Image Generation¶

Architectures¶

• MMDiT - Multi-Modal Diffusion Transformer architecture
• Flow Matching - Flow matching for diffusion models
• Block Causal Linear Attention - Block causal linear attention mechanism
• DC AE - Deep Compression Autoencoder
• SANA - SANA architecture
• SANA Denoiser Architecture - SANA denoiser design
• Transformers v5 - Transformers v5 for diffusion

FLUX Models¶

• FLUX Klein 9B Inference - FLUX Klein 9B inference guide and best practices
• FLUX Kontext - FLUX Kontext model

Training & Fine-tuning¶

• Diffusion LoRA Training - LoRA training for diffusion models
• LoRA Fine Tuning for Editing Models - LoRA fine-tuning for editing models
• Text to LoRA - Text-to-LoRA generation
• Paired Training for Restoration - Paired training for image restoration

Inference & Optimization¶

• Diffusion Inference Acceleration - Inference acceleration techniques
• Tiled Inference - Tiled inference for high-resolution generation
• Temporal Tiling - Tiles as temporal sequence
• Low VRAM Inference Strategies - Low-VRAM inference strategies
• Textual Latent Interpolation - Textual latent interpolation

Editing & Restoration¶

• Step1X Edit - Step1X-Edit model
• [[ACE++]] - ACE++ editing
• LaMa - Large Mask Inpainting
• Image Restoration Survey - Image restoration survey
• RealRestorer - RealRestorer model
• Color Checker and White Balance - Color checker and white balance correction
• grayscale overlay nn architectures - Neural networks for grayscale overlay prediction

Specialized Models¶

• Calligrapher - Calligrapher model
• PixelSmile - PixelSmile model
• X Dub - X-Dub model
• FLAIR - FLAIR model
• MACRO - MACRO model
• MARBLE - MARBLE model
• ATI - Any Trajectory Instruction

Segmentation¶

• in context segmentation - In-context segmentation