The Coral USB Accelerator is a USB device that provides an Edge TPU as a coprocessor for your computer. It accelerates inferencing for your machine learning models when attached to a Linux host computer.
This page is your guide to get started. All you need to do is download our Edge TPU runtime and Python library to the host computer where you'll connect the USB Accelerator. Then you can begin running TensorFlow Lite models.
If you want to learn more about the USB Accelerator, see the datasheet.
- Any Linux computer with a USB port
- Debian 6.0 or higher, or any derivative thereof (such as Ubuntu 10.0+)
- System architecture of either x86-64 or ARM32/64 with ARMv8 instruction set
And yes, this means Raspberry Pi is supported. However, we officially support only Raspberry Pi 2/3 Model B/B+ running Raspbian. Unofficially, support for Raspberry Pi Zero is also available (install the TAR from GitHub on your Pi Zero, instead of the one below).
Also note that you should use a USB 3.0 port for the best inference speeds (unfortunately, Raspberry Pi has USB 2.0 ports only, but it still works).
Set up on Linux or Raspberry Pi
To get started, perform the following steps on your Linux machine or Raspberry Pi that will connect to the USB Accelerator.
Install the Edge TPU runtime and Python library:
wget https://dl.google.com/coral/edgetpu_api/edgetpu_api_latest.tar.gz -O edgetpu_api.tar.gz --trust-server-names
tar xzf edgetpu_api.tar.gz
bash ./install.shCaution: During installation, you'll be asked, "Would you like to enable the maximum operating frequency?" Enabling this option improves the the inferencing speed but it also causes the USB Accelerator to become very hot to the touch during operation and might cause burn injuries. If you're not sure you need the increased performance, type
Nand press Enter to use the default operating frequency. (You can change this later by simply re-running the install script.)Help! If you see the message
./install.sh: line 116: python3.5: command not found, then the install failed because you don't have Python 3.5 installed. So type
python3 --versionand press Enter. If it prints
Python 3.6or higher, then open the
install.shscript and edit the very last line to use
python3.5. However, if your Python version is lower than 3.5, you need to install Python 3.5.
Plug in the Accelerator using the provided USB 3.0 cable. (If you already plugged it in, remove it and replug it so the just-installed udev rule can take effect.)
Run a model on the Edge TPU
Now that your USB Accelerator is setup, you can start running TensorFlow Lite models on the Edge TPU. Follow these steps to perform image classification with one of our pre-compiled models and sample scripts.
First, download our bird classifier model, labels file, and photo:
cd ~/Downloads/ wget https://dl.google.com/coral/canned_models/mobilenet_v2_1.0_224_inat_bird_quant_edgetpu.tflite \ https://dl.google.com/coral/canned_models/inat_bird_labels.txt \ https://coral.withgoogle.com/static/images/parrot.jpg
Now navigate to the directory where we've shared the sample scripts and run image classification with the parrot image (shown in figure 1):
# on Debian/Ubuntu Linux: cd /usr/local/lib/python3.6/dist-packages/edgetpu/demo # on Raspberry Pi: cd /usr/local/lib/python3.5/dist-packages/edgetpu/demo python3 classify_image.py \ --model ~/Downloads/mobilenet_v2_1.0_224_inat_bird_quant_edgetpu.tflite \ --label ~/Downloads/inat_bird_labels.txt \ --image ~/Downloads/parrot.jpg
You should see results like this:
--------------------------- Ara macao (Scarlet Macaw) Score : 0.761719
Congrats! You've just performed an inference on the Edge TPU.
ERROR: Failed to retrieve TPU context
ERROR: Node number 0 (edgetpu-custom-op) failed to prepare
Then either your USB Accelerator is not plugged in or you cannot access it because your user account is not in the
plugdevsystem group. Ask your system admin to add your account to the
plugdevgroup, and then restart your computer for it to take effect.
This demo uses a Python API we created that makes it easy to perform an image classification or object detection inference on the Edge TPU. To learn more about the API, see the Edge TPU API overview & demos.
For details about how to create compatible TensorFlow Lite models, read TensorFlow Models on the Edge TPU.
Is this content helpful?