FLUX.1-schnell-amdnpu

🚀 FLUX.1 [schnell] on AMD AI PC NPU

FLUX.1 [schnell] is a fast, distilled text-to-image model from the FLUX.1 family, designed for rapid generation with few inference steps. For more information about how FLUX.1 [schnell] functions, please see the Black Forest Labs documentation.

More details about this model can be found on the original Hugging Face model card: black-forest-labs/FLUX.1-schnell.

This model repo contains the optimized ONNX models required to run the image generation pipeline for FLUX.1 [schnell] on AMD NPUs.

Model Details

The folder structure is organized to mirror the main components of the diffusion pipeline (scheduler, text encoders, tokenizers, transformer, and VAE decoder).

├─ scheduler/
├─ text_encoder/
├─ text_encoder_3_gptq_v2/
├─ tokenizer/
├─ tokenizer_2/
├─ transformer/
└─ vae_decoder/

The scheduler folder contains the scheduler configuration (timesteps, betas, alphas, etc.) used during the diffusion sampling process.

The text_encoder folder contains the CLIP text encoder model used to convert the input prompt into conditioning embeddings.

The text_encoder_3_gptq_v2 folder contains the quantized T5 text encoder (GPTQ v2) used for additional prompt conditioning in the FLUX architecture.

The tokenizer and tokenizer_2 folders contain the tokenizer configuration and vocabulary files required to preprocess text prompts for the respective text encoders.

The transformer folder contains the flow-matching transformer model used in the diffusion process. The transformer is exported and structured specifically to leverage the AMD NPU accelerator for the denoising steps.

The vae_decoder folder contains the VAE decoder model used to map latent representations back to the image space. The VAE decoder is also structured to make use of the NPU accelerator for efficient image reconstruction.

Note: Transformer and VAE decoder models are optimized and structured to run on AMD NPUs. The other components (text encoders, tokenizers and scheduler) are shared between GPU and NPU pipelines, but are provided here for completeness.

⚡ Intended Use

Getting Started

To get started with this model, visit github.com/amd/sd-sandbox.

⚓ Ethical Considerations

AMD is committed to conducting our business in a fair, ethical and honest manner and in compliance with all applicable laws, rules and regulations. You can find out more at the AMD Ethics and Compliance page.

⚠️ Caveats and Recommendations

Please visit the original model card for more details: black-forest-labs/FLUX.1-schnell.

📌 Citation Details

Please cite the original FLUX.1 [schnell] model and refer to Black Forest Labs for official citation guidance.