VapourSynth-MLRT-ORT

This package contains the ONNX Runtime backend implementation of the vs-mlrt plugin.

Installation

To install the standard CPU/DirectML/CoreML package:

pip install vapoursynth-mlrt-ort

To install the CUDA-enabled package:

pip install vapoursynth-mlrt-ort-cuda

Building from source

Requirements

  • C++ Compiler: C++20 compatible (e.g. MSVC 2019+, GCC, Clang)
  • Dependencies:
    • onnxruntime (ONNX Runtime SDK)
    • ONNX
    • Protobuf
  • Optional Backend Dependencies:
    • DirectML (Windows): Requires DirectML SDK. Define the DML_DIR environment/CMake variable to point to the SDK directory.
    • CUDA: Requires CUDAToolkit and cuDNN SDKs. Ensure CUDA_PATH, CUDNN_PATH / CUDNN_HOME are set correctly.

Compilation

By default, the package builds the CPU backend (with CoreML on macOS and optionally DirectML on Windows):

uv build --package vapoursynth-mlrt-ort

To build the CUDA-enabled version, the package definition must first be updated using the helper script:

# Update pyproject.toml package configuration to target CUDA
uv run --script scripts/cuda_pyproject.py pyproject.toml

# Compile the CUDA package
uv build --package vapoursynth-mlrt-ort-cuda

Detailed parameter information from the parent project follows.

VapourSynth ONNX Runtime

The vs-onnxruntime plugin provides optimized CPU & CUDA runtime for some popular AI filters.

Building and Installation

To build, you will need ONNX Runtime, protobuf, ONNX and their dependencies.

Please refer to ONNX Runtime Docs for installation notes. Or, you can use our prebuilt Windows binary releases from AmusementClub.

Please refer to our github actions workflow for sample building instructions.

If you only use the CPU backend, then you just need to extract binary release into your vapoursynth/plugins directory.

However, if you also use the CUDA backend, you will need to download some CUDA libraries as well, please see the release page for details. Those CUDA libraries also need to be extracted into VS vapoursynth/plugins directory. The plugin will try to load them from vapoursynth/plugins/vsort/ directory or vapoursynth/plugins/vsmlrt-cuda/ directory.

Usage

Prototype: core.ort.Model(clip[] clips, string network_path[, int[] overlap = None, int[] tilesize = None, string provider = "", int device_id = 0, int verbosity = 2, bint cudnn_benchmark = True, bint builtin = False, string builtindir="models", bint fp16 = False, bint path_is_serialization = False, bint use_cuda_graph = False])

Arguments:

  • clip[] clips: the input clips, only 32-bit floating point RGB or GRAY clips are supported. For model specific input requirements, please consult our wiki.
  • string network_path: the path to the network in ONNX format.
  • int[] overlap: some networks (e.g. CNN) support arbitrary input shape where other networks might only support fixed input shape and the input clip must be processed in tiles. The overlap argument specifies the overlapping (horizontal and vertical, or both, in pixels) between adjacent tiles to minimize boundary issues. Please refer to network specific docs on the recommended overlapping size.
  • int[] tilesize: Even for CNN where arbitrary input sizes could be supported, sometimes the network does not work well for the entire range of input dimensions, and you have to limit the size of each tile. This parameter specify the tile size (horizontal and vertical, or both, including the overlapping). Please refer to network specific docs on the recommended tile size.
  • string provider: Specifies the device to run the inference on.
    • "CPU" or "": pure CPU backend
    • "CUDA": CUDA GPU backend, requires Nvidia Maxwell+ GPUs.
    • "DML": DirectML backend
    • "COREML": CoreML backend
  • int device_id: select the GPU device for the CUDA backend.'
  • int verbosity: specify the verbosity of logging, the default is warning.
    • 0: fatal error only, ORT_LOGGING_LEVEL_FATAL
    • 1: also errors, ORT_LOGGING_LEVEL_ERROR
    • 2: also warnings, ORT_LOGGING_LEVEL_WARNING
    • 3: also info, ORT_LOGGING_LEVEL_INFO
    • 4: everything, ORT_LOGGING_LEVEL_VERBOSE
  • bint cudnn_benchmark: whether to let cuDNN use benchmarking to search for the best convolution kernel to use. Default True. It might incur some startup latency.
  • bint builtin: whether to load the model from the VS plugins directory, see also builtindir.
  • string builtindir: the model directory under VS plugins directory for builtin models, default "models".
  • bint fp16: whether to quantize model to fp16 for faster and memory efficient computation.
  • bint path_is_serialization: whether the network_path argument specifies an onnx serialization of type bytes.
  • bint use_cuda_graph: whether to use CUDA Graphs to improve performance and reduce CPU overhead in CUDA backend. Not all models are supported.
  • int ml_program: select CoreML provider.
    • 0: NeuralNetwork
    • 1: MLProgram

When overlap and tilesize are not specified, the filter will internally try to resize the network to fit the input clips. This might not always work (for example, the network might require the width to be divisible by 8), and the filter will error out in this case.

The general rule is to either:

  1. left out overlap, tilesize at all and just process the input frame in one tile, or
  2. set all three so that the frame is processed in tilesize[0] x tilesize[1] tiles, and adjacent tiles will have an overlap of overlap[0] x overlap[1] pixels on each direction. The overlapped region will be throw out so that only internal output pixels are used.

