{"id":13456791,"url":"https://github.com/AmusementClub/vs-mlrt","last_synced_at":"2025-03-24T11:31:21.880Z","repository":{"id":40301542,"uuid":"434098140","full_name":"AmusementClub/vs-mlrt","owner":"AmusementClub","description":"Efficient CPU/GPU/Vulkan ML Runtimes for VapourSynth (with built-in support for waifu2x, DPIR, RealESRGANv2/v3, Real-CUGAN, RIFE, SCUNet and more!)","archived":false,"fork":false,"pushed_at":"2024-05-23T00:36:20.000Z","size":581,"stargazers_count":239,"open_issues_count":20,"forks_count":13,"subscribers_count":18,"default_branch":"master","last_synced_at":"2024-05-23T00:52:25.140Z","etag":null,"topics":["artificial-intelligence","cuda","deep-learning","directml","dpir","gpu","migraphx","ncnn","neural-network","onnx","onnxruntime","openvino","real-cugan","real-esrgan","rife","tensorrt","vapoursynth","vulkan","waifu2x"],"latest_commit_sha":null,"homepage":"","language":"C++","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"gpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/AmusementClub.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2021-12-02T05:48:44.000Z","updated_at":"2024-05-31T09:06:55.533Z","dependencies_parsed_at":"2023-02-15T23:15:33.274Z","dependency_job_id":"eea810df-5100-466a-a500-15071035a387","html_url":"https://github.com/AmusementClub/vs-mlrt","commit_stats":null,"previous_names":[],"tags_count":36,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AmusementClub%2Fvs-mlrt","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AmusementClub%2Fvs-mlrt/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AmusementClub%2Fvs-mlrt/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AmusementClub%2Fvs-mlrt/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/AmusementClub","download_url":"https://codeload.github.com/AmusementClub/vs-mlrt/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":245260855,"owners_count":20586484,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["artificial-intelligence","cuda","deep-learning","directml","dpir","gpu","migraphx","ncnn","neural-network","onnx","onnxruntime","openvino","real-cugan","real-esrgan","rife","tensorrt","vapoursynth","vulkan","waifu2x"],"created_at":"2024-07-31T08:01:27.794Z","updated_at":"2025-03-24T11:31:20.180Z","avatar_url":"https://github.com/AmusementClub.png","language":"C++","funding_links":[],"categories":["\u003ca name=\"cpp\"\u003e\u003c/a\u003eC++","C++"],"sub_categories":[],"readme":"# vs-mlrt\n\nThis project provides VapourSynth ML filter runtimes for a variety of platforms:\n - x86 CPUs: [vsov-cpu](#vsov-openvino-based-pure-cpu--intel-gpu-runtime), [vsort-cpu](#vsort-onnx-runtime-based-cpugpu-runtime)\n - Intel GPU (both integrated \u0026 discrete): [vsov-gpu](#vsov-openvino-based-pure-cpu--intel-gpu-runtime), [vsncnn-vk](#vsncnn-ncnn-based-gpu-vulkan-runtime)\n - NVidia GPU: [vsort-cuda](#vsort-onnx-runtime-based-cpugpu-runtime), [vstrt](#vstrt-tensorrt-based-gpu-runtime), [vsncnn-vk](#vsncnn-ncnn-based-gpu-vulkan-runtime)\n - AMD GPU: [vsncnn-vk](#vsncnn-ncnn-based-gpu-vulkan-runtime), [vsmigx](#vsmigx-migraphx-based-gpu-runtime)\n - Apple SoC: [vsort-coreml](#vsort-onnx-runtime-based-cpugpu-runtime)\n\nTo simplify usage, we also provide a Python wrapper [vsmlrt.py](https://github.com/AmusementClub/vs-mlrt/blob/master/scripts/vsmlrt.py)\nfor all bundled models and a unified interface to select different backends.\n\nPlease refer to [the wiki](https://github.com/AmusementClub/vs-mlrt/wiki) for supported models \u0026 usage information.\n\n## vsov: OpenVINO-based Pure CPU \u0026 Intel GPU Runtime\n\n[OpenVINO](https://docs.openvino.ai/latest/index.html) is an AI inference runtime developed\nby Intel, mainly targeting x86 CPUs and Intel GPUs.\n\nThe vs-openvino plugin provides optimized *pure* CPU \u0026 Intel GPU runtime for some popular AI filters.\nIntel GPU supports Gen 8+ on Broadwell+ and the Arc series GPUs.\n\nTo install, download the latest release and extract them into your VS `plugins` directory.\n\nPlease visit the [vsov](vsov) directory for details.\n\n## vsort: ONNX Runtime-based CPU/GPU Runtime\n\n[ONNX Runtime](https://onnxruntime.ai/) is an AI inference runtime with many backends.\n\nThe vs-onnxruntime plugin provides optimized CPU and CUDA GPU runtime for some popular AI filters.\n\nTo install, download the latest release and extract them into your VS `plugins` directory.\n\nPlease visit the [vsort](vsort) directory for details.\n\n## vstrt: TensorRT-based GPU Runtime\n\n[TensorRT](https://developer.nvidia.com/tensorrt) is a highly optimized AI inference runtime\nfor NVidia GPUs. It uses benchmarking to find the optimal kernel to use for your specific\nGPU, and so there is an extra step to build an engine from ONNX network on the machine\nyou are going to use the vstrt filter, and this extra step makes deploying models a little\nharder than the other runtimes. However, the resulting performance is also typically\n*much much better* than the CUDA backend of [vsort](vsort).\n\nTo install, download the latest release and extract them into your VS `plugins` directory.\n\nPlease visit the [vstrt](vstrt) directory for details.\n\n## vsmigx: MIGraphX-based GPU Runtime\n\n[MIGraphX](https://github.com/ROCm/AMDMIGraphX) is a highly optimized AI inference runtime\nfor AMD GPUs. It also uses benchmarking to find the optimal kernel, similar to vstrt.\n\nTo install, download the latest release and extract them into your VS `plugins` directory.\n\nPlease visit the [vsmigx](vsmigx) directory for details.\n\n## vsncnn: NCNN-based GPU (Vulkan) Runtime\n\n[ncnn](https://github.com/Tencent/ncnn) is a popular AI inference runtime. [vsncnn](vsncnn)\nprovides a vulkan based runtime for some AI filters. It includes support for on-the-fly\nONNX to ncnn native format conversion so as to provide a unified interface across all\nruntimes provided by this project. As it uses the device-independent\n[Vulkan](https://en.wikipedia.org/wiki/Vulkan) interface for GPU accelerated inference,\nthis plugin supports all GPUs that provides Vulkan interface (NVidia, AMD, Intel integrated \u0026\ndiscrete GPUs all provide this interface.) Another benefit is that it has a significant\nsmaller footprint than other GPU runtimes (both vsort and vstrt CUDA backends require \u003e1GB\nCUDA libraries.) The main drawback is that it's slower.\n\nTo install, download the latest release and extract them into your VS `plugins` directory.\n\nPlease visit the [vsncnn](vsncnn) directory for details.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FAmusementClub%2Fvs-mlrt","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FAmusementClub%2Fvs-mlrt","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FAmusementClub%2Fvs-mlrt/lists"}