Easy Diffusion (formerly cmdr2 UI) is a free program that installs Stable Diffusion with a single download and gives you a clean browser interface: prompt, a few sliders, a queue, image modifiers you can click, live preview and face fixing. It has none of the complexity of node graphs and does not require Python or Git to be installed first.

A good first app for beginners on any platform, and for low-end machines, since it can run on CPU and on 4 GB cards. If you grow out of it, Forge or ComfyUI pick up where it stops.

Tiada pemasangan diperlukan - jalankan model AI di awan

BitVector Prism ialah cara paling mudah untuk bermula: pilih model, taip prompt dan jana dalam aplikasi web yang bersih, tanpa apa-apa untuk dikonfigurasi. BitVector juga tersedia di Discord dan di web (SpyGlass).

Cuba BitVector

Apa yang anda perlukan

Windows 10 or 11, Linux or macOS. NVIDIA GPU with 4 GB VRAM or more recommended; AMD on Linux; Apple Silicon supported; it also runs (slowly) on CPU. 8 GB system RAM and about 25 GB of free disk space including the default model.

Cara mendapatkan dan memasang Easy Diffusion

Download the installer for your system from easydiffusion.github.io.
Windows: run the downloaded file and let it extract to a folder without spaces (for example C:\EasyDiffusion). Mac and Linux: unzip and run the start script.
Start Easy Diffusion with Start Stable Diffusion UI.cmd (Windows) or start.sh. The first start downloads the required files and a default model and opens http://localhost:9000.
Put extra checkpoints in models/stable-diffusion, LoRAs in models/lora, VAEs in models/vae and embeddings in models/embeddings; they appear in the dropdowns after a refresh.

Muat turun hanya daripada pautan rasmi di halaman ini. Pemasang yang ditemui di tempat lain mungkin lapuk atau telah diubah suai. ModelTrends.ai tidak menghoskan sebarang fail.

Tiada pemasangan diperlukan - jalankan model AI di awan

PirateDiffusion hanya di Telegram dan dibina untuk profesional: beribu-ribu model, LoRA dan aliran kerja yang dipacu oleh arahan sembang, dengan penjanaan tanpa had pada pelan harga tetap.

Cuba PirateDiffusion

Cara menggunakan Easy Diffusion

Type a prompt, optionally a negative prompt, and press Make Image. The image renders with a live preview in the queue.
Open Image Settings to pick the model, the size, the steps, the guidance scale, the sampler and the number of images.
Click Image Modifiers to add style words (artists, mediums, moods) to your prompt by clicking tiles.
Pick a LoRA in the Image Settings and add its trigger words from the model page; use the strength slider to tune its effect.
Drop an image onto the initial image box for image to image or inpainting; enable Fix Faces (GFPGAN) and Upscale (Real-ESRGAN) under the image.

Petua dan masalah biasa

Turn on Low VRAM mode in the settings on 4 GB cards; it is slower but avoids out of memory errors.
Use the beta channel in the settings to try new features, or stay on main for stability.
Prompts with | create several variations in one go; each gets its own image.
All settings are stored per browser; use the Save Settings option so they persist.

Model yang boleh dijalankan dalam Easy Diffusion

Easy Diffusion runs Stable Diffusion 1.5 and SDXL checkpoints, LoRAs and embeddings from this catalog; newer versions also load Flux models. It is not suited for video models.

Model Stable Diffusion 1.5

Model Stable Diffusion XL 1.0

Model Flux

Tiada pemasangan diperlukan - jalankan model AI di awan

Tiada GPU, tiada muat turun, tiada Python. Perkhidmatan ini menjalankan model sumber terbuka yang sama pada pelayan mereka sendiri, jadi anda boleh mula menjana dari pelayar, Discord atau Telegram dalam seminit.

Cuba BitVector

Cuba PirateDiffusion

Pautan

Pautan ini menuju ke laman pembuat sendiri. ModelTrends.ai tidak bergabung dengan mereka dan tidak menghoskan sebarang perisian.

Aplikasi tempatan lain

Cara memasang ComfyUI

The node-based powerhouse for Stable Diffusion, Flux, Wan video and almost every new open source model. Free, open source, for Windows, Mac and Linux.