Views
No views yet
pip install -r requirements.txt1# See the full inference script at: https://github.com/NVlabs/PixelDiT
2cd t2i/
3python inference.py \
4 --config configs/PixelDiT_1024px_pixel_diffusion_stage3.yaml \
5 --model_path PixelDiT-T2I-v1.pth \
6 --txt_file prompts.txt \
7 --custom_height 1024 --custom_width 1024 \
8 --cfg_scale 2.75 --seed 2025 \
9 --negative_prompt "low quality, worst quality, over-saturated, blurry, deformed, watermark" \
10 --work_dir "."| Parameter | Default | Description |
|---|---|---|
--cfg_scale | 3.5 | Classifier-free guidance scale |
--step | 50 | Number of sampling steps (25 for fast, 50 for quality) |
--seed | 0 | Random seed |
--negative_prompt | "" | Negative prompt for CFG |
--interval_guidance | [0, 1] | CFG application interval |
--sampling_algo | flow_dpm-solver | Sampling algorithm |
| Component | Value |
|---|---|
| Parameters | 1.3B |
| Patch size | 16 |
| Hidden size | 1536 |
| Attention heads | 24 |
| Patch-level depth | 14 |
| Pixel-level depth | 2 |
| Pixel hidden size | 16 |
| Pixel attention hidden size | 1152 |
| Text embedding dim | 2304 |
| Text max length | 300 |
| Text encoder | Gemma-2-2B-IT |
1@inproceedings{yu2026pixeldit,
2 title={PixelDiT: Pixel Diffusion Transformers for Image Generation},
3 author={Yongsheng Yu and Wei Xiong and Weili Nie and Yichen Sheng and Shiqiu Liu and Jiebo Luo},
4 booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
5 year={2026},
6}