Fix TensorFlow Lite micro_speech readme and Colab
PiperOrigin-RevId: 261575841
This commit is contained in:
Daniel Situnayake
TensorFlower Gardener
parent
ab8b627e0b
commit
80fa48dac2
@ -18,7 +18,6 @@ detection model, takes up a total of 22KB.
|
||||
* [Building for Ambiq Micro Apollo3Blue EVB using Make](#building-for-ambiq-micro-apollo3blue-evb-using-make)
|
||||
* [Additional Apollo3 Instructions](#additional-apollo3-instructions)
|
||||
* [Building for the Eta Compute ECM3531 EVB using Make](#Building-for-the-Eta-Compute-ECM3531-EVB-using-Make)
|
||||
* [Building for NXP FRDM K66F EVB using mbed](#Building-for-NXP-FRDM-K66F-using-mbed)
|
||||
|
||||
- [Goals](#goals)
|
||||
|
||||
@ -380,67 +379,6 @@ To flash a part with JFlash Lite, do the following:
|
||||
tensorflow/lite/experimental/micro/tools/make/targets/ecm3531 \
|
||||
./flash_program executable_name to load into flash.
|
||||
|
||||
## Building for NXP FRDM K66F using mbed
|
||||
|
||||
1. Follow the instructions at
|
||||
[Tensorflow Micro Speech](https://github.com/tensorflow/tensorflow/tree/master/tensorflow/lite/experimental/micro/examples/micro_speech#getting-started)
|
||||
to download the Tensorflow source code and the support libraries
|
||||
2. Follow instructions from [mbed website](https://os.mbed.com/docs/mbed-os/v5.13/tools/installation-and-setup.html) to setup and install mbed CLI
|
||||
3. Compile tensorflow with the following command to generate mbed project:
|
||||
|
||||
```
|
||||
make -f tensorflow/lite/experimental/micro/tools/make/Makefile TARGET=mbed TAGS="nxp_k66f" generate_micro_speech_mbed_project
|
||||
```
|
||||
4. Go to the location of the generated project. The generated project is usally
|
||||
in tensorflow/lite/experimental/micro/tools/make/gen/mbed_cortex-m4/prj/micro_speech/mbed
|
||||
5. Create a mbed project using the generated files: ```mbed new .```
|
||||
6. Change the project setting to use C++ 11 rather than C++ 14 using
|
||||
|
||||
```
|
||||
python -c 'import fileinput, glob;
|
||||
for filename in glob.glob("mbed-os/tools/profiles/*.json"):
|
||||
for line in fileinput.input(filename, inplace=True):
|
||||
print line.replace("\"-std=gnu++14\"","\"-std=c++11\", \"-fpermissive\"")'
|
||||
```
|
||||
7. To compile project, use the following command:
|
||||
|
||||
```
|
||||
mbed compile --target K66F --toolchain GCC_ARM --profile release
|
||||
```
|
||||
8. For some mbed compliers, you may get compile error in mbed_rtc_time.cpp.
|
||||
Go to mbed-os/platform/mbed_rtc_time.h and comment line 32 and line 37
|
||||
|
||||
```
|
||||
//#if !defined(__GNUC__) || defined(__CC_ARM) || defined(__clang__)
|
||||
struct timeval {
|
||||
time_t tv_sec;
|
||||
int32_t tv_usec;
|
||||
};
|
||||
//#endif
|
||||
```
|
||||
9. Look at helpful resources from NXP website such as [NXP FRDM-K66F User guide](https://www.nxp.com/docs/en/user-guide/FRDMK66FUG.pdf) and [NXP FRDM-K66F Getting Started](https://www.nxp.com/document/guide/get-started-with-the-frdm-k66f:NGS-FRDM-K66F)
|
||||
to understand information about the board.
|
||||
10. Connect USB cable to micro USB port. When ethernet port is face towards you,
|
||||
The micro USB port is left of the ethernet port.
|
||||
11. To compile and flash in a single step, add --flash option:
|
||||
|
||||
```
|
||||
mbed compile --target K66F --toolchain GCC_ARM --profile release --flash
|
||||
```
|
||||
12. Disconnect USB cable from the device to power down the device and connect
|
||||
back the power cable to start running the model
|
||||
13. Connect to serial port with baud rate of 9600 and correct serial device
|
||||
to view the output from the MCU. In linux, you can run the following screen
|
||||
command if the serial device is /dev/ttyACM0
|
||||
|
||||
```
|
||||
sudo screen /dev/ttyACM0 9600
|
||||
```
|
||||
14. Saying "Yes" will print "Yes" and "No" will print "No" on the serial port
|
||||
15. A loopback path from microphone to headset jack is enabled. Headset jack is
|
||||
in black color. If there is no output on the serial port, you can connect
|
||||
headphone to headphone port to check if audio loopback path is working
|
||||
|
||||
## Implement target optimized kernels
|
||||
|
||||
The reference kernels in tensorflow/lite/experimental/micro/kernels are
|
||||
|
||||
@ -20,8 +20,9 @@ kilobytes of Flash.
