Wideband datasets, models, etc. (#21)

* Wideband extensions to TorchSig

* add missing req and update dataset init

* Adding necessary transforms and visualizations.

Co-authored-by: Luke Boegner <luke.boegner@gmail.com>
Co-authored-by: Garrett Vanhoy <gmvanhoy@gmail.com>

.gitignore

@@ -0,0 +1,3 @@
+__pycache__/
+*.pyc
+*.mdb

README.md

@@ -5,22 +5,25 @@
 -----
 
 
-TorchSig is an open-source signal processing machine learning toolkit based on the PyTorch data handling pipeline. The user-friendly toolkit simplifies common digital signals processing operations, augmentations, and transformations when dealing with both real and complex-valued signals. TorchSig streamlines the integration process of these signals processing tools building on PyTorch, enabling faster and easier development and research for machine learning techniques applied to signals data, particularly within (but not limited to) the radio frequency domain. An example dataset based on many unique communication signal modulations is also included to accelerate the field of modulation recognition.
+TorchSig is an open-source signal processing machine learning toolkit based on the PyTorch data handling pipeline. The user-friendly toolkit simplifies common digital signal processing operations, augmentations, and transformations when dealing with both real and complex-valued signals. TorchSig streamlines the integration process of these signals processing tools building on PyTorch, enabling faster and easier development and research for machine learning techniques applied to signals data, particularly within (but not limited to) the radio frequency domain. An example dataset, Sig53, based on many unique communication signal modulations is included to accelerate the field of modulation classification. Additionally, an example wideband dataset, WidebandSig53, is also included that extends Sig53 with larger data example sizes containing multiple signals enabling accelerated research in the fields of wideband signal detection and recognition.
 
-*TorchSig is currently in beta (v0.1.0)*
+*TorchSig is currently in beta*
 
 ## Key Features
 ---
 TorchSig provides many useful tools to facilitate and accelerate research on signals processing machine learning technologies:
 - The `SignalData` class and its `SignalDescription` objects enable signals objects and meta data to be seamlessly handled and operated on throughout the TorchSig infrastructure.
-- The `Sig53` Dataset is a state-of-the-art static modulations-based RF dataset meant to serve as the next baseline for RFML development & evaluation.
+- The `Sig53` Dataset is a state-of-the-art static modulations-based RF dataset meant to serve as the next baseline for RFML classification development & evaluation.
 - The `ModulationsDataset` class synthetically creates, augments, and transforms the largest communications signals modulations dataset to date in a generic, flexible fashion.
+- The `WidebandSig53` Dataset is a state-of-the-art static wideband RF signals dataset meant to serve as the baseline for RFML signal detection and recognition development & evaluation.
+- The `WidebandModulationsDataset` class synthetically creates, augments, and transforms the largest wideband communications signals dataset in a generic, flexible fashion.
 - Numerous signals processing transforms enable existing ML techniques to be employed on the signals data, streamline domain-specific signals augmentations in signals processing machine learning experiments, and signals-specific data transformations to speed up the field of expert feature signals processing machine learning integration.
+- TorchSig also includes a model API similar to open source code in other ML domains, where several state-of-the-art convolutional and transformer-based neural architectures have been adapted to the signals domain and pretrained on the `Sig53` and `WidebandSig53` datasets. These models can be easily used for follow-on research in the form of additional hyperparameter tuning, out-of-the-box comparative analysis/evaluations, and/or fine-tuning to custom datasets.
 
 
 ## Documentation
 ---
-Documentation can be built locally by following the instructions below.
+Documentation can be found [online](https://torchsig.readthedocs.io/en/latest/) or built locally by following the instructions below.
 ```
 cd docs
 pip install -r docs-requirements.txt

docs/datasets.rst

@@ -28,12 +28,24 @@ Sig53
 .. autoclass:: Sig53
 
 
+WidebandSig53
+~~~~~~~~~~~~~~
+
+.. autoclass:: WidebandSig53
+
+
 ModulationsDataset
 ~~~~~~~~~~~~~~~~~~~~
 
 .. autoclass:: ModulationsDataset
 
 
+WidebandModulationsDataset
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. autoclass:: WidebandModulationsDataset
+
+
 DigitalModulationDataset
 ~~~~~~~~~~~~~~~~~~~~~~~~~~
 
@@ -70,6 +82,24 @@ FMDataset
 .. autoclass:: FMDataset
 
 
+WidebandDataset
+~~~~~~~~~~~~~~~~~~
+
+.. autoclass:: WidebandDataset
+
+
+SyntheticBurstSourceDataset
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. autoclass:: SyntheticBurstSourceDataset
+
+
+FileBurstSourceDataset
+~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. autoclass:: FileBurstSourceDataset
+
+
 RadioML2016
 ~~~~~~~~~~~~~~
 

docs/models.rst

@@ -12,10 +12,14 @@ The following utilities are available:
     :local:
 
