Perception Encoder

Perception Encoder maps images, text and finite video clips into a shared embedding space.

Tasks: classification, embeddings
Install: pip install "libreyolo[clip]"
Support tier: Sibling tier, since v1.6.0. A separate product surface with its own factory and contract.
Upstream: Perception Encoder by Meta AI, Apache-2.0. Paper, source
Licenses: Code MIT, weights Apache-2.0. Commercial use

Install

bash

pip install "libreyolo[clip]"

Predict

Python

from libreyolo import LibreYOLO, SAMPLE_IMAGE model = LibreYOLO("LibrePEt16-cls.pt", device="cpu")model.set_classes(["person", "building"])result = model(SAMPLE_IMAGE)print(result.probs)

Set classes before zero-shot classification. Select task="embed" when loading for embeddings. Video embedding samples clip_frames=8 frames by default, averages their vectors and L2-normalizes the result. Training is not supported. Exported video graphs take direct runtime input.

Export

Task	ONNX	TorchScript	ExecuTorch	TensorRT	OpenVINO	Paddle	MNN	RKNN	ncnn	TFLite	CoreML	Core AI
Classification	Classification to ONNX: supported	Classification to TorchScript: supported	Classification to ExecuTorch: supported	Classification to TensorRT: supported	Classification to OpenVINO: supported	Classification to Paddle: not supported	Classification to MNN: not supported	Classification to RKNN: not supported	Classification to ncnn: not supported	Classification to TFLite: not supported	Classification to CoreML: not supported	Classification to Core AI: not supported
Embeddings	Embeddings to ONNX: supported	Embeddings to TorchScript: supported	Embeddings to ExecuTorch: supported	Embeddings to TensorRT: supported	Embeddings to OpenVINO: supported	Embeddings to Paddle: not supported	Embeddings to MNN: not supported	Embeddings to RKNN: not supported	Embeddings to ncnn: not supported	Embeddings to TFLite: not supported	Embeddings to CoreML: not supported	Embeddings to Core AI: not supported

Python

from libreyolo import LibreYOLO, SAMPLE_IMAGE model = LibreYOLO("LibrePEt16-cls.pt", device="cpu")model.export(format="onnx")

Export setup lists format dependencies and loading exported artifacts.

Checkpoints

File	Weights license
Classification
LibrePEt16-cls.pt	apache-2.0
LibrePEs16-cls.pt	apache-2.0
LibrePEb16-cls.pt	apache-2.0
LibrePEl14-cls.pt	apache-2.0
LibrePEg14-cls.pt	apache-2.0

Every file above exists in the LibreYOLO org today and downloads on first use.

Licensing

Check the license on the Hugging Face repository of the specific weights you download. Every checkpoint in the LibreYOLO org carries one, and they are not always the same across a family. That repository is the authoritative source; the summary below describes what applied when this page was last verified.

This is a description of the licenses involved, not legal advice. If the answer matters commercially, read the licenses yourself and take your own counsel.

Original work: Perception Encoder, Meta AI
Upstream license: Apache-2.0
Upstream source: github.com/facebookresearch/perception_models
LibreYOLO code: MIT
Weights: Apache-2.0, republished at huggingface.co/LibreYOLO
Interpretation: The publisher declares Apache-2.0 for the encoder weights.

Citation

@article{bolya2025PerceptionEncoder,
  title={Perception Encoder: The best visual embeddings are not at the output of the network},
  author={Daniel Bolya and Po-Yao Huang and Peize Sun and Jang Hyun Cho and Andrea Madotto and Chen Wei and Tengyu Ma and Jiale Zhi and Jathushan Rajasegaran and Hanoona Rasheed and Junke Wang and Marco Monteiro and Hu Xu and Shiyu Dong and Nikhila Ravi and Daniel Li and Piotr Doll{\'a}r and Christoph Feichtenhofer},
  journal={arXiv:2504.13181},
  year={2025}
}

Copied from the authors' citation block at raw.githubusercontent.com/facebookresearch/perception_models/main/README.md.