fix: CI never ran tests for 3 of 4 packages, or even triggered on them - #33

23a2fa30 fix: eliminate Python 3.14 terminal noise from multiprocessing and ONNX
REVERT: ed2a0df8 Pull request #97: added soundfile so that the correct backend is installed to support audio_applications
REVERT: 3a38e077 added soundfile so that the correct backend is installed to support audio_applications
REVERT: 506eb9e3 Pull request #96: Did nan check for loss to check training instability TINYML_ALGO-730
REVERT: 539fe129 Did nan check for loss to check training instability TINYML_ALGO-730
REVERT: f2ff07be Pull request #95: range normalise preprocessing function removed
REVERT: 3110abeb range normalise preprocessing function removed
REVERT: d34a25f2 Pull request #92: MSPM0 V2
REVERT: 9980157b resolved the comment, added range_normalization argument parser in correct place
REVERT: 69e200df all changes due to range normalisation argument addition
REVERT: 54a07f27 Pull request #93: Fix: added check for integration in AutoML
REVERT: 9e95da6e Added subset = None check to enable integration in AutoML flow
REVERT: f0412f8a Merge branch 'MSPM0_V2' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into MSPM0_V2
REVERT: f6b44a45 feature extraction function and its flag is added
REVERT: 94251b38 Bug fix
REVERT: 19b438f8 Pull request #91: incorporated Qodo suggestions
REVERT: f54f0c53  incorporated Qodo suggestions
REVERT: b08b4058 Pull request #90: audio and image data loader scripts
REVERT: 618b5c8a no message
REVERT: 43281d16 no message
REVERT: 5da3efd7 no message
REVERT: e6170a80 Merge branch 'main-dev' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into MSPM0_V2
REVERT: 9755f8e3 fixes for audio_dataset
REVERT: 442dfe4f support for audio added
REVERT: eea264cc Pull request #80: Added training data exporting functionality to tinyverse.
REVERT: be2564c1 Handled the cases where the raw dataset can be 3d or 4d depending on whether feature extraction is applied or not
REVERT: d11f9e01 Added support to export conv2d layer for ondevice learning
REVERT: e090edb3 TINYML_ALGO-559: Added training data exporting functionality to tinyverse. Data will be exported in ondevice_training_data.c,.h
REVERT: e28db628 Pull request #89: TINYML_ALGO-586: Adding hawq based automatic mixed precision flow in regression, classification, anomaly detection, forecasting train.py
REVERT: 92165e59 spacing
REVERT: 5d2e78bc adding tolerance flow from params.py, utils.py flow through kwargs, and changes in four flows train.py
REVERT: 98869ac4 changing variable name to auto_quantization
REVERT: b2247eeb forecasting mixed precision flow
REVERT: 73c6929d Changes in train.py for all 3 flows regression, classification and anomaly detection to support automatic mixed precision when partial_quantization is True
REVERT: 9580c3e7 adding hawq based partial quantization flow in regression train.py
REVERT: 3c02f8d5 Pull request #88: Image Classification transforms + augmentation
REVERT: c9733e30 enabled training only online augmentation for image datasets
REVERT: 42c4a64f added enable_online_augmentation flag as suggested
REVERT: 1da8c217 no message
REVERT: 2a525fd0 Merge branch 'main-dev' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into MSPM0_V2
REVERT: ece4ebec no message
REVERT: 6e467c8c added opencv-python package for image classification tasks
REVERT: fdea5c96  Added GenericImageDataset support for image classification tasks - Added deterministic image preprocessing transforms:   - GRAYSCALE   - RGB   - RESIZE   - RESIZE_PAD   - AUTOCONTRAST   - EQUALIZE   - INVERT   - CLAHE   - SOBEL   - LAPLACIAN   - BINARIZE
REVERT: d8feb816 TINYML_ALGO-706
REVERT: ceadc440 TINYML_ALGO-705
REVERT: 57f17ac7 TINYML_ALGO-703, TINYML_ALGO-704
REVERT: 2f852929 Pull request #87: Added description on how to read the reconstruction error graph
REVERT: 8b94009e Added description on how to read the reconstruction error plot
REVERT: 6def19ec TINYML_ALGO-678
REVERT: ce0619db TINYML_ALGO-663
REVERT: bb68e886 Pull request #83: edgeai studio regx fix for mspm0
REVERT: 34bab9ef added support for q15-scale-factor( consumed by the rfft preprocessing even for regression task) in train.py and test_onnx.py for timeseries_regression task aswell
REVERT: bf30cb3e bug fix causing a difference in feature extraction with RAW based presets
REVERT: 02252138 Pull request #82: Updated logs to print the average mse of epoch  as well along with the first batch mse
REVERT: d7f07bea Updated logs to print the average mse of epoch  as well along with the first batch mse
REVERT: fb1b3c26 TINYML_ALGO-640
REVERT: 95d00fa2 Pull request #81: Added user_input_config.h flags for pir_fixed point feature extraction
REVERT: 189c25fa no message
REVERT: d5148d0e TINYML_ALGO-532
REVERT: 1fe83124 TINYML_ALGO-638
REVERT: 6828df1f Pull request #79: Adding LR logs using Metric logger in regression
REVERT: e777c8b1 changing log for test to match other flows
REVERT: bd15248e lr in logs using metric logger
REVERT: 79cf4e87 minor
REVERT: 7cd33257 Pull request #78: TINYML_ALGO-609: fix: resolve torch tensor and numpy compatibility issues in timeseries forecasting
REVERT: 41837b43 fix: resolve torch tensor and numpy compatibility issues in timeseries forecasting
REVERT: a2da1c3f TINYML_ALGO-583
REVERT: 8c55e5b1 TINYML_ALGO-571
REVERT: 47d35155 TINYML_ALGO-573
REVERT: 58553315 Pull request #77: Added fixed point feature extraction fns for MSPM0
REVERT: 5e6e20f4 Resolve qodo comments
REVERT: c8f56ed2 Typos fixed
REVERT: 2cc925ca Merge branch 'main-dev' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into MSPM0_V2
REVERT: 905ec313 Bug fix
REVERT: 1af367e5 Pull request #76: Add PIR Feature Extraction - Fixed point
REVERT: 5092ed20 Add PIR Feature Extraction - Fixed point
REVERT: 980464fc Pull request #75: 2026/adithya/modelzoo
REVERT: 05b58ad7 updated the 55k generic TS models architecture to make sure all the layers are optimized to run on the NPU
REVERT: feae6571 minor
REVERT: 2d508f97 Pull request #74: args.output_int set by user is None by default
REVERT: b29f4df0 args.output_int set by user is None by default
REVERT: 30f427e8 Pull request #73: Fixed few datatypes issues in ondevice_training.py and updated the confusion matrix similar to classification style.
REVERT: 7efa609f Fixed few datatypes issues in ondevice_training.py and updated the confusion matrix similar to classification style.
