Tar -xvf ILSVRC2012_img_train.tar # This took 48 minutesįind. This took some 2.5 hours for me.įor good measure get the test files too. It took some 1585 minutes (26.4 hours) to download the training images.Note also that nonpub has changed to nnoupd. Note: You need to use the username and password of your account.The path names for downloading the test and validation images are slightly wrong and wget needs your username and password.Instructions for downloading the various image data sets are here: īut there are a couple of things you have to do to make it work. This tells how to download the images.Here are instructions for training the network. The examples that use the GUI have an error unless you run gedit first. Use with care!Įxec=sudo bash -c "cd /home/debian/exercises/x15/tidl/tidl-api/examples/classification gedit &. Now debian doesn't need to enter a password when using sudo. Do this by added a line to the /etc/sudoers file. Here are some notes that I hope will lead up to the examples auto starting.įirst allow user debian to run sudo without a password. Layer_ouput and mcbench look like handy tools. ssd_multibox -d 2 -e 2 -i camera0 -w 1200 2> /dev/null Others The ssd_multibox example takes an image as input and detects multiple objects with bounding boxes according to pre-trained cd. SSD is the abbreviation for Single Shot multi-box Detector. segmentation -d 2 -e 2 -i camera0 -w 1200 2> /dev/null ssd_multibox The segmentation example takes an image as input and performs pixel-level classification according to pre-trained cd. It is looking for and highlights the last object it found. The following takes live video from a camera (/dev/video0) and displays it on the host. Stream_config_mobilenet.txt runs but it looks like the color channels are switched Stream_config_j11_v2.txt runs but gets the error "Corrupt JPEG data: 2 extraneous bytes before marker 0xd4". Stream_config_inceptionnet.txt seems to have a file missing. You need to login to the x15 as root for the X-Windows authentication to cd lsĪvg_fps_window.h imagenet1001.txt Makefile stream_config_mobilenet.txtĬlasslist.txt imagenet.txt readme.md tidl_classificationĬlips images stream_config_inceptionnet.txt tidl-sw-stack-small.pngįindclasses.cpp main.cpp stream_config_j11_v2.txt The classification demo is looking for one (or two if you set TWO_ROIs) object out of a small list of 12 or so things. The imagenet demo is looking for one object out of a list of 1000 things. Here are instructions for setting a root password, etc. You need to ssh as root for the X-Windows authentication to work. Install the image viewer "eye of gnome" for viewing images on the x15.įrom the host computer you need to ssh with the -XC flags so the x15 can access the host's X-windows toĭisplay things. X15$ sudo apt install libcanberra-gtk-module libcanberra-gtk3-module If you get Gtk-Message: Failed to load module "canberra-gtk-module", run X15$ cd /opt/scripts/tools/ git pull sudo. Here are a few other handy extras to install. User 1000 (should be debian) permission to read/write it. The next build puts things in /usr/share/ti/tidl so create it and assume give Install time 38s.Ĭheckout the most current branch and compile. Most were already installed and up to date. X15$ sudo apt install ti-opencl libboost-dev libopencv-core-dev libopencv-imgproc-dev libopencv-highgui-dev libjson-c-dev Now follow the instructions in the readme.md file.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |