anira-project/simple-gain

Simple gain network for anira

A minimal test model: audio times a gain, plus the peak, for 1, 2, 4 and 16 channels, for LibTorch, ONNX Runtime, LiteRT and ExecuTorch. anira's test and example model.

Files

File	Engine	Size	SHA256
onnx/simple_gain_network_16ch.onnx	ONNX Runtime	476	aa3e9e5770b677891ffb1ac86a5288c4c27e2e7e408e30a7bbe3a9cf57c0c27b
onnx/simple_gain_network_4ch.onnx	ONNX Runtime	476	f7423f6b5be612d5e1f23819f9ae3787b30f16ee7ec7e448269db31c49f45dcf
onnx/simple_gain_network_mono.onnx	ONNX Runtime	476	58906a580de0b9e96815095bb651477378cfcfad75d1e70c71d43399ad17b1cb
onnx/simple_gain_network_stereo.onnx	ONNX Runtime	476	d62ad47c12689c9299e780bc15ba964ec7235423099246d6ad347b029e102221
litert/simple_gain_network_16ch.tflite	LiteRT, TensorFlow Lite	1,732	6fa3ed73c222c1d6a770a06910d4ad07f9d9d87398fbd7734aab8aae626bbdeb
litert/simple_gain_network_4ch.tflite	LiteRT, TensorFlow Lite	1,732	2247913d8d1d404706dc55a9c148e6d206ddce96f347920f3f915fb9b5e9488b
litert/simple_gain_network_mono.tflite	LiteRT, TensorFlow Lite	1,728	55b3bc26d6f9231c110d3e24718095d93261a2f2f271e2b9f4cffb80fb5e6273
litert/simple_gain_network_stereo.tflite	LiteRT, TensorFlow Lite	1,732	94bbbb0a572e0e7bc089b1bb8ec1ea4a416629dba84fbf8b0f1e45c28ef2bdd4
libtorch/simple_gain_network_16ch.pt	LibTorch	3,610	6e3a73b68d309bd92ea9fad9c3885f4ab71cf8f23f8e899379079cfdbcdcf49f
libtorch/simple_gain_network_4ch.pt	LibTorch	3,600	4134794d969d767448bba22e01070d8920d61f51022e4efd917b03c6e27fcfca
libtorch/simple_gain_network_mono.pt	LibTorch	3,510	df6cd2117baacf09248ce473c91fa8fd9d5f5ccb3ab1b9dae100d33dc6bfef47
libtorch/simple_gain_network_stereo.pt	LibTorch	3,630	0e861205881adc88c6d85133b09642dead539b7ed2826eebe4e33a21d5721f5c
executorch/simple_gain_network_16ch.pte	ExecuTorch 1.3.1, CPU	1,720	643da6f63f385b6e08841373ecc1d854151b23a8335740507ca2aa85df12ae65
executorch/simple_gain_network_16ch_coreml.pte	ExecuTorch 1.3.1, Core ML (Apple)	7,254	aec97c4fbeffc341d15e5e1a3ebb58cc8cae802a2acee7aae995258ceb0a4961
executorch/simple_gain_network_16ch_vulkan.pte	ExecuTorch 1.3.1, Vulkan (GPUs)	2,816	b948665a3ffb388b3894eb23713863505c6893760aae5de403ef2bb3d6e99dbf
executorch/simple_gain_network_4ch.pte	ExecuTorch 1.3.1, CPU	1,720	dd353942783acad85470ab16916d896d5a21726044ab82b37b2526c384411383
executorch/simple_gain_network_4ch_coreml.pte	ExecuTorch 1.3.1, Core ML (Apple)	7,254	48b8cd4b8cc1f70ea9d319ee7e433dffd28990b1a596b2b32e061c9e363cd18d
executorch/simple_gain_network_4ch_vulkan.pte	ExecuTorch 1.3.1, Vulkan (GPUs)	2,816	e089dfc964de0028ceafb49870b9d5b024ecb279c8a376f0955eec860309d0db
executorch/simple_gain_network_mono.pte	ExecuTorch 1.3.1, CPU	1,720	b523d28fbbaef5812218bcc44167f5985d43aaa8855ab83e9ec8823d8370b8e7
executorch/simple_gain_network_mono_coreml.pte	ExecuTorch 1.3.1, Core ML (Apple)	7,254	6f68d929098675c9ab4fcaea2a9eeca728354136a2c426e854dc26f15cb8f01a
executorch/simple_gain_network_mono_vulkan.pte	ExecuTorch 1.3.1, Vulkan (GPUs)	2,816	2026aaa9b03d163c7734198b2f5f6d34519b566f8c07b38407cc0ac7bdc0a1b8
executorch/simple_gain_network_stereo.pte	ExecuTorch 1.3.1, CPU	1,720	9bc965e169e935ac32685cadd147b96a9509adf5d56b0d7711ab99351fcff590
executorch/simple_gain_network_stereo_coreml.pte	ExecuTorch 1.3.1, Core ML (Apple)	7,254	eabdc6d6f3af76c4a9b79cefa9858d5358ae0aa6eadae93e51302f84fd78f98f
executorch/simple_gain_network_stereo_vulkan.pte	ExecuTorch 1.3.1, Vulkan (GPUs)	2,816	b63ddbdf5904ff967d2b4f3847752a159ea3d901f76841c1b489b2e85930895b
coreml/simple_gain_network_16ch.mlpackage	Core ML (native)	1,918	per file in the catalog
coreml/simple_gain_network_4ch.mlpackage	Core ML (native)	1,918	per file in the catalog
coreml/simple_gain_network_mono.mlpackage	Core ML (native)	1,918	per file in the catalog
coreml/simple_gain_network_stereo.mlpackage	Core ML (native)	1,918	per file in the catalog

