Посмотреть как Markdown

Speridlabs Iris-Z

Speridlabs Iris-Z is FLUX.2 fine-tuned to denoise an image and its depth in one sequence. The same weights generate images from a prompt, estimate depth from a photo in 4 steps, paint a prompt onto a given depth, and return an image together with its depth. LibreYOLO runs it through the LibreGen factory in two sizes.

Задачи
image generation, depth
Размеры
9b, 32b at 1024 px
Установка
pip install "libreyolo[gen]"
Уровень поддержки
Смежный уровень, начиная с v1.8.0. Отдельная часть продукта со своей фабрикой и контрактом.
Исходный проект
Speridlabs Iris-Z от Speridlabs, FLUX Non-Commercial License. Статья, исходный код
Лицензии
Код: MIT and Apache-2.0, веса: FLUX Non-Commercial License. Коммерческое использование

Install

bash
pip install "libreyolo[gen]"

Predict

The task is fixed when the model loads: task="depth" for image to depth, the default generate for the other three modes. A depth model never loads a text encoder.

Image to depth
from libreyolo import LibreGen, SAMPLE_IMAGE model = LibreGen("speridlabs-iris-z", task="depth")   # 9b; loads no text encoderresult = model(SAMPLE_IMAGE) depth = result.depth_mapprint(depth.encoding)          # "log_depth": larger is fartherprint(depth.data.shape)        # (H, W) at the photo's sizeprint(depth.valid_mask.shape)  # False on skyresult.save("depth.png")       # coloured depth
Text to image
from libreyolo import LibreGen model = LibreGen("speridlabs-iris-z")result = model("a lighthouse at dusk", seed=0)result.generated.save("lighthouse.png")
Depth to image
from libreyolo import LibreGen, SAMPLE_IMAGE model = LibreGen("speridlabs-iris-z")# The photo's depth is estimated first, then a new scene is painted on it.result = model("a sunken ruin under the sea", source=SAMPLE_IMAGE, seed=0)result.generated.save("ruin.png")
Image plus depth
from libreyolo import LibreGen model = LibreGen("speridlabs-iris-z")result = model("a grand old library, a spiral staircase", depth=True, seed=0)result.generated.save("library.png")print(result.depth_map.data.shape)
CLI
libreyolo generate prompt="a lighthouse at dusk" model=speridlabs-iris-zlibreyolo generate source=https://raw.githubusercontent.com/LibreYOLO/libreyolo/release/libreyolo/assets/guggenheim-bilbao.jpg task=depthlibreyolo generate source=photos/ task=depth ensemble=10 imgsz=1024

Image to depth. Sources are what predict takes. result.depth_map is a DepthMap with encoding="log_depth" (relative log depth, larger is farther) at the source's size, and valid_mask is False on sky. imgsz is the working resolution, 512 or 1024 (default 512 under half a megapixel, else 1024). Knobs: steps (4) and ensemble (1), the per-pixel median over that many seeds. The default seed is 0, so depth repeats. save=True writes the coloured map under runs/depth/predict.

Text to image. Prompts, imgsz, seed and n work as for every image generation model. Knobs: steps (50) and guidance (4.0 for 9b, 3.5 for 32b).

Depth to image. source= on the generate task is the depth the image is painted on: a photo (its depth is estimated first), a Results or DepthMap from any depth model, or a grayscale depth image with depth_format="depth" (bright is far) or "disparity" (bright is near). An image is only read as a depth map when depth_format says so. The output keeps the depth's aspect ratio, and Results.depth_map is the depth it was painted on.

Image plus depth. depth=True samples image and depth jointly, then re-estimates the depth from the finished image, as upstream does by default. The result carries both result.generated and result.depth_map.

The CLI covers text to image and image to depth; prompt= and source= stay mutually exclusive there.

Variants

SizeAliasParametersText encoderGPU memory (library figure)
9bsperidlabs-iris-z, -9b, -small9.1B, from FLUX.2 [klein] Base 9BQwen3 8B FP8about 18 GB, plus 9 GB text encoder and 0.2 GB VAE
32bsperidlabs-iris-z-32b, -big32.2B, from FLUX.2 [dev]Mistral Small 3.2 24Babout 64 GB, plus 48 GB text encoder and 0.2 GB VAE

offload=True keeps the text encoder on the CPU between prompts. On CUDA without the FP8 kernel that transformers asks for (the kernels package), and on MPS or CPU, the 9b text encoder runs dequantized in bfloat16, close to but not identical with upstream's embeddings. A local snapshot or variant folder loads with LibreGen(path).

