The NVIDIA Qwen-Image-Flash model generates images from text prompts using a four-step, DMD2-distilled version of Qwen/Qwen-Image. The distillation used DMD2 from NVIDIA FastGen, NVIDIA Model Optimizer, and NVIDIA AutoModel while retaining the base model architecture. The packaged scheduler is configured for the four-step, shift-3 trajectory.
Developers and researchers evaluating few-step text-to-image generation, including creative-content prototyping and latency-sensitive image-generation workflows. Image editing, image understanding, and safety- or life-critical decision-making are outside the intended scope of this checkpoint.
Number of Model Parameters: 20.43B parameters in the denoising transformer; 28.85B learned parameters in the full pipeline when the text encoder and VAE are included.
The full pipeline contains a Qwen2.5-VL text encoder, Qwen tokenizer, 60-layer Qwen-Image transformer, Qwen-Image VAE, and FlowMatch Euler scheduler. The transformer has the same architecture as the base Qwen-Image transformer; its weights are replaced by the distilled student weights.
Input:
Input Type(s): Text
Input Format(s): String
Input Parameters: One-Dimensional (1D) sequence
Other Properties Related to Input: The release was distilled with English captions. Any inherited Chinese-language capability has not been evaluated for this model. The pipeline allows at most 1,024 prompt tokens and defaults to 512.
Output:
Output Type(s): Image
Output Format: Red, Green, Blue (RGB)
Output Parameters: Two-Dimensional (2D)
Other Properties Related to Output: The pipeline returns generated images, typically as Python Imaging Library (PIL) images. The tested output setting is 1024 × 1024 pixels. Use width and height values divisible by 16 to avoid automatic resizing.
Our AI models are designed and/or optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA hardware and software frameworks such as CUDA and PyTorch, the model can achieve faster inference than CPU-only execution.
The integration of foundation and fine-tuned models into AI systems requires additional testing using use-case-specific data to ensure safe and effective deployment. Following the V-model methodology, iterative testing and validation at both unit and system levels are essential to mitigate risks, meet technical and functional requirements, and ensure compliance with safety and ethical standards before deployment.
Model Version(s):
Qwen-Image-Flash was distilled with NVIDIA Model Optimizer 0.45.0.
Internal Training Dataset Identifier:qwen-image-synthetic-dalle3-1m
Data Modality: Text, Image
Text Training Data Size: [Less than a Billion Tokens]
Image Training Data Size: [Less than a Million Images]
Data Collection Method by dataset: Hybrid: Automated, Synthetic
Labeling Method by dataset: Automated
Properties: Only text prompts from the ProGamerGov dataset were used; its original images were not used for training. Each prompt was passed to the original Qwen-Image model to generate a new synthetic image. The resulting dataset contains 1,000,000 English prompt–image pairs processed at 1024 × 1024, and a subset was used for Distribution Matching Distillation (DMD2) training.
Data Collection Method by dataset: Manually-collected
Labeling Method by dataset: Automated
Properties: Qwen-Image-Bench: 1,000 prompts; OneIG-Bench-EN: 1,120 English prompts across five categories.
Inference:
Acceleration Engine: Hugging Face Diffusers, SGLang Diffusion, vLLM-Omni, and TensorRT-LLM VisualGen
Test Hardware: NVIDIA GB200
DMD2 Distillation
Qwen-Image-Flash was trained using multi-step DMD2 (Distribution Matching Distillation) to approximate the output distribution of Qwen-Image in four denoising steps. The student retains the base transformer's architecture and parameter count. The four-step design reduces the number of transformer evaluations but does not reduce the checkpoint's parameter count or memory requirements.
The teacher target used classifier-free guidance (CFG) 4.0 during distillation. That guidance is internalized by the student, so inference uses true_cfg_scale=1.0 to avoid applying guidance a second time. The packaged static shift-3 FlowMatch Euler scheduler produces effective sigmas [1.0, 0.9, 0.75, 0.5, 0.0] over the required four steps.
1import torch
2from diffusers import QwenImagePipeline
34pipe = QwenImagePipeline.from_pretrained(5"nvidia/Qwen-Image-Flash",6 torch_dtype=torch.bfloat16,7).to("cuda")89image = pipe(10 prompt="A red fox in a snowy pine forest at golden hour, photorealistic, sharp focus, soft bokeh",11 width=1024,12 height=1024,13 num_inference_steps=4,14 true_cfg_scale=1.0,15 guidance_scale=None,16 negative_prompt=None,17 generator=torch.Generator(device="cuda").manual_seed(42),18).images[0]1920image.save("qwen-image-flash-diffusers.png")
* OneIG-Bench-EN publishes no aggregate score; the value shown is the unweighted mean of its five dimensions (alignment, text, reasoning, style, diversity), each scored over the full 1,120-prompt EN set.
Ethical Considerations
NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. Developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.
For more detailed information on ethical considerations for this model, please see the Model Card++ Explainability, Bias, Safety & Security, and Privacy Subcards below.
Please report model quality, risk, security vulnerabilities or NVIDIA AI Concerns here.
SUBCARDS:
Explainability
Field
Response
Intended Task/Domain
Few-step text-to-image generation.
Model Type
DMD2-distilled diffusion transformer (MMDiT) packaged as a full Diffusers pipeline.
Intended Users
Developers and researchers evaluating or integrating latency-sensitive image generation.
Output
Generated RGB image.
Describe how the model works
A Qwen2.5-VL text encoder converts the prompt to conditioning embeddings. A DMD2-distilled Qwen-Image transformer denoises image latents in four deterministic FlowMatch Euler steps, and the VAE decodes the final latents into an image.
Name the adversely impacted groups this has been tested to deliver comparable outcomes regardless of
Not Applicable
Technical Limitations & Mitigation
The model is designed for the packaged four-step schedule; changing the number of inference steps or scheduler trajectory may degrade image quality or produce unexpected results. It was distilled with English captions at 1024 × 1024, so performance with non-English prompts has not been evaluated and may be less reliable. The pipeline accepts other resolutions when the width and height are divisible by 16, but only 1024 × 1024 has been tested and quality may vary at other sizes. The model may inherit biases and visual artifacts from the base model and training data and does not include a safety checker; deployers should conduct use-case-specific testing and add appropriate safeguards.
Verified to have met prescribed NVIDIA quality standards
Yes
Performance Metrics
Image quality, prompt adherence, throughput, and end-to-end latency.
Potential Known Risk
This model can generate synthetic images and may produce content that is inaccurate, offensive, or otherwise inappropriate. Users should implement robust safety guardrails — including content filtering, abuse monitoring, and access controls — to reduce the risk of harmful outputs. Users are responsible for ensuring that their use of the model complies with all applicable laws and regulations, and for regularly reviewing and updating their guardrails as risks evolve.
Additional Information:Apache License 2.0 , and applicable law. Image editing, image understanding, high-stakes decisions, and illegal or abusive content generation are outside the intended scope. |
| Model and Dataset Restrictions | The Principle of least privilege (PoLP) is applied limiting access for dataset generation and model development. Restrictions enforce dataset access during training, and dataset license constraints adhered to. |
Privacy
Field
Response
Generatable or reverse-engineerable personal data
No
Was consent obtained for any personal data used
Not Applicable
Personal data used to create this model
None Known
How often is the dataset reviewed
Before Release
Was data from user interactions with the AI model used to train the model
No
Is there provenance for all datasets used in training
Yes
Does data labeling comply with privacy laws
Yes
Is the data compliant with data-subject requests for correction or removal
