Ming-Image

Deploy Ming-Image Design and Design-Layer with SGLang Diffusion for text-to-image, RGBA image editing, and ordered transparent layer decomposition.

1. Quick start

Use a nightly Docker image containing this integration, or install from source, following the SGLang Diffusion installation guide. Run the commands below inside that environment.

Choose Design for generation or editing and Design-Layer for decomposition. The two checkpoints require separate servers. Start with one H200 on Linux/CUDA; the picker marks unverified hardware and feature combinations explicitly.

Requests return base64 PNGs, preserving all four RGBA channels. For decomposition, each returned image is one ordered layer, not an independent generation. The internal composite frame is omitted from the response. Save all entries in data, rather than only the first image. n selects independent generations (or sets of layers), processed sequentially.

2. Model capabilities

Ming-Image Design generates images from text and edits a single reference image. It targets design assets, layouts, and transparent RGBA output. Use the prompt prefix RGBA, 4-channel, transparent background to request transparency; RGBA output alone does not guarantee a transparent background.

Design-Layer decomposes one reference into ordered transparent layers. Supply num_layers, or specify a count with into N layers or Number of layers: N in the prompt; the prompt count takes precedence. Defaults are 12 steps with guidance 1 for Design and guidance 2 for Design-Layer. The model uses zero unconditional embeddings and does not support custom negative prompts.

Both checkpoints use the native Bailing multimodal MoE encoder, query connector, joint DiT, and four-channel Qwen-style VAE. No remote modeling code is executed. The original checkpoint directory layout is accepted without conversion.

3. Offline generation

Design

sglang generate --model-path inclusionAI/Ming-Image-0.1-Design \
  --prompt "A minimalist exhibition poster with a red geometric sculpture" \
  --height 1024 --width 1024 --save-output

Layer decomposition

sglang generate --model-path inclusionAI/Ming-Image-0.1-Design-Layer \
  --image-path /path/to/input.png --num-layers 4 --save-output

Design defaults to 2048 by 2048; 1024 by 1024 reduces latency and activation memory. Decomposition defaults to the official 1024-resolution aspect buckets; use --height 512 --width 512 for the 512 buckets. Reference-image requests select the closest official bucket and restore the input aspect ratio in the returned PNGs. Multiple reference images are not supported.

For a renamed local checkpoint directory, also provide --model-id inclusionAI/Ming-Image-0.1-Design or the corresponding Layer ID.

4. Runtime features

DiT parallelism reuses the native TP and sequence-parallel attention layers. The DiT has 30 attention heads: TP multiplied by Ulysses must divide 30. Encoder folding has its own head constraints; two-way folding is compatible with the 16-head language tower, 16-head vision tower, and 12-head connector. Do not infer encoder compatibility from the DiT’s head count.

Use the standard component residency controls for transformer, text_encoder, and vae. Layerwise offload trades host memory and transfer time for lower VRAM. The default eager path does not use approximate caching or quantized attention. Cache-DiT and SageAttention require workload-specific quality validation. VAE tiling is disabled by default to match the official inference path. Enable --vae-tiling true when decode memory is constrained; tiling can change pixels.

Breakable CUDA graphs preserve Ming’s checkpoint-specific padding and RoPE offsets. They capture exact prompt shapes; an uncaptured shape runs eagerly. Requests with different prompts are not merged into a dynamic batch. A fixed seed does not imply bit-identical output across attention backends, GPU types, or parallel topologies.

Full-checkpoint validation covers generation, editing, and layer decomposition on H200, including repeated HTTP requests before and after warmup. Tested resolutions include 2048-square generation and four-layer 1024-square decomposition, plus repeated 512-square HTTP generation and editing with one or two sequential generations per request.

The validated runtime scope is:

• One H200: DiT and encoder layerwise offload, opt-in VAE tiling, Cache-DiT, SageAttention, breakable CUDA graph replay, and dynamic LoRA loading/removal with a synthetic adapter.
• Two H200s: DiT TP2, Ulysses2, Ring2 with FlashAttention, CFG parallelism for Design-Layer, encoder folding, and spatial VAE decoding.

These are functional checks, not quality-equivalence claims for lossy optimizations. Other GPU families and quantized checkpoints remain unverified.

See performance optimization for shared runtime controls and the official repository for model examples.

5. Run in ComfyUI

Use the generic image node for Design generation. The ComfyUI integration has not been validated for ordered multi-layer output; use the HTTP API above for Design-Layer decomposition.