Project links

Data verified by PyPI on Jul 26, 2026
Data provided by the project maintainers, verified at the time the release was uploaded to PyPI.

Key dates

PyPI data
Data sourced directly from PyPI's database.
  • Released:
Newer release available (16.2)

1 maintainer

PyPI data
Data sourced directly from PyPI's database.
Avatar for Ichunjo from gravatar.com Ichunjo

Credits

Author: AmusementClub

Maintainer: Vardë

GitHub Statistics

Data verified by PyPI on Jul 26, 2026
The GitHub source repository was provided by the project maintainers and verified by PyPI at the time of upload. Stars, forks, and open issues/PRs are derived from that repository and have not been independently verified.

License expression

GPL-3.0-or-later
View SPDX License List

Requires

Python >=3.12

Provides Extra

cufft

Report project as malware

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

vapoursynth_mlrt_ort_cuda-16.1.tar.gz (676.6 kB view details)

Uploaded Source

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

vapoursynth_mlrt_ort_cuda-16.1-py3-none-win_amd64.whl (106.2 MB view details)

Uploaded Python 3Windows x86-64

vapoursynth_mlrt_ort_cuda-16.1-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (102.8 MB view details)

Uploaded Python 3manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

vapoursynth_mlrt_ort_cuda-16.1-py3-none-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl (100.6 MB view details)

Uploaded Python 3manylinux: glibc 2.27+ ARM64manylinux: glibc 2.28+ ARM64

File details

Details for the file vapoursynth_mlrt_ort_cuda-16.1.tar.gz.

File metadata

  • Download URL: vapoursynth_mlrt_ort_cuda-16.1.tar.gz
  • Upload date:
  • Size: 676.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.14

File hashes

Hashes for vapoursynth_mlrt_ort_cuda-16.1.tar.gz
Algorithm Hash digest
SHA256 89693bf16977860cf3e26539d8a17fc8c13526a66baaf1cf6e07bc284fc16cbe
MD5 4f1c5d6edd8b5455e5aa5a7e79fc8a17
BLAKE2b-256 cde036a644575ef16f42a104b6dfe87133a057fb02d56297b5f0998a5934d857

See more details on using hashes here.

Provenance

The following attestation bundles were made for vapoursynth_mlrt_ort_cuda-16.1.tar.gz:

Publisher: cd-publish.yml on Jaded-Encoding-Thaumaturgy/vs-wheels

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file vapoursynth_mlrt_ort_cuda-16.1-py3-none-win_amd64.whl.

File metadata

File hashes

Hashes for vapoursynth_mlrt_ort_cuda-16.1-py3-none-win_amd64.whl
Algorithm Hash digest
SHA256 69a161e9e07c3e519eac5b5fbca55c6d30f67413034d6674cdc13cfa678b6e53
MD5 f97918af67f9ad0b2fdcb30fbf7a97a3
BLAKE2b-256 415d6048f8869dc29a49158bfe3830905bc21dd81e4871f1b9ade8053d3654a5

See more details on using hashes here.

Provenance

The following attestation bundles were made for vapoursynth_mlrt_ort_cuda-16.1-py3-none-win_amd64.whl:

Publisher: cd-publish.yml on Jaded-Encoding-Thaumaturgy/vs-wheels

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file vapoursynth_mlrt_ort_cuda-16.1-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.

File metadata

File hashes

Hashes for vapoursynth_mlrt_ort_cuda-16.1-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 3cb29c41a17cfc8ff3cac8c55921cc2fc7109621aa9bbf2391559fdd6a786c6b
MD5 50f6822ecb15ef7aac8015f950d67393
BLAKE2b-256 3d7b1fc86c05ebbc6cfd39402dc9cfd9f8de3f724de5b65f8e7d6120812b6015

See more details on using hashes here.

Provenance

The following attestation bundles were made for vapoursynth_mlrt_ort_cuda-16.1-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl:

Publisher: cd-publish.yml on Jaded-Encoding-Thaumaturgy/vs-wheels

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file vapoursynth_mlrt_ort_cuda-16.1-py3-none-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl.

File metadata

File hashes

Hashes for vapoursynth_mlrt_ort_cuda-16.1-py3-none-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl
Algorithm Hash digest
SHA256 79ab60373754ec2ee5069ddadc66c66a7cf54d029d9e32e8f7c151cae139f0d5
MD5 3cac63358677309bfc9734fac54fdb81
BLAKE2b-256 bceb96df0d0e48b445828adc62566193ab1fed343f29343e138f464163882238

See more details on using hashes here.

Provenance

The following attestation bundles were made for vapoursynth_mlrt_ort_cuda-16.1-py3-none-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl:

Publisher: cd-publish.yml on Jaded-Encoding-Thaumaturgy/vs-wheels

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Release history Release notifications | RSS feed

16.2

3 files

This release

16.1 This release

4 files

16.0

4 files

15.16.2

4 files

15.16.1.post2

1 file

15.16.1.post1

1 file

15.16.1

1 file

15.16.post2

1 file

15.16.post1

1 file

15.16

1 file

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page