ARM-software/ML-KWS-for-MCU issues and pull requests

#149 - Requirements.txt

Issue - State: open - Opened by huseyin-karaca 11 months ago

#148 - reproduction in keras doesn't give same accuracy

Issue - State: open - Opened by asadisina over 2 years ago - 1 comment

#147 - How to run a DNN model with 436 neurons on the MCU?

Issue - State: open - Opened by Bosch936 about 3 years ago

#146 - link

Issue - State: open - Opened by lisem777 over 3 years ago

#145 - mbed error

Issue - State: open - Opened by aktoey over 3 years ago - 1 comment

#144 - mbed error

Issue - State: closed - Opened by aktoey over 3 years ago

#143 - Completely wrong predictions on own wav files

Issue - State: open - Opened by oezguensi almost 4 years ago - 1 comment

#142 - error to run label_wav

Issue - State: open - Opened by TheMasterRoot over 4 years ago - 1 comment

#141 - windows fatal exception when run label.py

Issue - State: closed - Opened by Ibrahfaye over 4 years ago - 1 comment

#140 - update to TensorFlow 2

Issue - State: open - Opened by upcoders over 4 years ago - 5 comments

#139 - problem with the Speed of training

Issue - State: open - Opened by wangohaha over 4 years ago

#138 - No output received after porting this example to tmpm46b cortex m4 mcu

Issue - State: closed - Opened by vinuraj1010 over 4 years ago - 1 comment

#137 - Problem with the DS-CNN model

Issue - State: open - Opened by wangohaha over 4 years ago - 1 comment

#136 - How to implement pb file in Android App

Issue - State: open - Opened by Xxfore over 4 years ago - 1 comment

#135 - Required tensorflow version?

Issue - State: open - Opened by sailor88128 over 4 years ago - 1 comment

#134 - Pre-defined Models Quantization

Issue - State: closed - Opened by mehrdadh over 4 years ago - 1 comment

#133 - false detection on new keywords

Issue - State: open - Opened by dimanshu over 4 years ago

#132 - I have a problem about Accuracy of models

Issue - State: open - Opened by hungptit123 over 4 years ago

#131 - train on custom dataset

Issue - State: open - Opened by dimanshu over 4 years ago

#130 - mbed deploy failed.

Issue - State: closed - Opened by xwJohn over 4 years ago - 2 comments

#129 - undefined reference to `__VECTOR_RAM' when build simple_test_k64f_gcc

Issue - State: closed - Opened by Mountain-with-sea over 4 years ago

#128 - Defined names in weights.h different from ds_cnn_weights.h

Issue - State: open - Opened by arvoelke over 4 years ago - 1 comment

#127 - details about ds_cnn.cpp and ds_cnn.h

Issue - State: open - Opened by XuLw over 4 years ago

#126 - Cannot write weights to weights.h for DS_CNN

Issue - State: closed - Opened by Ankitj03 over 4 years ago

#125 - Quantization. How to decide the best settings for act_max parameters?

Issue - State: open - Opened by jessonwtp almost 5 years ago

#124 - Quantization, what's the differenct between label_wav.py & C code

Issue - State: open - Opened by jasonwu1977 almost 5 years ago

#123 - act_max

Issue - State: open - Opened by ShintaZachariah almost 5 years ago - 2 comments

#122 - quant_test with different act_max parameters but produces same content weights.h ?

Issue - State: closed - Opened by jessonwtp almost 5 years ago - 3 comments

#121 - Deployment of GRU on ARM

Issue - State: open - Opened by pramodjayram almost 5 years ago

#120 - ValueError : The passed save_path is not a valid checkpoint

Issue - State: open - Opened by Shenfanlin almost 5 years ago - 3 comments

#119 - configs for all model are needed to do re-train and test

Issue - State: closed - Opened by nejyeah almost 5 years ago

#118 - 【Help】How to modify the training configuration parameters if my sample duration is 1.5 seconds? Thanks

Issue - State: open - Opened by xjj210130 almost 5 years ago - 2 comments

#117 - Optimal parameters for DS-CNN

Issue - State: open - Opened by idsparco97 about 5 years ago

#116 - How to convert .pb file into C colde like "ds_cnn_weights.h" file?

Issue - State: closed - Opened by gilbertjuly about 5 years ago

#115 - What does it mean that the memory refered in your paper?