REVERT: 46a631a6 TINYML_ALGO-457
REVERT: 369d98ae Pull request #72: Similar for image classifiication, forecasting, regression and anomaly detection
REVERT: ff4218f5 Similar for image classifiication, forecasting, regression and anomaly detection
REVERT: 68796da9 Pull request #71: TINYML_ALGO-566 Enable test_vector for mspm33
REVERT: 3afb4821 TINYML_ALGO-566 Enable test_vector for mspm33
REVERT: 2be6c3f0 Pull request #68: ecg classification model and feature extraction added
REVERT: 49352411 Pull request #70: TINYML_ALGO-513 : adding partial quantization variable and flow
REVERT: 6de8871e Merge branch 'main-dev' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into MSPM0_V2
REVERT: fc21948f adding parser for partial_quantization
REVERT: 0431a19a adding partial quantization variable and flow
REVERT: 551f9fe7 changes regarding naming convention of pre-processing function
REVERT: 106930c6 Pull request #69: Added 1k model for regression TINYML_ALGO-243 Regression Models
REVERT: e529143b Added 1k model for regression TINYML_ALGO-243 Regression Models
REVERT: 36b83571 flag input_batch_norm is removed in the newer version, so i removed it too
REVERT: 78d06ef2 ecg classification model and feature extraction added
REVERT: 332ab419 Bug fix, installing onnx-graphsurgeon
REVERT: a0b5bdc6 Pull request #67: Added support to export the trainable model by specifying k layers from last.
REVERT: 9f373c79 Added support to export the trainable model by specifying k layers from last. Used onnx_graphsurgeon to extract and export the last k layers
REVERT: 7e13fd7e TINYML_ALGO-542
REVERT: 707f8734 Pull request #66: TINYML_ALGO-538 Tinyverse: Support for SKIP_NORMALIZE, when feature extraction isn't used
REVERT: bedcfde7 TINYML_ALGO-538 Tinyverse: Support for SKIP_NORMALIZE, when feature extraction isn't used
REVERT: 1883e300 TINYML_ALGO-537
REVERT: e0e5b2a6 Pull request #59: Added RNN,LSTM,GRU layers and LSTM(hidden size=5) model
REVERT: bd6d7e6e Bug Fix for TINYML_ALGO-531
REVERT: 4c858e07 golden vector generation in forecasting, code structural change that follows latest code changes in main-dev, Removed with_input_batchnorm in the LSTM models
REVERT: 99807e5a Merge branch 'main-dev' of bitbucket.itg.ti.com:tinyml-algo/tinyml-tinyverse into 2025/fasna
REVERT: a5441771 TINYML_ALGO-531
REVERT: f7738872 Bug fix reported by Fasna
REVERT: 9341adea Bug fix reported by Fasna
REVERT: 98ac91bb TINYML_ALGO-530
REVERT: a080b508 TINYML_ALGO-530
REVERT: 71bef15a Removed with_input_bathnorm from LSTM models
REVERT: 8c02ea3b rebase
REVERT: d9576898 TINYML_ALGO-529
REVERT: 21ab3380 TINYML_ALGO-528
REVERT: 31ad11f2 Code Cleanup
REVERT: c9eecaa2 TINYML_ALGO-523
REVERT: f4654213 minor
REVERT: 15f4356d Added LSTM8 forecasting model
REVERT: a13b8b70 Code Cleanup
REVERT: c4a6f5b7 TINYML_ALGO-517, TINYML_ALGO-519, TINYML_ALGO-520
REVERT: 3ce34f72 TINYML_ALGO-522
REVERT: a042db6a TINYML_ALGO-522
REVERT: ac077f7c TINYML_ALGO-522
REVERT: 33e673b4 On Device Learning flow made available as an argument for all train scripts
REVERT: b132787d Pull request #65: Added support for on device learning in Tinyverse.
REVERT: 62f01b0e Added support for on device learning in Tinyverse. Added a model with frozen and trainable parts and exported frozen model with tvm and created trainable_parameters.c/.h for the trainable parts of the model
REVERT: 945fd011 Pull request #63: 2025/tushar
REVERT: 862771e9 TINYML_ALGO-515 Tinyverse: Incorrect raw_test_input when not using FE
REVERT: 97845ffc Pull request #62: fixed a small bug, when num_frame_concat is greater than 1, the feature_size_per_frame and stack frame width are getting populated wrongly
REVERT: 455ebd5d fixed a small bug, when num_frame_concat is greater than 1, the feature_size_per_frame and stack frame width are getting populated wrongly
REVERT: eb0d7951 max pool layer of kernel_size (1, 1) doesn't make sense
REVERT: de63b73d Merge branch 'main-dev' of bitbucket.itg.ti.com:tinyml-algo/tinyml-tinyverse into 2025/fasna
REVERT: 0ee46974 Pull request #61: Correction in default value of skip_normalize
REVERT: edd5ac28 Correction
REVERT: 03f80db3 Pull request #60: TINYML_ALGO-494 ModelMaker: Compilation arguments of skip_normalize and output_int required in user_input_config
REVERT: 659a2abc Included model_test_input in test_vector.c so as to align it with newer feature extraction library
REVERT: 7c5b8207 corrected output tensor shape
REVERT: 12871613 Added RNN,LSTM,GRU layers and LSTM(hidden size=5) model
REVERT: 3091af25 TINYML_ALGO-494 ModelMaker: Compilation arguments of skip_normalize and output_int required in user_input_config
REVERT: 4b5fd8f3 Pull request #58: mspm0 bin_size Fix
REVERT: 23a54813 fixed the bin_size getting generated for bin_q15. earlier the flag was capturing the correct value but the self.bin_size was wrong
REVERT: 3164a182 Pull request #54: fix for mnist classification example
REVERT: a54577e1 Pull request #55: added support for PIR-net
REVERT: 0859ef7c removed print statements in test_onnx.py
REVERT: 520c4a14 removed commented out statements
REVERT: 49a157ce removed print statements
REVERT: f45b88b2 fix after evaluate_classification function was changed
REVERT: 65ca0a24 added support for PIR-net
REVERT: 1e9123d9 Pull request #52: Added a new model with linear layers and update the existing CNN model to support any length input
REVERT: 6a1b079a Merge branch 'main-dev' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into MSPM0
