Nitro-AR is a compact masked autoregressive (AR) model for high-quality text-to-image generation. With just ~0.3B parameters, Nitro-AR matches the performance of diffusion-based counterparts while significantly reducing inference latency. By combining continuous token prediction with efficient sampling strategies, it achieves diffusion-level image quality with ultra-low latency. On a single AMD Radeon™ RX 7900 XT GPU, Nitro-AR can generate a 512×512 image in as little as 0.23 seconds (standard) or 74ms (optimized). The release consists of:
Nitro-AR-512px-GAN.safetensors: A masked autoregressive model with a diffusion MLP head, optimized via adversarial fine-tuning for few-step diffusion head sampling.
Nitro-AR-512px-Joint-GAN.safetensors: A high-performance variant using Joint Sampling (transformer-based head) and adversarial fine-tuning, enabling single-step AR generation with superior consistency.
Model architecture: Nitro-AR is built on the E-MMDiT (Efficient Multimodal Diffusion Transformer) backbone, adapted to a masked modeling autoregressive framework. Unlike traditional AR models that predict discrete tokens, Nitro-AR employs a diffusion prediction head to sample continuous tokens, ensuring compatibility with VAE models.
To achieve high efficiency and quality, we introduce several key optimizations:
Joint Sampling: Replaces the standard MLP head with a small transformer head to model token dependencies, significantly improving coherence in few-step generation.
Global Token: Enhances diversity by using a global token as the initial state for masked tokens.
Adversarial Fine-Tuning: Treats the diffusion prediction head as a generator trained against a discriminator, compressing the diffusion process to just 3–6 denoising steps.
Dataset: Our models were trained on a dataset of ~25M images consisting of both real and synthetic data sources that are openly available on the internet. We make use of the following datasets for training: Segment-Anything-1B, JourneyDB, DiffusionDB and DataComp as prompt of the generated data.
Training cost: The models were trained using AMD Instinct™ MI325X GPUs. The efficient architecture allows for rapid convergence comparable to the Nitro-E family.
Quickstart
Environment
When running on AMD Instinct™ GPUs, it is recommended to use the public PyTorch ROCm images for optimized performance.
Download the checkpoints (Nitro-AR-512px-GAN.safetensors and Nitro-AR-512px-Joint-GAN.safetensors) from Hugging Face and place them in the ckpts/ directory.
We support two optimized model variants: gan and joint_gan.
python
1from core.tools.inference_pipe import init_pipe
2import torch
34device = torch.device('cuda')5dtype = torch.bfloat16
67# Initialize pipeline for 'gan' model (Adversarial Fine-tuned, 3 steps)8pipe = init_pipe(device, dtype, model_type='gan')910# Run inference11prompt ="a photo of a dog, with a white background"12output = pipe(prompt=prompt)13output.images[0].save("output_gan.png")1415# Initialize pipeline for 'joint_gan' model (Joint Sampling + Adversarial Fine-tuned, 6 steps)16pipe_joint = init_pipe(device, dtype, model_type='joint_gan')1718# Run inference19output_joint = pipe_joint(prompt=prompt)20output_joint.images[0].save("output_joint_gan.png")
License
Copyright (C) 2026 Advanced Micro Devices, Inc. All Rights Reserved.