Export

ЗадачаONNXTorchScriptExecuTorchTensorRTOpenVINOPaddleMNNRKNNAxelerancnnTFLiteCoreMLCore AI
Image generationImage generation to ONNX: не поддерживаетсяImage generation to TorchScript: не поддерживаетсяImage generation to ExecuTorch: не поддерживаетсяImage generation to TensorRT: не поддерживаетсяImage generation to OpenVINO: не поддерживаетсяImage generation to Paddle: не поддерживаетсяImage generation to MNN: не поддерживаетсяImage generation to RKNN: не поддерживаетсяImage generation to Axelera: не поддерживаетсяImage generation to ncnn: не поддерживаетсяImage generation to TFLite: не поддерживаетсяImage generation to CoreML: не поддерживаетсяImage generation to Core AI: не поддерживается
DepthDepth to ONNX: не поддерживаетсяDepth to TorchScript: не поддерживаетсяDepth to ExecuTorch: не поддерживаетсяDepth to TensorRT: не поддерживаетсяDepth to OpenVINO: не поддерживаетсяDepth to Paddle: не поддерживаетсяDepth to MNN: не поддерживаетсяDepth to RKNN: не поддерживаетсяDepth to Axelera: не поддерживаетсяDepth to ncnn: не поддерживаетсяDepth to TFLite: не поддерживаетсяDepth to CoreML: не поддерживаетсяDepth to Core AI: не поддерживается

export(), train() and val() raise, as for every LibreGen family.

Licensing

Проверяйте лицензию в репозитории конкретных весов на Hugging Face. Она указана для каждого чекпойнта в организации LibreYOLO и может различаться даже внутри одного семейства. Этот репозиторий считается авторитетным источником, а сводка ниже описывает условия на момент последней проверки страницы.

Это описание соответствующих лицензий, а не юридическая консультация. Если ответ важен для коммерческого использования, прочитайте лицензии самостоятельно и проконсультируйтесь с юристом.

Оригинальная работа
Speridlabs Iris-Z, Speridlabs
Лицензия исходного проекта
FLUX Non-Commercial License
Исходный код проекта
github.com/speridlabs/iris-z-sdk
Код LibreYOLO
MIT
Веса
FLUX Non-Commercial License, распространяются авторами. LibreYOLO не размещает и не зеркалирует их.
Толкование
The Iris-Z weights are derived from FLUX.2 [dev] and FLUX.2 [klein] Base 9B and are under Black Forest Labs' FLUX Non-Commercial License, a custom agreement that is not OSI-approved: no commercial or production use, and Section 2(e) asks for content filtering or human review of generated images before they are shared. LibreYOLO's adapter is MIT; the transformer, samplers and prompt embedders it ports from the Speridlabs Iris-Z SDK keep Apache-2.0, as does the FLUX.2 code the SDK builds on, with their notices in libreyolo/models/gen/speridlabs_irisz/NOTICE. LibreYOLO does not redistribute any weights. They download at runtime, unmodified, from their publishers at pinned revisions: speridlabs/iris-z at 16239e1a5444c484b11a04b61d633eceb768574c (9b from Iris-Z-small, 32b from Iris-Z-big, not gated), the FLUX.2 VAE from black-forest-labs/FLUX.2-klein-4B (Apache-2.0), and the text encoders Qwen/Qwen3-8B-FP8 for 9b and mistralai/Mistral-Small-3.2-24B-Instruct-2506 for 32b (both Apache-2.0).

Speridlabs Iris-Z is by Speridlabs. The weights are under Black Forest Labs' FLUX Non-Commercial License: no commercial or production use, and section 2(e) asks for content filtering or human review of generated images before they are shared. LibreYOLO's adapter is MIT; the transformer, samplers and prompt embedders ported from the Iris-Z SDK keep Apache-2.0.

LibreYOLO does not redistribute any weights. They download at runtime, unmodified, from speridlabs/iris-z at a pinned revision, not gated. The FLUX.2 VAE comes from the Apache-2.0 black-forest-labs/FLUX.2-klein-4B repository, and the text encoders (Qwen/Qwen3-8B-FP8 for 9b, mistralai/Mistral-Small-3.2-24B-Instruct-2506 for 32b) are Apache-2.0.

Citation

@article{canada2026irisz,
  title   = {Iris-Z: Scaling One Generative Model Past
             Specialist Geometry Models},
  author  = {Ca{\~n}ada, Alejandro and Garabito, Chema},
  journal = {arXiv preprint arXiv:2611.16010},
  year    = {2026},
  url     = {https://arxiv.org/abs/2611.16010}
}

Скопировано из блока цитирования авторов на странице speridlabs.com/research/iris-z#citation.

Проверено с LibreYOLO v1.8.0. Таблицы поддержки, чекпойнты и результаты бенчмарков на этой странице сгенерированы из выпущенной библиотеки и опубликованных весов, а не написаны вручную.