Issue - State: open - Opened by Haryslee over 5 years ago

#114 - fft_len_padded representation

Issue - State: closed - Opened by marjanemd over 5 years ago

#113 - Ranges of the values in WAVE_DATA

Issue - State: open - Opened by marjanemd over 5 years ago - 1 comment

#112 - Error run quant_test.py to quantize ds_cnn

Issue - State: open - Opened by Haryslee over 5 years ago - 4 comments

#111 - Saving checkpoints in test.py

Issue - State: closed - Opened by marjanemd over 5 years ago - 4 comments

#110 - Understanding the calculation of Qm.n and bias_shift / out_shift for DS-CNN

Issue - State: open - Opened by devzeb over 5 years ago - 1 comment

#109 - Run inference on live audio on STM32F746NG

Issue - State: open - Opened by eera-l over 5 years ago

#108 - How to test trained model on custom testing dataset

Issue - State: open - Opened by saraalemadi over 5 years ago

#107 - Training the model with GFCCs

Issue - State: open - Opened by saichand07 over 5 years ago - 5 comments

#106 - How to decide the value of MFCC_DEC_BITS

Issue - State: closed - Opened by rayeh5 over 5 years ago - 2 comments

#105 - How to Find weight min/max by python

Issue - State: open - Opened by ccu1tn over 5 years ago - 1 comment

#104 - deployment on a mcu with no fpu

Issue - State: open - Opened by sacchi73 over 5 years ago - 1 comment

#103 - Bug in C++ pooling layer code affecting DS-CNN classification results

Issue - State: open - Opened by tpeet over 5 years ago

#102 - about training samples.

Issue - State: open - Opened by ccnankai over 5 years ago - 7 comments

#101 - Update Quant_guide.md

Pull Request - State: open - Opened by BabaJaguska over 5 years ago

#100 - Run KWS inference on live audio on Other development kit

Issue - State: open - Opened by romaindonze over 5 years ago - 1 comment

#99 - The missing pre-emphasis

Issue - State: closed - Opened by majianjia over 5 years ago - 1 comment

#98 - How to change Qm.n format of the layers

Issue - State: closed - Opened by saichand07 over 5 years ago - 1 comment

#97 - Reasons for the Random keyword detection on board ?

Issue - State: closed - Opened by saichand07 over 5 years ago - 11 comments

#96 - Calculate a negative value in bias_shift (shift right)

Issue - State: closed - Opened by ccnankai over 5 years ago - 15 comments

#95 - change wav file to 1500ms ,get poor result

Issue - State: closed - Opened by ccnankai over 5 years ago - 11 comments

#94 - multiple definitions of math functions in kws_simple_test

Issue - State: open - Opened by jeremyholleman over 5 years ago - 7 comments

#93 - the accuracy is very low when using my own data training models(on board)

Issue - State: open - Opened by lancezhangsf over 5 years ago - 8 comments

#92 - How to quant dnn

Issue - State: closed - Opened by barrycool over 5 years ago - 5 comments

#91 - [Error] ds_cnn_weights.h@1,28988: too many initializers for 'const q7_t [2560] {aka const sign

Issue - State: closed - Opened by lancezhangsf over 5 years ago - 6 comments

#90 - How to decide the bias_shift and out_shift for arm_fully_connected_q7?

Issue - State: closed - Opened by bkanakietac over 5 years ago

#89 - Memory sizes of the MFCC variables

Issue - State: closed - Opened by saichand07 over 5 years ago - 4 comments

#88 - Memory allocation Dynamic to Static

Issue - State: closed - Opened by saichand07 over 5 years ago

#87 - Deployment STM32F746NG

Issue - State: closed - Opened by saichand07 over 5 years ago - 1 comment

#87 - Deployment STM32F746NG

Issue - State: closed - Opened by saichand07 over 5 years ago - 1 comment

#86 - WARNING: MBED_ARM_PATH set as environment variable but doesn't exist

Issue - State: closed - Opened by saichand07 almost 6 years ago - 2 comments

#85 - Code conversion to cpp

Issue - State: closed - Opened by saichand07 almost 6 years ago

#84 - Change of wanted_words resulting very low recognition accuracy (only 24%)

Issue - State: closed - Opened by auvilink almost 6 years ago - 15 comments

#83 - Deployment of CRNN on STM32F746 mcu

Issue - State: open - Opened by saichand07 almost 6 years ago - 10 comments

#82 - what is the FFT size used in spectrogram/mfcc?