|
||||
- [Deploy to Arduino](#deploy-to-arduino)
|
||||
- [Deploy to SparkFun Edge](#deploy-to-sparkfun-edge)
|
||||
- [Deploy to STM32F746](#deploy-to-STM32F746)
|
||||
- [Deploy to NXP FRDM K66F](#deploy-to-nxp-frdm-k66f)
|
||||
- [Calculating the input to the neural network](#calculating-the-input-to-the-neural-network)
|
||||
- [Create your own model](#create-your-own-model)
|
||||
- [Train your own model](#train-your-own-model)
|
||||
|
||||
|
||||
## Getting started
|
||||
@ -31,6 +32,7 @@ This code has been tested on the following devices:
|
||||
* [SparkFun Edge](https://sparkfun.com/products/15170)
|
||||
* [Arduino Nano 33 BLE Sense](https://store.arduino.cc/usa/nano-33-ble-sense-with-headers)
|
||||
* [ST Microelectronics STM32F746G Discovery kit](https://os.mbed.com/platforms/ST-Discovery-F746NG/)
|
||||
* [NXP FRDM K66F](https://www.nxp.com/design/development-boards/freedom-development-boards/mcu-boards/freedom-development-platform-for-kinetis-k66-k65-and-k26-mcus:FRDM-K66F)
|
||||
|
||||
This readme contains instructions for building the code on Linux and macOS, and
|
||||
deploying the code to the above microcontroller platforms and macOS.
|
||||
@ -112,7 +114,6 @@ to [Arduino](https://www.arduino.cc/) devices.
|
||||
The sample has been tested with the following devices:
|
||||
|
||||
- [Arduino Nano 33 BLE Sense](https://store.arduino.cc/usa/nano-33-ble-sense-with-headers)
|
||||
(this ).
|
||||
|
||||
The Arduino Nano 33 BLE Sense is currently the only Arduino with a built-in
|
||||
microphone. If you're using a different Arduino board and attaching your own
|
||||
@ -136,7 +137,7 @@ modifications to the code), run this command to generate a zip file containing
|
||||
the required source files:
|
||||
|
||||
```
|
||||
make -f tensorflow/lite/experimental/micro/tools/make/Makefile TARGET=arduino TAGS="" generate_micro_speech_arduino_library_zip
|
||||
make -f tensorflow/lite/experimental/micro/tools/make/Makefile TARGET=arduino TAGS="portable_optimized" generate_micro_speech_arduino_library_zip
|
||||
```
|
||||
|
||||
A zip file will be created at the following location:
|
||||
@ -438,6 +439,69 @@ see will be at least 200.
|
||||
To stop viewing the debug output with `screen`, hit `Ctrl+A`, immediately
|
||||
followed by the `K` key, then hit the `Y` key.
|
||||
|
||||
## Deploy to NXP FRDM K66F
|
||||
|
||||
The following instructions will help you build and deploy the sample to the
|
||||
[NXP FRDM K66F](https://www.nxp.com/design/development-boards/freedom-development-boards/mcu-boards/freedom-development-platform-for-kinetis-k66-k65-and-k26-mcus:FRDM-K66F)
|
||||
using [ARM Mbed](https://github.com/ARMmbed/mbed-cli).
|
||||
|
||||
1. Download [the TensorFlow source code](https://github.com/tensorflow/tensorflow).
|
||||
2. Follow instructions from [mbed website](https://os.mbed.com/docs/mbed-os/v5.13/tools/installation-and-setup.html) to setup and install mbed CLI.
|
||||
3. Compile TensorFlow with the following command to generate mbed project:
|
||||
|
||||
```
|
||||
make -f tensorflow/lite/experimental/micro/tools/make/Makefile TARGET=mbed TAGS="nxp_k66f" generate_micro_speech_mbed_project
|
||||
```
|
||||
4. Go to the location of the generated project. The generated project is usally
|
||||
in `tensorflow/lite/experimental/micro/tools/make/gen/mbed_cortex-m4/prj/micro_speech/mbed`
|
||||
5. Create a mbed project using the generated files: `mbed new .`
|
||||
6. Change the project setting to use C++ 11 rather than C++ 14 using:
|
||||
|
||||
```
|
||||
python -c 'import fileinput, glob;
|
||||
for filename in glob.glob("mbed-os/tools/profiles/*.json"):
|
||||
for line in fileinput.input(filename, inplace=True):
|
||||
print line.replace("\"-std=gnu++14\"","\"-std=c++11\", \"-fpermissive\"")'