 
+IQ Models
+------------------
+
+
 EfficientNet
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. currentmodule:: torchsig.models.efficientnet
+.. currentmodule:: torchsig.models.iq_models.efficientnet.efficientnet
 
 .. autoclass:: efficientnet_b0
 
@@ -27,8 +31,84 @@ EfficientNet
 XCiT
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. currentmodule:: torchsig.models.xcit
+.. currentmodule:: torchsig.models.iq_models.xcit.xcit
 
 .. autoclass:: xcit_nano
 
 .. autoclass:: xcit_tiny12
+
+
+Spectrogram Models
+------------------
+
+
+YOLOv5
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. currentmodule:: torchsig.models.spectrogram_models.yolov5
+
+.. autoclass:: yolov5p
+
+.. autoclass:: yolov5n
+
+.. autoclass:: yolov5s
+
+.. autoclass:: yolov5p_mod_family
+
+.. autoclass:: yolov5n_mod_family
+
+.. autoclass:: yolov5s_mod_family
+
+
+DETR
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. currentmodule:: torchsig.models.spectrogram_models.detr
+
+.. autoclass:: detr_b0_nano
+
+.. autoclass:: detr_b2_nano
+
+.. autoclass:: detr_b4_nano
+
+.. autoclass:: detr_b0_nano_mod_family
+
+.. autoclass:: detr_b2_nano_mod_family
+
+.. autoclass:: detr_b4_nano_mod_family
+
+
+PSPNet
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. currentmodule:: torchsig.models.spectrogram_models.pspnet
+
+.. autoclass:: pspnet_b0
+
+.. autoclass:: pspnet_b2
+
+.. autoclass:: pspnet_b4
+
+.. autoclass:: pspnet_b0_mod_family
+
+.. autoclass:: pspnet_b2_mod_family
+
+.. autoclass:: pspnet_b4_mod_family
+
+
+Mask2Former
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. currentmodule:: torchsig.models.spectrogram_models.mask2former
+
+.. autoclass:: mask2former_b0
+
+.. autoclass:: mask2former_b2
+
+.. autoclass:: mask2former_b4
+
+.. autoclass:: mask2former_b0_mod_family
+
+.. autoclass:: mask2former_b2_mod_family
+
+.. autoclass:: mask2former_b4_mod_family

docs/transforms.rst

@@ -72,6 +72,22 @@ CutOut
 ^^^^^^^^^
 .. autoclass:: CutOut
 
+PatchShuffle
+^^^^^^^^^^^^^
+.. autoclass:: PatchShuffle
+
+DatasetWidebandMixUp
+^^^^^^^^^^^^^^^^^^^^^
+.. autoclass:: DatasetWidebandMixUp
+
+DatasetWidebandCutMix
+^^^^^^^^^^^^^^^^^^^^^^^^^
+.. autoclass:: DatasetWidebandCutMix
+
+SpectrogramRandomResizeCrop
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+.. autoclass:: SpectrogramRandomResizeCrop
+
 
 Expert Feature Transforms
 -------------------------
@@ -159,6 +175,10 @@ RandomFrequencyShift
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 .. autoclass:: RandomFrequencyShift
 
+RandomDelayedFrequencyShift
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+.. autoclass:: RandomDelayedFrequencyShift
+
 LocalOscillatorDrift
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 .. autoclass:: LocalOscillatorDrift
@@ -239,3 +259,32 @@ ImpulseInterferer
 RandomPhaseShift
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 .. autoclass:: RandomPhaseShift
+
+
+Spectrogram Transforms
+----------------------------
+.. currentmodule:: torchsig.transforms.spectrogram_transforms.spec
+
+SpectrogramResize
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+.. autoclass:: SpectrogramResize
+
+SpectrogramDropSamples
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+.. autoclass:: SpectrogramDropSamples
+
+SpectrogramPatchShuffle
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+.. autoclass:: SpectrogramPatchShuffle
+
+SpectrogramTranslation
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+.. autoclass:: SpectrogramTranslation
+
+SpectrogramMosaicCrop
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+.. autoclass:: SpectrogramMosaicCrop
+
+SpectrogramMosaicDownsample
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+.. autoclass:: SpectrogramMosaicDownsample

docs/utils.rst

@@ -52,4 +52,16 @@ Signal Visualizers
 
 .. autoclass:: TimeSeriesVisualizer
 
-.. autoclass:: ImageVisualizer
+.. autoclass:: ImageVisualizer
+
+.. autoclass:: PSDVisualizer
+
+.. autoclass:: MaskVisualizer
+
+.. autoclass:: MaskClassVisualizer
+
+.. autoclass:: SemanticMaskClassVisualizer
+
+.. autoclass:: BoundingBoxVisualizer
+
+.. autoclass:: AnchorBoxVisualizer