REVERT: d1c463b9 Added a new linear model and update the existing CNN model to support any length input
REVERT: 8851a8a2 Pull request #51: washing machine example changes and reg flow
REVERT: 21c84f1f washing machine example changes and reg flow
REVERT: 8986645e Minor bug fix on window-count
REVERT: 2819a433 TINYML_ALGO-468
REVERT: d230ff77 Pull request #50: added support for PIR-net
REVERT: ad30f116 ammended __init_.py script to enable PIR model
REVERT: 3f8d3191 modified compilation.py script for print statement for CC2755
REVERT: ee1b8a55 added support for PIR-net
REVERT: bffe3a61 TINYML_ALGO-463: All flows changed to not reload dataset for quantization
REVERT: 86ef2b68 TINYML_ALGO-467- Torchinfo summary breaks Model Composer (but not Modelmaker)
REVERT: 7d986e94 Pull request #49: adding 4k washing machine model
REVERT: 61f3f994 washing machine 4K params model
REVERT: 571098da adding 4k washing machine model
REVERT: 970fc80f TINYML_ALGO-465: Bug fix that caused IP models to be printed in model.summary
REVERT: 4a8f4de0 TINYML_ALGO-463- Dont re-load dataset for quantization in all flows
REVERT: bb8e10b4 Pull request #48: 2025/abhijeet
REVERT: 93be44ed relevant changes in regression flow
REVERT: 79735ee0 Minor changes in reg train.py
REVERT: 972d47f3 Based on discussions with MCE team, the environment variable for MSPM0's compiler is renamed from MSPM0_CGT_PATH to ARM_LLVM_CGT_PATH. This maintains consistency with other MCEs as well as CCS
REVERT: 98250ec0 washing machine models and regression flow
REVERT: 630299d0 Changes 29th Sept washing machine
REVERT: 293725ec TINYML_ALGO-423 : August Example Battery_RUL Regression Model
REVERT: 7461f213 Pull request #45: Added anomaly detection flow
REVERT: 52d54aab Pull request #47: fixes
REVERT: b69081ea fixed the accidental removal of  self.preprocessing_flags.append('SKIP_NORMALIZE')
REVERT: ddb56ffb Merge branch 'main-dev' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into MSPM0
REVERT: 10450aca Fixed the issues mentioned in the Anomaly detection flow PR
REVERT: 226c5b66 Added anomaly detection flow
REVERT: 3dff5b52 no message
REVERT: d3e6a97a Pull request #46: Correction
REVERT: 3b43b964 Correction
REVERT: bdf8d2b5 Pull request #44: Minor fixes in forecasting flow
REVERT: 8e9a14b9 Flattening axes in scatter plot to prevent indexing issues
REVERT: a3c2810c minor
REVERT: bdc4f6a6 minor changes
REVERT: ff984086 TINYML_ALGO-454
REVERT: e4d81e39 Pull request #43: 2025/tushar
REVERT: 2b3868b4 For backward compatability using value of FE_LOG_BASE as evaluated value of e
REVERT: 7c368cb8 TINYML_ALGO-452 Tinyverse: Feature extraction params and preprocessing flags for new AI Library
REVERT: ba1c2a53 Merge branch 'main-dev' of bitbucket.itg.ti.com:tinyml-algo/tinyml-tinyverse into 2025/fasna
REVERT: 2c32eace minor
REVERT: 0723948c Pull request #41: TINYML_ALGO-444: Training process to provide feedback on which data files caused false alarms
REVERT: 4636a96a Release prep for v1.2
REVERT: 8bb3fe56 Pull request #39: Additional feature extraction functions added for MSPM0 + image_dataset.py and LENET5 model
REVERT: 3ef22225 no message
REVERT: 10c652d1 changed FE_SCALE nomeclature to FE_COMPLEX_MAG_SCALE and FE_COMPLEX_MAG_SCALE_FACTOR
REVERT: 3807389a no message
REVERT: 70e51d3b comment out unnecessary info print
REVERT: ba23a26a removed blank lines
REVERT: 7b0a72d7 renamed MnistDataset to GenericImageDataset and added logger info statements.
REVERT: 4dbc4868 removed unncessary parameters from train.py and test_onnx.py for image classification. renamed MnistDataset to GenericImageDataset
REVERT: 0e026f04 minor
REVERT: 8e87e24b minor changes
REVERT: dc890465 minor changes
REVERT: 1e77f792 Changed logging configuration of file_level_classification_summary
REVERT: 9618dfe5 Training process to provide feedback on which data files caused false alarms
REVERT: f0b35589 added torchvision==0.22.1 for image_classification
REVERT: a8225167 Merge branch 'main-dev' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into MSPM0
REVERT: f5a3b6e3 no message
REVERT: bbe1e091 no message
REVERT: 03b7376e minor fix for mnist
REVERT: f8d5326d merged mnist branch into mspm0 branch
REVERT: def9876c Merge branch 'mspm0-lenet5-mnist' into MSPM0
REVERT: 138a59da no message
REVERT: 4ad59997 Pull request #40: Removed model_test_input from test_vector.c since forcasting does not use feature extraction transforms (concat) for now
REVERT: d22a5d87 no message
REVERT: 708fb469 Fixed error in log transformation , now users can choose their own log base (earlier only 10 and e were allowed), updated preprocessing flags for classification and forecasting so that it takes correct user input from config
REVERT: 22e37c17 Removed model_test_input from test_vector.c since forcasting does not use feature extraction transforms (concat) for now
REVERT: ddd01d2c temp changes
REVERT: 8b2f5a48 minor cosmetic fixes for the pr
REVERT: 9d31643b cleaned up the to q15 fn
REVERT: 78817869 fix for pytoml to not download from github as we have observed some issues there
REVERT: e59e1e09 basic support for lenet5
REVERT: bc2470d7 all modifications done to support timeseries applications for mspm0
REVERT: 639a9a86 Pull request #38: Correction for non availability of feature size per frame
REVERT: c1f2e167 Correction for non availability of feature size per frame
REVERT: 8e4d584f fixed a tiny bug
REVERT: bd393b8a Pull request #36: Running only a model without feature extraction requires these flags in user input config