Issue - State: open - Opened by jenweilee almost 6 years ago - 2 comments

#81 - About negative samples

Issue - State: closed - Opened by phecda-xu almost 6 years ago - 1 comment

#80 - How to convert our Python code into Cpp for deployement on Arm_cortex MCU ?

Issue - State: closed - Opened by saichand07 almost 6 years ago

#79 - Is there any company or providers who will create a voice commands Dataset with user specified "wanted words" ?

Issue - State: closed - Opened by saichand07 almost 6 years ago

#78 - Error training ds_cnn model

Issue - State: closed - Opened by auvilink almost 6 years ago

#77 - the accuracy is very low when using bnfused ckpt

Issue - State: closed - Opened by remme123 almost 6 years ago - 1 comment

#76 - Are the dimensions entered in the model.py file and the deployment files different?

Issue - State: open - Opened by Hust-ShuaiWang almost 6 years ago - 1 comment

#75 - A Bug in quant_models.py

Issue - State: open - Opened by Holded almost 6 years ago - 4 comments

#74 - 12

Issue - State: closed - Opened by vrushalibhokare almost 6 years ago

#73 - Remove hardware dependencies

Issue - State: closed - Opened by ccnankai almost 6 years ago

#72 - mbed.h: No such file or directory

Issue - State: closed - Opened by yanyanem almost 6 years ago - 1 comment

#71 - How to convert these models to TF-Lite models and run it

Issue - State: open - Opened by rayeh5 almost 6 years ago - 4 comments

#71 - How to convert these models to TF-Lite models and run it

Issue - State: open - Opened by rayeh5 almost 6 years ago - 4 comments

#70 - How to determine the value of the parameter In the process of quantization

Issue - State: closed - Opened by ccnankai almost 6 years ago - 1 comment

#69 - How to execute smart mic with KWS

Issue - State: open - Opened by vrushalibhokare almost 6 years ago - 9 comments

#68 - act_max parameter for ds-cnn model

Issue - State: open - Opened by leiming0225 about 6 years ago - 3 comments

#67 - bug in quant_models.py

Issue - State: open - Opened by leiming0225 about 6 years ago

#66 - error run quant_test.py

Issue - State: closed - Opened by Huangxingjx about 6 years ago - 1 comment

#65 - error run quant_test.py

Issue - State: closed - Opened by OptimusPrime0 about 6 years ago - 2 comments

#64 - Deployment on other cortex chips

Issue - State: open - Opened by dpny518 about 6 years ago

#63 - Errors when running freeze.py

Issue - State: closed - Opened by muskedunder about 6 years ago - 2 comments

#62 - Generation of ds_cnn_weights.h

Issue - State: open - Opened by gamnes about 6 years ago - 14 comments

#61 - Accuracy on testing_list.txt

Issue - State: closed - Opened by LeighSu about 6 years ago - 1 comment

#60 - k64f build failure

Issue - State: closed - Opened by JammyZhou about 6 years ago - 2 comments

#59 - Mandarin

Issue - State: closed - Opened by xxoospring about 6 years ago - 2 comments

#58 - BSP and AUDIO_DISCO libraries for other mbed targets

Issue - State: closed - Opened by sharmaprakhar over 6 years ago - 1 comment

#57 - How do I freeze a quantized model

Issue - State: closed - Opened by XiangZheng2017 over 6 years ago

#56 - Examples\realtime_test\main.cpp line:79&95

Issue - State: closed - Opened by xujieuse over 6 years ago - 2 comments

#55 - Quantize weight

Issue - State: closed - Opened by XiangZheng2017 over 6 years ago

#54 - Bad error flow in input_data.py

Issue - State: open - Opened by OrDavidi over 6 years ago - 1 comment

#53 - Quantization and scaling

Issue - State: closed - Opened by pooyaww over 6 years ago - 7 comments

#52 - quant_test returns incorrect number of fraq bits?

Issue - State: closed - Opened by varteev over 6 years ago - 3 comments

GitHub / ARM-software/ML-KWS-for-MCU issues and pull requests