ollewelin/pytorch-training-resnet50

PyTorch-Training-Resnet50. AUTO save cat (or other) images on external USB hard drive

Youtube Auto cat flap examples

https://www.youtube.com/watch?v=VDU2EMxDElc&t=34s

https://www.youtube.com/watch?v=emWBobbEPwY

More examples

Here with other texts in lable.txt

    class1   = Background
    class2   = Cat with prey
    class3   = Cat no prey

New cat flap (updated 2022-04-24) Stereo view 3D input with a tunnel of mirrors.

New installation

Older examples with other scale factor on the probablility output text (more then 100 %)

    $ python load_test_export.py

Related links

Installation guide Anaconda for Trainig on Ubuntu PC

Only need to do the installation step 1 to 6

https://github.com/ollewelin/Installing-and-Test-PyTorch-C-API-on-Ubuntu-with-GPU-enabled

Trainig on Ubuntu PC with or without GPU PyTorch

https://pytorch.org/tutorials/beginner/transfer_learning_tutorial.html

Interferance on Jetson Nano

https://github.com/dusty-nv/jetson-inference

Training and export to ONNX model on PC

Number of class settings in train.py file

...
# Here the size of each output sample is set to 3.
# Alternatively, it can be generalized to nn.Linear(num_ftrs, len(class_names)).
model_ft.fc = nn.Linear(num_ftrs, 3)  # change here number of classes
...

Data Path structure

$ ./data/hymenoptera_data/train/Background/imgXX.jpg
$ ./data/hymenoptera_data/train/Cat1/imgXX.jpg
$ ./data/hymenoptera_data/train/Cat2/imgXX.jpg
$ ./data/hymenoptera_data/val/Background/imgXX.jpg
$ ./data/hymenoptera_data/val/Cat1/imgXX.jpg
$ ./data/hymenoptera_data/val/Cat2/imgXX.jpg

Start train command

$ python train.py

Test result on PC test

Then close the two test windows to export to ONNX model

Install Jetson Nano Device

Follow instructions from :

Setting up Jetson with JetPack

https://github.com/dusty-nv/jetson-inference/blob/master/docs/jetpack-setup-2.md

Getting Started with Jetson Nano Developer Kit

https://developer.nvidia.com/embedded/learn/get-started-jetson-nano-devkit

Or

Getting Started with Jetson Nano 2GB Developer Kit

https://developer.nvidia.com/embedded/learn/get-started-jetson-nano-2gb-devkit

Run on Jetson Nano

imagenet3.cpp

Will collect image if class 1 > thresshold probablility and store on external USB hard drive ..

Copy over the Cat with prey model

Combine splitted gz file on PC and convert to ONNX

Back on the PC, get the splitted model

    olle@olle-TUF-Gaming-FX505DT-FX505DT:~$
    $ git clone https://github.com/ollewelin/PyTorch-Training-Resnet50

Combine splitted files

    $ cd PyTorch-Training-Resnet50
    $ cat model_31* > model_31.tar.gz
    $ gunzip model_31.tar.gz
    $ tar -xf model_31.tar
    $ cp model_31.pt model_saved.pt
    $ python load_test_export.py

Close windows..

    (base) olle@olle-TUF-Gaming-FX505DT-FX505DT:~/PyTorch-Training-Resnet50$ python load_test_export.py
    Hejvdsgdg
    Export model to ONNX 
    (base) olle@olle-TUF-Gaming-FX505DT-FX505DT:~/PyTorch-Training-Resnet50$ 

Copy over the model to Jetson Nano over ethernet

    $ cp model_saved.onnx model_31.onnx
    $ scp ~/PyTorch-Training-Resnet50/model_28.onnx [email protected].<xxx>.<xxx>:~/jetson-inference/examples/my-recognition

Start Docker container comand

Check IP address on local network

        $ ifconfig
        $ arp -a

Check IP address at Jetson nano module

 jetson@jetson-desktop:~$ ip addr 
 
 192.168.1.<xxx>

SSH from PC

olle@olle-TUF-Gaming-FX505DT-FX505DT:~$ ssh [email protected].<xxx>.<xxx>    

jetson@jetson-desktop:~$ sudo apt-get update

jetson@jetson-desktop:~$ git clone --recursive https://github.com/dusty-nv/jetson-inference

jetson@jetson-desktop:~$ git clone https://github.com/ollewelin/PyTorch-Training-Resnet50

jetson@jetson-desktop:~$ cp ~/PyTorch-Training-Resnet50/labels.txt ~/jetson-inference/examples/my-recognition

jetson@jetson-desktop:~$ cp ~/PyTorch-Training-Resnet50/CMakeLists.txt ~/jetson-inference/examples/my-recognition

jetson@jetson-desktop:~$ cp ~/PyTorch-Training-Resnet50/imagenet2.cpp ~/jetson-inference/examples/my-recognition

jetson@jetson-desktop:~$ cp ~/PyTorch-Training-Resnet50/imagenet3.cpp ~/jetson-inference/examples/my-recognition