REVERT: 3ac67ee9 Replicating in all files
REVERT: 591e8275 Pull request #37: 2025/fasna
REVERT: c650fb75 Generate golden vectors for forecasting
REVERT: 7b7ba9ba Generate golden vectors for forecasting
REVERT: f8dcf1fa Running only a model without feature extraction requires these flags in user input config
REVERT: e065b55a TINYML_ALGO-442
REVERT: d3055aae TINYML_ALGO-442: Gain variations for different classes in a classification task
REVERT: bbcf5efd Merge remote-tracking branch 'origin' into MSPM0
REVERT: 7bf44162 no message
REVERT: 31a57635 updated timeseries_dataset.py to replicare kilby's ac arc feature extraction flow
REVERT: a3ed3fcf Pull request #35: Bug Fix: Malformed node or string error for certain files
REVERT: 1414d49a Merge branch 'main-dev' of bitbucket.itg.ti.com:tinyml-algo/tinyml-tinyverse into 2025/fasna
REVERT: 5f674c57 Minor changes
REVERT: afd5d745 Minor edit in logger placement
REVERT: 7b4bc89a Minor bug fix
REVERT: 0e936f8c Pull request #34: Use tensors to calculate r2_score and smape
REVERT: 04907b09 Merge branch 'main-dev' of bitbucket.itg.ti.com:tinyml-algo/tinyml-tinyverse into 2025/fasna
REVERT: 5273ee45 added metric logger statements
REVERT: d9683239 TINYML_ALGO-412, TINYML_ALGO-413, TINYML_ALGO-414
REVERT: 784c6fe8 Merge branch 'main-dev' of bitbucket.itg.ti.com:tinyml-algo/tinyml-tinyverse into 2025/fasna
REVERT: e8adc402 use tensors to calculate r2_score and smape
REVERT: c5dd15be Merge branch 'main-dev'
REVERT: bb69d679 TINYML_ALGO-404: Quantization accuracy on Windows is generally poor
REVERT: 4dad6028 Merge branch 'main-dev'
REVERT: 4f1981b0 Pull request #33: TINYML_ALGO-372: Timeseries Forecasting (tinyverse) TINYML_ALGO-405: (GoF now handles multiple file types)
REVERT: bb155b12 updated r2_score function
REVERT: fc8f8cf0 GoF now handles multiple file types
REVERT: 36fff4f3 cleaned code
REVERT: 4db482f9 cleaned utils.py
REVERT: 8f52d9ce cleaned timeseries_dataset.py
REVERT: f7de8731 r2 score bug
REVERT: 1b2be41c minor
REVERT: b3570fef minor
REVERT: f471eabd minor
REVERT: a78991df minor
REVERT: 8c8d3161 minor
REVERT: e4f5a614 removed rnn code
REVERT: 5083b626 removed rnn code
REVERT: 4685fd35 cleaned code
REVERT: ab6955c1 minor
REVERT: 79416ff4 minor
REVERT: a6e840a5 minor
REVERT: aceb7ec1 minor
REVERT: bbb0e047 minor changes
REVERT: 36ff7a91 deleted files
REVERT: 047fd468 Merge branch 'main-dev'
REVERT: dbb19fe8 TINYML_ALGO-406: Fixed torch not compiled with CUDA issue
REVERT: 1cba1a90 Changed submodules to pip install from specific version instead of main
REVERT: 67f0c9ab Changed submodules to pip install from specific version instead of main
REVERT: 317a03c9 Pull request #30: Minor bug fix
REVERT: 50e01da8 Minor bug fix
REVERT: 65b59fb4 Pull request #29: Main dev to main merge
REVERT: d51eeaaa TINYML_ALGO-364 - Cleanup of unnecesasry packages and cross platform support
REVERT: a0dace5d TINYML_ALGO-397
REVERT: a9145a49 TINYML_ALGO-397-Evaluate TimeSeries_Generic_6k model
REVERT: 2fb09496 TINYML_ALGO-398: Dataset Loader fails when file size is small (but larger than frame size)
REVERT: 16d2d5a0 Even quantization training now reports best epoch instead of last epoch
REVERT: 7707eac6 Pull request #28: 2025/tushar
REVERT: 8ce7c65f Enable normalization calculations only when needed
REVERT: 1b1ffd9e Added lambda reg param
REVERT: 1e9bd1f0 TINYML_ALGO-394, TINYML_ALGO-395 : Support for F29, SDK dependency removed from MSPM0
REVERT: 73b59d2c Merge branch '2025/tushar' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into 2025/tushar
REVERT: 24251e83 TINYML_ALGO-243 Regression Models
REVERT: 2018d5e9 TINYML_ALGO-367 Enabling golden vectors for  regression
REVERT: ea45f4dd TINYML_ALGO-243 Regression Models
REVERT: e404ba57 TINYML_ALGO-367 Enabling golden vectors for  regression
REVERT: af397741 TINYML_ALGO-243 Regression Models
REVERT: 62a34f2b TINYML_ALGO-367 Enabling golden vectors for  regression
REVERT: e3269719 TINYML_ALGO-243 Regression Models
REVERT: 4dbf4f9e TINYML_ALGO-367 Enabling golden vectors for  regression
REVERT: fb11d23a Smooth exit to NAS not finding a GPU
REVERT: d53a2622 Pull request #27: Adjusments for NAS
REVERT: be8bc663 Merge branch '2025/soum' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into 2025/soum
REVERT: f29c492c Adjusments for NAS
REVERT: a119c972 Adjusments for NAS
REVERT: 8309d69e added scalling
REVERT: 6c31bfe5 Backup
REVERT: 92aac8ce initial commit: Forecasting
REVERT: bff06220 TINYML_ALGO-385, TINYML_ALGO-386 - Version upgrade to 1.1 & C2000Ware 6.0
REVERT: a5e2b7c7 Merge branch '2025/tushar' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into 2025/tushar
REVERT: ba0cadb9 TINYML_ALGO-243 Regression Models
REVERT: 6ea655e4 TINYML_ALGO-367 Enabling golden vectors for  regression
REVERT: d3d5fefc TINYML_ALGO-243 Regression Models
REVERT: bab246b4 TINYML_ALGO-367 Enabling golden vectors for  regression
REVERT: 54645d0e TINYML_ALGO-380: Modelmaker to be compatible with ti-mcu-nnc-2.0.0
REVERT: 3301c344 TINYML_ALGO-381: Moving away from requirements.txt
REVERT: e925666e Merge branch '2025/tushar' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into 2025/tushar
REVERT: 97f0b966 TINYML_ALGO-243 Regression Models
REVERT: 45a6c967 TINYML_ALGO-367 Enabling golden vectors for  regression
REVERT: ed25ff1f TINYML_ALGO-243 Regression Models
REVERT: c794d418 Added decimation, high pass filter, q15 transformation and fixed point fft (CMSIS DSP Q15) as part of feature extraction.
