to‑image generation, built on the newly released MHP dataset with extensive human annotations across aesthetic, semantic alignment, detail quality, and overall preference dimensions, and demonstrates its superior performance through comprehensive experiments and RLHF integration.
Kwai (Kuaishou) has eight papers accepted at CVPR 2024 covering multi‑dimensional human preference for text‑to‑image generation, short‑video quality assessment, efficient video quality assessment, compressed video enhancement, conditional unsigned distance fields, universal cross‑domain retrieval, perception‑oriented frame interpolation, and test‑time energy adaptation.
This article explains how AI‑generated content (AIGC) technologies such as text‑to‑image diffusion models, Lora fine‑tuning, and IP‑Adapter style transfer are applied to interior design, dramatically reducing design time, cutting costs, and enabling personalized, high‑quality visualizations for both consumers and furniture merchants.
This guide walks you through generating eye‑catching clay‑style artwork using Stable Diffusion, covering model selection, prompt engineering, sampling settings, image‑to‑image techniques, and iterative refinements to achieve high‑quality, realistic results.
OPPO details how it deploys multimodal pretrained models on resource‑constrained edge devices by compressing CLIP‑based image‑text retrieval, adapting Chinese text‑to‑image generation with LoRA and adapters, and lightweighting diffusion models through layer pruning and progressive distillation, achieving sub‑3‑second generation while preserving cloud‑level quality.
SDXL‑Lightning, a new diffusion‑based text‑to‑image model from ByteDance, uses Progressive Adversarial Distillation to cut inference steps to as few as 2 while maintaining high resolution and fidelity, offering ten‑fold speed gains, open‑source access, and compatibility with SDXL, ControlNet, and ComfyUI.
This article explains how contextual learning can enhance text‑to‑image models by incorporating example image‑text pairs, redesigning the UNet architecture, building large in‑context training datasets, and training the SuTI model to achieve fast, controllable, and high‑quality personalized image generation.
In this interview, Baidu's search architect Tianbao explains the evolution of text‑to‑image generation since 2022, discusses data preparation, model quality, prompt engineering, multi‑style support, evaluation methods, and predicts when fully AI‑generated video and movies might become mainstream.
In a recent InfoQ Geek Talk, Baidu Search chief architect Tianbao discussed the rapid evolution of AI text‑to‑image technology—highlighting Chinese‑language data preparation, prompt‑engineering challenges, evaluation methods combining human feedback and metrics, and future video‑generation prospects—while announcing openings for visual algorithm engineers.
This guide introduces Tencent Hunyuan's upgraded text‑to‑image model, explains its technical innovations, provides detailed prompt engineering advice, showcases example prompts and generated images across various styles, and highlights real‑world applications and performance metrics for developers and creators.
The article demystifies Stable Diffusion by explaining its low‑cost latent‑space design and conditioning mechanisms, comparing it to autoregressive, VAE, flow‑based and GAN models, detailing the iterative noise‑to‑image process, token‑based text‑to‑image control, version differences, common generation issues, and providing implementation code examples.
Rapid Diffusion introduces a knowledge‑enhanced, high‑speed Chinese text‑to‑image diffusion model with one‑click deployment, achieving superior image quality and up to 1.73× faster inference through FlashAttention and BladeDISC optimizations, and demonstrates strong performance across e‑commerce, traditional painting, and food datasets.
QQGC, Tencent’s two‑stage text‑to‑image model that separates CLIP‑based Prior mapping from a Stable Diffusion Decoder, leverages T5‑enhanced text embeddings and a suite of efficiency tricks—including FP16, flash attention, ZeRO and GPU‑RDMA—to train over‑2 B‑parameter models on 64 GPUs, achieving state‑of‑the‑art FID and CLIP scores while supporting image variation, semantic img2img, precise CLIP‑vector edits and unsafe‑content filtering, and now powers the company’s Magic Painting Room.
This article provides a comprehensive overview of Stable Diffusion, covering its latent diffusion architecture, training data and procedures, model components such as autoencoder, CLIP text encoder and UNet, as well as practical usage examples including text‑to‑image generation, image‑to‑image, inpainting, and advanced extensions like ControlNet and SD‑2.x.
This article surveys the key challenges of controllable text‑to‑image generation and explains the architectures, components, and training details of major diffusion‑based models such as DALL‑E‑2, Google Imagen, Stability AI's Latent Stable Diffusion, and the ControlNet extension.
This article reviews the key components and design choices of recent text‑controlled image generation systems—including DALL·E‑2, Google Imagen, Stability AI's Latent Stable Diffusion, and the ControlNet extension—highlighting how diffusion models, text encoders, prior modules, super‑resolution, and conditioning mechanisms enable high‑quality, controllable visual synthesis.
This guide explores how to craft effective prompts for Stable Diffusion image creation, ChatGPT article writing, and GitHub Copilot code generation, covering prompt evolution, negative prompts, ControlNet enhancements, model selection, and practical tips for iterative refinement and context building.
The article surveys AI‑generated art, explaining how GANs’ limitations gave way to diffusion models and the open‑source Stable Diffusion platform, which offers text‑to‑image, img2img, inpainting, DreamBooth fine‑tuning, and widespread commercial and DIY deployments via cloud or local WebUI setups.
Google’s new Muse model, a Transformer‑based text‑to‑image system running on TPUv4, claims to generate 256×256 images in 0.5 seconds—far faster than Imagen—while delivering unprecedented photorealism and deep language understanding through parallel decoding and large‑scale LLM‑conditioned training.
This article introduces Alibaba Cloud’s open‑source PAI‑Diffusion series, detailing its Latent Diffusion Model foundation, Chinese CLIP alignment, super‑resolution components, and showcases diverse artistic and real‑world text‑to‑image generation scenarios, while providing guidance on accessing the models via Alibaba Cloud AI Center, PAI‑DSW, and HuggingFace Space.
The article surveys the evolution of AI‑generated content from early art programs to modern diffusion‑based text‑to‑image and text‑to‑video models, outlines key milestones such as Stable Diffusion and DALL‑E 2, explores gaming applications, and highlights limitations, ethical concerns, and copyright risks of open‑source generative AI.
This article introduces EasyNLP’s newly integrated Chinese text‑to‑image generation framework, explains the underlying Transformer‑VQGAN architecture, provides model specifications, code snippets, performance benchmarks on multiple datasets, and step‑by‑step tutorials for fine‑tuning and inference using open‑source checkpoints.
This article introduces EasyNLP's Chinese text‑to‑image generation framework, explains the underlying Transformer‑VQGAN architecture, provides model specifications, showcases sample outputs, and offers step‑by‑step code and command‑line instructions for fine‑tuning and inference.
This guide introduces four free web‑based utilities—hexed.it for quick hex editing, text2image for converting text into images, Carbon for turning code snippets into stylish pictures, and draw.io for creating professional diagrams—highlighting their key features, interfaces, and practical use cases for developers.
Jack Cui presents three recently open‑sourced CVPR 2022 algorithms—DualStyleGAN for high‑resolution portrait style transfer, RQ‑VAE Transformer for improved text‑to‑image generation, and VFD for deep‑fake detection—detailing their functionality, usage options, and providing links to code repositories and demo platforms.
This article introduces the Seq‑SG2SL framework, which tackles the challenge of granting AI visual imagination by converting scene graphs into semantic layouts, discusses the limitations of existing text‑to‑image methods, proposes the SLEU metric for automatic evaluation, and presents experimental results demonstrating its effectiveness.