jetson@jetson-desktop:~$ cp ~/PyTorch-Training-Resnet50/start.sh ~/jetson-inference/examples/my-recognition

jetson@jetson-desktop:~$ cd jetson-inference

jetson@jetson-desktop:~/jetson-inference$ docker/run.sh --volume /:/olle_test1

jetson@jetson-desktop:~/jetson-inference$ cd ../olle_test1/home/jetson/jetson-inference/examples/my-recognition

root@jetson-desktop:/olle_test1/home/jetson/jetson-inference/examples/my-recognition# cmake CMakeLists.txt  
-- The C compiler identification is GNU 7.5.0   
-- The CXX compiler identification is GNU 7.5.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Looking for pthread_create
-- Looking for pthread_create - not found
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE  
-- Found CUDA: /usr/local/cuda (found version "10.2") 
-- Configuring done
-- Generating done
-- Build files have been written to: /olle_test1/home/jetson/jetson-inference/examples/my-recognition
root@jetson-desktop:/olle_test1/home/jetson/jetson-inference/examples/my-recognition# make
Scanning dependencies of target imagenet2
[ 50%] Building CXX object CMakeFiles/imagenet2.dir/imagenet2.cpp.o
[100%] Linking CXX executable imagenet2
[100%] Built target imagenet2
root@jetson-desktop:/olle_test1/home/jetson/jetson-inference/examples/my-recognition# 

Or start docker with a start.sh script inside container

    $ cd jetson-inference

    $ docker/run.sh --volume /:/olle_test1 -r "bash ../olle_test1/home/jetson/jetson-inference/examples/my-recognition/start.sh"

Or all from bash start_script.sh

    $ cd jetson-inference

    $ ~/jetson-inference/bash start_script.sh      

Run program (first time TensorRT import the ONNX model) Start On Jetson Target

Cat example model_31.onnx

   $ root@jetson-desktop:/olle_test1/home/jetson/jetson-inference/examples/my-recognition#./imagenet2 --model=./model_31.onnx --input_blob=input --output_blob=output --labels=./labels.txt csi://0

Or less printout's and other thresshold Cat with prey

    $ ./imagenet3 --model=./model_31.onnx --input_blob=input --output_blob=output --labels=./labels.txt csi://0

Or first example

root@jetson-desktop:/olle_test1/home/jetson/jetson-inference/examples/my-recognition#./imagenet2 --model=./resnet50.onnx --input_blob=input --output_blob=output --labels=./labels.txt csi://0

Auto save image on external USB disk

Auto create image folder on external USB disk

Jetson Nano build from Source and auto start

Build from source (Instead of using Docker)

Jetson nano Build from source

https://github.com/dusty-nv/jetson-inference/blob/master/docs/building-repo-2.md

$ sudo apt-get update
$ sudo apt-get install git cmake libpython3-dev python3-numpy
$ git clone --recursive https://github.com/dusty-nv/jetson-inference
$ cd jetson-inference
$ mkdir build
$ cd build
$ cmake ../
$ make -j$(nproc)
$ sudo make install
$ sudo ldconfig

Start Jetson Nano directly after boot

Enable Auto login Jetson Nano module

Run the “User Accounts” app, then click the “Unlock” button in the top right of the window and enter your sudo password, and then you can enable/disable Automatic Login for your account.

Edit bash_completion.sh with inside Vim editor from PC SSH

SSH From PC side

ssh [email protected]
sudo vim ~/../../etc/profile.d/bash_completion.sh

Add start_d.sh script at bash_completion.sh with inside Vim editor

Vim guide

https://coderwall.com/p/adv71w/basic-vim-commands-for-getting-started

Edit Commands

R - to replace more than one character

R

Add line directly from Source (Not Docker) With streaming to explicitly IP number on LAN

bash ~/jetson-inference/examples/my-recognition/start_rtp.sh

Or line Docker use

bash ~/jetson-inference/start_script.sh

Save, Exit and reboot Jetson Nano

:wq - to save and exit

press ESC to set VIM to command mode

:wq

sudo reboot

Install gstreamer on PC Ubuntu

https://gstreamer.freedesktop.org/documentation/installing/on-linux.html?gi-language=c

sudo apt-get update
sudo apt-get install libgstreamer1.0-dev libgstreamer-plugins-base1.0-dev libgstreamer-plugins-bad1.0-dev gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-bad gstreamer1.0-plugins-ugly gstreamer1.0-libav gstreamer1.0-doc gstreamer1.0-tools gstreamer1.0-x gstreamer1.0-alsa gstreamer1.0-gl gstreamer1.0-gtk3 gstreamer1.0-qt5 gstreamer1.0-pulseaudio

Start gstreamer video view on PC

gst-launch-1.0 -v udpsrc port=1234  caps = "application/x-rtp, media=(string)video, clock-rate=(int)90000, encoding-name=(string)H264, payload=(int)96" !  rtph264depay ! decodebin ! videoconvert ! autovideosink