REVERT: 567b1bb5 TINYML_ALGO-376
REVERT: cabbe55a Merge branches '2025/tushar' and '2025/tushar' of ssh://bitbucket.itg.ti.com/tinyml-algo/tinyml-tinyverse into 2025/tushar
REVERT: 4b3c6c2d TINYML_ALGO-367 Enabling golden vectors for  regression
REVERT: e30a3960 Pull request #25: TINYML_ALGO-352: Dataset header handling fix
REVERT: 772886e4 TINYML_ALGO-374 : AUC ROC & confusion matrix computation breaks if test dataset does not have ground truth representing all the classes
REVERT: ac64597b TINYML_ALGO-373: If every epoch's accuracy is 0 or nan, then it wont export any checkpoint and breaks the flow
REVERT: 04804b85 Dataset header handling fix
REVERT: d3c72bc5 dataset header handling
REVERT: 05973c7c dataset header handling fix
REVERT: 5630633a TINYML_ALGO-367 Enabling golden vectors for  regression
REVERT: 688560ed fixed typo
REVERT: 8d76b14b TINYML_ALGO-290: Updated toml with newer packages
REVERT: fdca7005 TINYML_ALGO-290: Python package version updated
REVERT: 35d97005 Pull request #24: Model Addition and correction in golden output dtype
REVERT: 9c12b6e2 TINYML_ALGO-20 Addition of Models from STZoo TINYML_ALGO-353 dtype of golden output based on quantization type and output_dequantize
REVERT: 47405172 Pull request #18: TINYML_ALGO-347 Tinyverse: Scale inputs before feature extraction and after loading datafile
REVERT: b2c79bf8 TINYML_ALGO-347 Tinyverse: Scale inputs before feature extraction and after loading datafile
REVERT: 4e8f13ca Pull request #17: TINYML_ALGO-343 Added argument to get quantized/dequantized output from model
REVERT: 687d26e5 TINYML_ALGO-343 Added argument to get quantized/dequantized output from model
REVERT: ca3b281a TINYML_ALGO-342
REVERT: b558a717 Minor bug fixes
REVERT: c85acec8 Fix for quantization error logging
REVERT: 768d678c Exception Handling if Multiclass ROC plots encounter NaN error
REVERT: f7afbdcc Updated for quantization_error_logging
REVERT: 3c01b294 GOF Utils doesnt display the plots. Just saves them
REVERT: 44298e23 Pull request #14: TINYML_ALGO-239: Set default value of frame_skip to 1
REVERT: d9d067f2 made frame_skip internal for gof
REVERT: 1b43783f frame_skip set internally inside gof
REVERT: 538b7ba4 Merge branch 'main' of https://bitbucket.itg.ti.com/scm/tinyml-algo/tinyml-tinyverse into 2025/fasna
REVERT: 4798b93d Minor bug fix if CGT path isnt found
REVERT: 20929a74 Updated regression flow bugs
REVERT: a24fe384 Minor update to setup files
REVERT: 1197abe1 Updated YAML structure
REVERT: 3c9d0662 Merge branch 'main' of https://bitbucket.itg.ti.com/scm/tinyml-algo/tinyml-tinyverse into 2025/fasna
REVERT: 0f16fb4d Updated YAML structure
REVERT: 89c382c6 Minor bug fix
REVERT: 22170254 Pull request #15: correction in residual model, assigning values of config dict to self
REVERT: 0b6a49b4 correction in residual model, assigning values of config dict to self
REVERT: 42179ceb Merge branch 'main' of https://bitbucket.itg.ti.com/scm/tinyml-algo/tinyml-tinyverse into 2025/fasna
REVERT: 408e2e0c Set default value of frame_skip to 1
REVERT: bf6b2135 Minor bug fix
REVERT: 3eb17c9c Bug fix
REVERT: 941a3f47 Minor bugfixes
REVERT: 1a243066 Pull request #12: TINYML_ALGO-280 March App Example for slicing model
REVERT: caa4c69e Pull request #13: TINYML_ALGO-239: Goodness of Fit Test Integration in tinyverse
REVERT: b6c21659 Renamed gof_test.py to gof_utils.py
REVERT: f0672103 Modified GoF notes and updated output directory of plot
REVERT: 5159945b Merge branch 'main' of https://bitbucket.itg.ti.com/scm/tinyml-algo/tinyml-tinyverse into 2025/fasna
REVERT: d0071c81 Goodness of Fit Test Integration
REVERT: 3b461b48 removing copy of GenericTSDataset
REVERT: cbdde8ab TINYML_ALGO-280 March App Example for slicing model
REVERT: f92aa218 Pull request #11: 2025/adithya autoencoder
REVERT: 2c294adf Feature Update for Anomaly Detection: TINYML_ALGO-265, TINYML_ALGO-266, TINYML_ALGO-267, TINYML_ALGO-268
REVERT: 13aa6ce5 Flowflush completion
REVERT: 597af0dd TINYML_ALGO-282 Models are separated out into different files for better UX. Addition of Autoencoder Model Examples. TINYML_ALGO-265, TINYML_ALGO-266, TINYML_ALGO-267, TINYML_ALGO-268
REVERT: 4faaeabf Draft commit
REVERT: 6b5d535a Initial commit
REVERT: 8d1dbb16 Pull request #10: 2025/adithya cnn preprocessing
REVERT: 716fcc79 updated feature extraction model
REVERT: 50b8aa40 change linear to conv_bn_relu as linear is expensive for computation
REVERT: acd7f4c5 using FE lInear model and adam optimizer
REVERT: 235d73e9 More bug fixes
REVERT: 7b4a2875 Draft commit
REVERT: ad985f6e Floating point training works and is able to export a model. QAT fails. TVM also fails
REVERT: dbfb3e3d Minor update
REVERT: 5184f662 Draft commit
REVERT: 8e4b3888 Draft
REVERT: 6a7c201e Lot of bug fixes
REVERT: 75654907 Minor bug fix
REVERT: ef0afec9 Minor bug fix
REVERT: a6db9a3b Support for gpu compute from tinyml-mlbackend
REVERT: a26968d1 Bug Fix
REVERT: 106c1f1d Minor bug fixes
REVERT: 94a6df5f TINYML_ALGO-250, TINYML_ALGO-251 - Torch mps backend support
REVERT: 1f367061 PTQ and QAT support in toolchain, Ternary and 4 bit support in Toolchain: TINYML_ALGO-253, TINYML_ALGO-254, TINYML_ALGO-255, TINYML_ALGO-256, TINYML_ALGO-257
REVERT: 19dd4255 Pull request #9: 2025/tushar
REVERT: 3a8dbab9 correction
REVERT: b1f139ae working towards support for regression
REVERT: 39656031 Minor bug fix
REVERT: 1ee07e9c test-bench requires artifacts generated at compilation stage
REVERT: 21bce30e correction in FFTBIN 1D, 2D stacking
REVERT: ea6d408c TINYML_ALGO-242 , General Cleanup
REVERT: ea2f4ac2 Minor changes
REVERT: 0c8d235f Feature Support for Timeseries Regression. TINYML_ALGO-235, TINYML_ALGO-233
REVERT: 41055955 Minor readme update
REVERT: 6cf28609 Minor bug with redundant statements
REVERT: a2d4edc7 Preparation for v1.0 release
REVERT: da91ce9a Amended the temporary commit
REVERT: 2631abcd Temporary commit
REVERT: 4d783ae2 TINYML_ALGO-230
REVERT: 471a2c56 TINYML_ALGO-228, TINYML_ALGO-230
REVERT: b52afb15 TINYML_ALGO-215
REVERT: e6951987 Minor bug fixes
REVERT: b37c679b Pull request #8: Enhancements
REVERT: b9f4520d using np.sum instead of python inbuilt sum
REVERT: cfa29f2c increasing the speed 5x by using different approach for iteration
REVERT: 0a9c1b96 Commented out debug statements to speed up data loading significantly
REVERT: 2c36a1de Minor readme edit
REVERT: d66a283c Pull request #7: 2024/post training analysis
REVERT: eb912bf7 TINYML_ALGO-140: Data Augmentation feature
REVERT: dde19d79 Stray lines cleaned up
REVERT: 3b920413 TINYML_ALGO-147, TINYML_ALGO-226
REVERT: e7ebb480 Code Cleanup
REVERT: 450e4901 TINYML_ALGO-225
REVERT: 2dfc4bd3 Minor bug fixes
REVERT: f0d1d73b Minor code cleanup
REVERT: 44aa682f TINYML_ALGO-224 GPU run support and code cleanup.