Source

anira-project/example-models @ d1bc5cf, SimpleGainNetwork/models/; the export script is SimpleGainNetwork/simple_gain_network.py there.

Formats and checks

Every model here runs on ONNX Runtime, LiteRT and TensorFlow Lite, LibTorch, ExecuTorch (XNNPACK or portable kernels on the CPU, Core ML on Apple, Vulkan on GPUs) and as a native Core ML ML program, with the ONNX models' tensor names. anira foundation models streams every format of every description over twelve consecutive calls with the state fed back and checks that they agree (tools/check_streaming.py). Deviations below are from ONNX Runtime, per model, in the order of the files.

Format	Files	Max deviation
ExecuTorch, Core ML (Apple, float32)	executorch/*_coreml.pte	0.0e+00, 0.0e+00, 0.0e+00, 0.0e+00
ExecuTorch, Vulkan (GPUs)	executorch/*_vulkan.pte	exported; not run here (no Vulkan runtime)
Core ML ML program (anira's native engine, coming with v3; macOS 13, iOS 16)	coreml/*.mlpackage	0.0e+00, 0.0e+00, 0.0e+00, 0.0e+00

uv run --no-project export/export_litert.py <model.onnx> <out.tflite> [<description>] [--guitar-lstm <pt>]
uv run --no-project export/export_executorch.py <model.onnx> <out.pte> [<description>] [--guitar-lstm <pt>] [--coreml | --vulkan]
uv run --no-project export/export_coreml.py <model.onnx> <out.mlpackage> [<description>] [--guitar-lstm <pt>]
uv run --no-project export/export_libtorch.py <model.onnx> <out.pt> [<description>] [--guitar-lstm <pt>]

<description> is the model's description in anira foundation models, for the input shapes where the ONNX model's time axis is dynamic. The PyTorch module comes from the ONNX model (export/onnx_to_torch.py: onnxsim to static shapes, onnx2torch with traceable converters); each export is checked against ONNX Runtime over eight consecutive calls of random input with the state fed back, relative to the output's peak (export/check.py).

This repository holds the weights, the export scripts and the licence. The anira model descriptions live with the code in anira foundation models (models/simple-gain), versioned with git. Pin a revision and check each file's SHA256 when you download:

https://huggingface.co/anira-project/simple-gain/resolve/<revision>/onnx/simple_gain_network_16ch.onnx

Licence

Apache-2.0: LICENSE, copied from the source repository.