|
||||
```
|
||||
7. To compile project, use the following command:
|
||||
|
||||
```
|
||||
mbed compile --target K66F --toolchain GCC_ARM --profile release
|
||||
```
|
||||
8. For some mbed compliers, you may get compile error in mbed_rtc_time.cpp.
|
||||
Go to `mbed-os/platform/mbed_rtc_time.h` and comment line 32 and line 37:
|
||||
|
||||
```
|
||||
//#if !defined(__GNUC__) || defined(__CC_ARM) || defined(__clang__)
|
||||
struct timeval {
|
||||
time_t tv_sec;
|
||||
int32_t tv_usec;
|
||||
};
|
||||
//#endif
|
||||
```
|
||||
9. Look at helpful resources from NXP website such as [NXP FRDM-K66F User guide](https://www.nxp.com/docs/en/user-guide/FRDMK66FUG.pdf) and [NXP FRDM-K66F Getting Started](https://www.nxp.com/document/guide/get-started-with-the-frdm-k66f:NGS-FRDM-K66F)
|
||||
to understand information about the board.
|
||||
10. Connect USB cable to micro USB port. When ethernet port is face towards you,
|
||||
The micro USB port is left of the ethernet port.
|
||||
11. To compile and flash in a single step, add the `--flash` option:
|
||||
|
||||
```
|
||||
mbed compile --target K66F --toolchain GCC_ARM --profile release --flash
|
||||
```
|
||||
12. Disconnect USB cable from the device to power down the device and connect
|
||||
back the power cable to start running the model.
|
||||
13. Connect to serial port with baud rate of 9600 and correct serial device
|
||||
to view the output from the MCU. In linux, you can run the following screen
|
||||
command if the serial device is `/dev/ttyACM0`:
|
||||
|
||||
```
|
||||
sudo screen /dev/ttyACM0 9600
|
||||
```
|
||||
14. Saying "Yes" will print "Yes" and "No" will print "No" on the serial port.
|
||||
15. A loopback path from microphone to headset jack is enabled. Headset jack is
|
||||
in black color. If there is no output on the serial port, you can connect
|
||||
headphone to headphone port to check if audio loopback path is working.
|
||||
|
||||
## Calculating the input to the neural network
|
||||
|
||||
The TensorFlow Lite model doesn't take in raw audio sample data. Instead it
|
||||
|
||||
@ -36,12 +36,12 @@
|
||||
"\n",
|
||||
"The model is designed to be used with [Google Colaboratory](https://colab.research.google.com).\n",
|
||||
"\n",
|
||||
"<table class=\\\"tfo-notebook-buttons\\\" align=\\\"left\\\">\n",
|
||||
"<table class=\"tfo-notebook-buttons\" align=\"left\">\n",
|
||||
" <td>\n",
|
||||
" <a target=\\\"_blank\\\" href=\\\"https://colab.research.google.com/github/tensorflow/tensorflow/blob/master/tensorflow/lite/experimental/micro/examples/micro_speech/train_speech_model.ipynb\\\"><img src=\\\"https://www.tensorflow.org/images/colab_logo_32px.png\\\" />Run in Google Colab</a>\n",
|
||||
" <a target=\"_blank\" href=\"https://colab.research.google.com/github/tensorflow/tensorflow/blob/master/tensorflow/lite/experimental/micro/examples/micro_speech/train_speech_model.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />Run in Google Colab</a>\n",
|
||||
" </td>\n",
|
||||
" <td>\n",
|
||||
" <a target=\\\"_blank\\\" href=\\\"https://github.com/tensorflow/tensorflow/blob/master/tensorflow/lite/experimental/micro/examples/micro_speech/train_speech_model.ipynb\\\"><img src=\\\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\\\" />View source on GitHub</a>\n",
|
||||
" <a target=\"_blank\" href=\"https://github.com/tensorflow/tensorflow/blob/master/tensorflow/lite/experimental/micro/examples/micro_speech/train_speech_model.ipynb\"><img src=\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\" />View source on GitHub</a>\n",
|
||||
" </td>\n",
|
||||
"</table>\n",
|
||||
"\n",
|
||||
@ -116,7 +116,7 @@
|
||||
"source": [
|
||||
"# Install the nightly build\n",
|
||||
"!pip install -q tf-nightly-gpu==1.15.0.dev20190729\n",
|
||||
"!git clone https://github.com/dansitu/tensorflow"
|
||||
"!git clone https://github.com/tensorflow/tensorflow"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": []
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user