REVERT: a23090d1 Still editing
REVERT: b36137a7 Changed test_onnx.py as well
REVERT: afffd072 Work in progress to add AUC ROC score. Yet to add to test_onnx.py
REVERT: 935520f8 Draft commit
REVERT: 426d0f26 Pull request #6: 2024/tushar concat TS Dataset
REVERT: 9f1a6918 removed original dataset loader and changed x_temp to x
REVERT: 8f39d429 using dont-train-just-feat-ext instead of test-bench, breakdown of generate_golden_vectors function
REVERT: 1c6bbda1 generalized output of basic transforms
REVERT: a94aa7d4 added test_bench for tinyml-firmware
REVERT: dc17581d correction in bin_size, addition of FE_RAW in preprocessing flags
REVERT: 779fefc9 checked difference btw basic_binning and dataloader binning, removed basic binning and bin_size
REVERT: 9583f0cc corrected the transform name in preprocessing_flags
REVERT: 288e59df Commit message
REVERT: 41f2bfde removed unused variable
REVERT: cc66eba7 code cleanup and comments
REVERT: d8a17c21 added offset and reframed variables
REVERT: 4e512800 adding offset and scaling for GenericTSDataset
REVERT: 6b397abb code cleanup
REVERT: 653d0018 incorporating new transforms
REVERT: 6d011e32 Merge branch 'main' into 2024/tushar_concat_TS_Dataset
REVERT: 0905e78b changed simple timeseries parameters
REVERT: 3f53f8a0 only values that are required will be present in feature extractions
REVERT: f6242eff Support for Haar and Hadamard Transforms. TINYML_ALGO-219, TINYML_ALGO-220
REVERT: 2585764a Clones only torchmodelopt directory instead of edgeai-modelopt repo
REVERT: 5241c4f3 Changed test_vectors.c to test_vector.c for the sake of 5.04 C2000Ware SDK having it wrongly
REVERT: 1c1f5bf0 TINYML_ALGO-214
REVERT: 2fb5df3a cleaning up code
REVERT: 26f7ba25 introduced a generic timeseries dataset as GenericTSDataset, including testing functionality for ArcFault, MotorFault
REVERT: 2e0631ff TINYML_ALGO-160: Additional generic models
REVERT: e8af9e6b TINYML_ALGO-213
REVERT: 81b6eed6 TINYML_ALGO-212
REVERT: 5cae1207 Renamed generic models appropriately
REVERT: 09a6c528 renaming of variables
REVERT: 4e8a4d9e added transformations Downsample and SimpleWindow of SimpleTSDataset
REVERT: 94182736 all cases of mf, af can work simultaneously with kwargs
REVERT: b1f5563e Updated tvm wheel to the final release
REVERT: 3f90c0fc few edge cases of arc fault left
REVERT: 9db70875 all 4 cases of motor fault are working
REVERT: 2cf79554 restructuring the feature extraction in a single function
REVERT: 1e329944 generic models changed. padding is now used.
REVERT: ff768ed4 restructuring the feature extraction in a single function
REVERT: 0545d61f Pull request #5: generic models changed.
REVERT: da21e31c generic models changed. padding is now used.
REVERT: 262ce115 modifying feature_extraction to accomodate variabe channels for future
REVERT: 88407b34 arcfault and motorfault dataset combined
REVERT: 98a05733 Converted the different fault detection in single class, compressed the shape changer of vax to data_packed using np
REVERT: 99167fce Updated with tvm rc5 package. Removed a few unused features in timeseries_dataset
REVERT: 3c43e5cc Minor updates
REVERT: cc841e61 TINYML_ALGO-200
REVERT: a176ac3f Preprocessing feature updates
REVERT: 7ffba99e Added AdaptiveAverage Pool Layer support
REVERT: 322d74ea Updated path to latest model optimization repo
REVERT: eaa8c747 Preparation for 0.9 release
REVERT: d62a616b TINYML_ALGO-197
REVERT: 36224c34 TINYML_ALGO-188: Version number on the log
REVERT: 3663b300 Added Average Pool support for tinynn.py. w.r.t: TINYML_ALGO-187
REVERT: 57358093 TINYML_ALGO-192: Integrate with ti_mcu_nnc 1.3.0rc4
REVERT: 5bcdf2d9 Code Clean up
REVERT: 6873065c Removed default optimizer and scheduler
REVERT: fbd05531 More removal of unused imports, variables and code cleanup
REVERT: cf8e8be1 Removed unused imports, code cleanup
REVERT: d837c68a Moved time.time() to a platform independent method
REVERT: 1448ed9a TINYML_ALGO-186, added support for dual_op
REVERT: 52a82fe2 Preparation for 0.9 release, TINYML_ALGO-186, added support for dual_op, TINYML_ALGO-187 : added new motor fault model
REVERT: a500ea96 Replaced '..' with os.path.dirname to not assume symlinks
REVERT: d8edd418 TINYML_ALGO-171: IP Protection for TI-developed arc-fault and motor-fault models on desktop
REVERT: cdf08b0d Added another exception
REVERT: 9057f5d3 TINYML_ALGO-179: Invalid datafiles in dataset result in training error
REVERT: a1f6d5a2 Change in preprocessing file generated and its content
REVERT: c743df34 Total elimination of forward slashes
REVERT: 27f8b61a TINYML_ALGO-176: Bug fix for training error during training data of incompatible dimensions
REVERT: 589d94d9 Added import of tinyml_proprietary models
REVERT: b23979cf Moved references under tinyml_tinyverse package to enable pyinstaller
REVERT: 86488573 Updated requirements with nnc-1.3.0rc2 package
REVERT: 18b9da84 Doesn't break flow if devc.o is not generated while compilation
REVERT: 89aee0a3 Changed workers to 0 for Windows and 16 for Linux
REVERT: 919aaf5e Platform dependent workers in train/test.py
REVERT: 5158fc4b Onnx version downgrade for Windows
REVERT: 64eef918 Downgraded torch from 2.4.0 to 2.3.0 because of fbgemm.dll being unavailable on Windows
REVERT: 7056e148 Used pep517 to remove a deprecated warning during pip
REVERT: 19e33c0c Moved back to Py3.10 because of onnx-sim, onnx-optimizer packages not supported on Windows for Py3.12
REVERT: f83d33c8 Typo in requirements
REVERT: 1196cfb7 Updated README
REVERT: dea66c36 v0.8 release prep
REVERT: 650863a7 Dynamic Preprocessing updates
REVERT: 8d6a308d 'org_sr' changed to 'sampling_rate'
REVERT: afc5a2f8 TINYML_ALGO-169, TINYML_ALGO-170 : Moved to Python3.12
REVERT: c227e060 Replaced forward slashes for cross OS compatibility
REVERT: fad18cdf Stray space in requirements file was causing packages to not be installed
REVERT: 366b9fca TINYML_ALGO-161- Changes for preprocessing handoff
REVERT: cc97891c TINYML_ALGO-158 : Ability to let the user choose to keep the libc files for compilation or not TINYML_ALGO-159	: Integrate Modelmaker with TI MCU NNC 1.3.0
REVERT: 347ef83d Added FFT transform in MotorFaultDataset
REVERT: b30c3f8d TINYML_ALGO-154: BYOM for testing an external model
REVERT: 10ca2f40 Working code of BYOM for testing. (Not without training)
REVERT: a2104124 Addressing EDGEST-954
REVERT: 02e67366 TINYML_ALGO-148 . Progress bar added for datafile loading.
REVERT: f878fd8b TINYML_ALGO-146 . Graceful handling of incompatible files in the dataset
REVERT: ca53a101 Updated confusion matrix display format to be a lot more intuitive
REVERT: dffe80b4 Version update from 0.6 to 0.7. Logger update in train.py
REVERT: e8103a4f Best Epoch metrics are printed at the end of each train run.
REVERT: 5da38536 Added ArcFaultDataset, MotorFaultDataset. TINYML_ALGO-136
REVERT: 9b97c914 Generates a global.h header file in golden_test_vectors directory for easier verification in C code
REVERT: 35a02224 TINYML_ALGO-131, TINYML_ALGO-141: Ability to verify a test set either separately/alongwith the dataset
REVERT: dd9a371d Saving last frame of raw data for motor bearing fault as well
REVERT: 864aa798 Changed TinyML to Tiny ML
REVERT: 2b156b2c Bug Fix for TINYML_ALGO-75 in motor fault section. All frames are kept in raw data
REVERT: 85dda4a7 Bug Fix for TINYML_ALGO-75 in arc fault section. Now only last frame is kept in raw data
REVERT: d9f91907 Pull request #3: TINYML_ALGO-75 : Added raw golden vectors
REVERT: 0a88d500 TINYML_ALGO-75 : Added raw golden vectors
REVERT: e2f09ec1 Preparation for v0.6 release
REVERT: 96afbe59 Minor update in generic_models.py
REVERT: 3f18a26c Golden Test Vectors: Raw adc data was not capturing correctly when num_frame_concat was >1
REVERT: a16396c7 export_model avoid warning message
REVERT: 8b3fcae6 export_model  - avoid warning - changed default opset_version to 17
REVERT: 27edfb23 np.savetxt uses {} instead of []
REVERT: 2f297de5 .npy golden test vectors aren't generated anymore
REVERT: 7480cfb1 Updated timeseries_dataset.py with raw adc test vectors. Updated train.py with golden text vectors now being .txt instead of .npy
REVERT: f3fc7d53 Code aesthetics cleanup
REVERT: f9912949 Updated NNC with Software DL link
REVERT: 42216096 Minor logger text change
REVERT: acebf97b class labels are in sorted order. Changed in timeseries_dataset.py. TINYML_ALGO-97
REVERT: 17a507e9 Version number update v0.2 to "0.5.0"
REVERT: 75f2eb70 Minor change in using kwargs.get() in timeseries_dataset.py
REVERT: fbfe2a4f LICENSE updated
REVERT: fe1605e7 LICENSE updated
REVERT: 20f5fd97 generic models - bug fixed and updated DictPlus to same layer name repeating
REVERT: b12b5514 generic models updated - avoiding 1x1 convolution
REVERT: 5ccd9c23 generic models updated (no padding for conv/maxpool)
REVERT: 4328ce4b bugfix in model name
REVERT: e5bf3be4 generic model definitions updated
REVERT: 5ec56386 easy dict update using DictPuls
REVERT: fdaf1c77 Minor bug fixes
REVERT: 893b8c2f Feature Addition: Enabled IP protection using model_spec
REVERT: cea17b9f Updated for Thomas's preprocessing modifications. Infer ch,wl,hl from stacking parameter addition
REVERT: 3a0ba4d2 Updated for Thomas's preprocessing modifications. Infer ch,wl,hl from stacking parameter addition
REVERT: 43d49fa2 TINYML_ALGO-81 -feature_size_per_frame is taken as input instead of feature_size. Changed CNN_AF_3L_LARGE to CNN_AF_3L_1400. Added a new motor fault kilby model
REVERT: 036e6c1d generic_model_spec - can be a dict or list, write out model to yaml
REVERT: f26335da changes and fixes in generic_model_spec
REVERT: d25c5573 TINYML_ALGO-78 -> Kilby models are neither packaged nor shown in logs
REVERT: 5fff5e29 Changed TINIE to TINPU, tinie to ti_npu. Updated TVM wheel to final 1.2.0 tag release in requirements
REVERT: aa4983c1 daatset.samples_in_sequence has been replaced with input_features (dataset.X.shape[2])
REVERT: 3c281a2d Temporary Fix for a bug. args.variables gets changed based on dataset.X.shape[1] to accommodate concatenation of motor fault data
REVERT: cac3c8da Mid way progress in integrating motor fault. Arc fault code restabilised
REVERT: 7f364e58 1. Added changes to enable motor fault pre-processing
REVERT: ec7abd3b Pull request #2: added model with spec
REVERT: 61a777f5 generic models are working now
REVERT: d69351a3 generic_model_spec separated out into a different file
REVERT: b77719b7 Added generation of golden vectors for model verification. By default it is generated
REVERT: 9670333d changed create_ functions to init_ in ModelSpec class
REVERT: a9b662be layer types restricted to what is in tinynn
REVERT: cfd5bede layer_types support added to create_model_spec
REVERT: ce09e456 added model with spec
REVERT: e13f294f Model summary shown during training. Added support for CNN_TS_GEN_BASE_13K in generic models. Moved kilby models to a separate kilby_models.py file. Cleaner reporting of confusion matrix
REVERT: 11c6f2db Minor logging bug
REVERT: c7ac9610 Training saves best validation accuracy checkpoint instead of all checkpoints/last checkpoint. Prints dataset label map. Confusion matrix description is much more cleaner. TINYML_ALGO-73 TINYML_ALGO-74
REVERT: 348bd063 app.c will not be geenrated anymore until proper code is developed. Changed TinyML to Tiny ML to avoid copyright issues. Changed defaults of CNN_AF_3L (num_channel_l1=32->12, num_channel=16->12). x_temp_raw_out is also saved in feat_ext_data
REVERT: 1261db3b TINYML_ALGO-72 bug fix
REVERT: f02d468a Updated requirements with nnc rc3
REVERT: 9ba3ba0d Added models: ArcFault_cnn_largest_t(CNN_AF_3L_LARGE), ArcFault_cnn_200(CNN_AF_3L_200), ArcFault_cnn_300(CNN_AF_3L_300), ArcFault_cnn_700(CNN_AF_3L_700). Updated documentation accordingly
REVERT: 26408283 timeseries_dataset.py had been hard coded to support arc fault labels. Removed this and made it generalised
REVERT: cacae58d Replaced dc-remove with min-fft-bin. Added fft-bin-size in train.py. Corresponding changes in timeseries_dataset. Additional feature support to remove dataset's column which has 'time' mentioned in it
REVERT: 2d40b183 Updated with TVM v1.2
REVERT: 6fadd91f Updated with the latest TVM wheel v1.2
REVERT: 14c790f5 version updated
REVERT: 97470336 Bug fix for class imbalance when dataset.Y is a list
REVERT: 607ab40c app.c in compilation is written inside artifacts dir. train.py/load_data() has a separate utility for loading just test data. Better way to print confusion matrix
REVERT: eabcee80 Store Feature Extracted Data (+ Store only without training), Class Imbalance Wwights included for training. Resolved bug causing default optimise to be adam instead of sgd. Multiclass confusion matrix is now reporting normalized values
REVERT: 18c45405 Automate deletion of intermittent .bin .txt  Tinie files
REVERT: cb5ceec4 Updated with the latest TVM Compilation API
REVERT: b2b79b37 Changes requested by Lei to preprocessing
REVERT: 4b9726ab Version update for torcheval in requirements
REVERT: 6bcfe0f0 Added basic Jittering code. Not tested
REVERT: 33b0bbc3 Added confusion matrix and f1 score metrics
REVERT: 46232f96 requirements updated
REVERT: 5cb867da removed dependencies that are not needed
REVERT: 9068ca2a torchscript model export added
REVERT: f30eaf5b Removed accelerator mode argument, code cleanup
REVERT: 74da451e Updated installation of gpu version of tinyverse
REVERT: 8e8ea6b9 Support for setup_gpu
REVERT: 285889da Compilation options updated to latest version of tvm
REVERT: c569cfe7 Added resampling factor support
REVERT: f8ffd4c3 Logger name changes for Float and Quant Training
REVERT: 9a64a121 Appends float train and quant logs together
REVERT: 0ea150f5 setup file with cuda support
REVERT: 5e8b55db model export - changes to ensure that batch size is 1 in onnx model
REVERT: 8cd1583d LICENSE file added
REVERT: 859cf803 Added input_features to be passed from train.py to model. Bug fixes in Motor Fault model
REVERT: e2a9dcb6 fix for small batch size
REVERT: f8e2bc73 dataset generalization fix
REVERT: 7f2e77df quant measure_stats is supported only by some quant modes
REVERT: e6d17ede quantization_error_logging - convert error stats simplification
REVERT: 1545cc84 Generalized into single dataset loader
REVERT: 77cb8644 Figured a bug introduced during export model which caused the batch size to be present in the model
REVERT: 7882b359 Updated logger to allow printing messages outside the process
REVERT: f5e4a4f4 quantization-error-logging option added
REVERT: d021caca use actual input for model export
REVERT: 087e0692 removed folders that are not needed now, updated copyright headers
REVERT: 0b2201ed Commented out unnecessary printing of Dataset Loaders
REVERT: d5e4b164 with_input_batchnorm option to enable same model for TINIE v/s non TINIE devices
REVERT: 0e470533 Added with_input_batchnorm as an argument to models
REVERT: 811fb567 with_input_batchnorm option added for models
REVERT: 92bddd3f model-quant-format options is not required, quantization option is used
REVERT: 8d5b3a40 quantization choices uodated in training script
REVERT: f4db4e33 Pull request #1: updates for TINIE compatible QAT
REVERT: 28e329f4 TinyMLModelQuantFormat name change
REVERT: 242e3e2b qat interface changes
REVERT: ab851b23 model export cleanup
REVERT: 2ed100fe model export using tinyml-modeloptimization
REVERT: 6081fb6a updates for TINIE compatible QAT
REVERT: 5a5423c1 Feature Extraction Preset Addition, Model naming conventions updated
REVERT: 32c9d2e9 compilation.py added arg: tinie_mode
REVERT: f04b0f26 Added kilby models for Arc Fault and Motor Fault detection. Renamed models as well
REVERT: 8dadef3a Feature Update: Added Kilby Models for Arc Fault Detection, Added ability to choose data loaders, Added argumnets for feature extraction
REVERT: 5bb89b62 Feature Update: Inter compatibility of feature extractors with Models. Models can also be given a config to change the parameters of the model without touching the definitions - model-config
REVERT: 834d602a Better handoff between preprocessings and transforms. Updated a parameter in tvm_input_config.py to handle latest tvm. Support for output_ndim in audio_dataset.py . There is an export problem for post qat matchboxnet.py .
REVERT: b65e23c2 Named version as 0.1
REVERT: 4b134738 Bug fix: setup_cpu.sh installs tinyverse package
REVERT: b28352a5 TVM build version updated to be latest always. (480-> lastSuccessfulBuild)
REVERT: 1ab6960f TVM build version updated
REVERT: 33e6ca8c Ease of use to setup.
REVERT: 48b0e6e3 Ease of use to setup.
REVERT: 14e49d28 Ease of use to setup.
REVERT: 54802c6f Git ignorable files removed
REVERT: c83e0d35 Backs up QDQ model during QAT
REVERT: 586234ae Feature update: Conversion of QDQ model to INT8 model
REVERT: 98c3889c Feature update for QAT enabled training and model compilation
REVERT: 74cb253c Feature update for QAT enabled training and model compilation
REVERT: 6f62c199 Updated README.md
REVERT: 9a343799 Moved audio transforms to a separate transforms directory to maintain consistency
REVERT: 8b13815c Updated with README and minor bugs in audio models
REVERT: 1ea5da67 Added Haar Wavelet Transform
REVERT: 88e8b8d6 Tested STFT and MFCC to be working. Added few more combinations of TCResnet and MatchboxNet
REVERT: 67df8421 Enabled dataset loading for audio data
REVERT: afc7aa49 Enabled dataset loading for timeseries data
REVERT: 973018e5 Added MatchboxNet
REVERT: b0e50651 Updated with compilation support for models with multi input/multi output. Also minor debugs in train.py for audio and timeseries.
REVERT: ca8b2916 Updated with compilation support for C28 + Soft TINIE
REVERT: 074e0556 Added missing file that contains models added
REVERT: 349a76ce Tested Multivariate Timeseries Classification
REVERT: bd198a7a Feature Added: Added STFT, MFCC Preprocessing and Temporal Convolution Resnet Model
REVERT: 995fe41b Feature Added: Audio Classification
REVERT: f1b31fe8 Updated compilation.py for minor argparse bugs
REVERT: b2616547 Updated compilation script support for cmsis-nn mode
REVERT: 372d07f1 Updated code for multivariate time series data support
REVERT: ec2bd45d Train.py works even if transforms are not chosen
REVERT: 423e720f Updated train.py and compilation.py scripts with Loggers
REVERT: 9e1e85ed Updated app.c for better UX
REVERT: a808c34f First cut version of TinyVerse that can run arc detection
REVERT: 5b3297dc Initial Commit
REVERT: 4d52a124 Initial Commit
REVERT: 0dd276fa Initial Commit

git-subtree-dir: tinyml-tinyverse
git-subtree-split: 23a2fa30f5e5cecb201045a3a2a4dca365f7465d