What does “Column names in .csv do not match the sensors you selected. (See QeexoAutoML_User_Guide)” mean?
When uploading CSV files, you will be asked to create an Environment and select the sensors and their configurations. What you select here must match with what is described in the CSV file.
Unable to perform Live Testing through Bluetooth connection. Error message shows “Bluetooth is not connected” even though the device is successfully paired.
Please make sure the “Discovery” option is not enabled in the Bluetooth Settings: Start -> Bluetooth & Other Device Setting-> More Bluetooth options -> Options tab -> Select Discovery (Allow Bluetooth devices to find this PC).
Training is stuck on “Calculating Latency” after everything else is complete in the “Real-Time Training Progress” view.
If the training appears to be stuck on Calculating Latency, it is likely due to the target hardware being disconnected from your machine. Since we calculate the latency by flashing the ML model onto the device to run live inferences and taking the average time, the process cannot be completed if the device is disconnected. You can exist out of the pop-up screen and still see Model training results without the latency. Click on the “Error” icon under the Latency column to try again after reconnecting the target hardware.
Unable to skip the “Group Labels” step when starting a new training.
If you are in a Single-class classification project, multiple labels must be grouped into one single group before training can continue. If you are in a Multi-class classification project, you are able to skip this step.
The “START NEW TRAINING” button is disabled and error shows “Training may not begin until the previous training or upload has completed.”
Please make sure that there are no other trainings in session.
In the Models Page, “Latency” column shows “Error”.
Make sure the hardware is connected, then click the icon to measure latency again.
I fail to start a new training and see “Insufficient amount of data”.
The Notifications page will tell you how much more data is needed to start a training.
The button “FLASH DATA COLLECTION APP” is still greyed out and not clickable even though sensors have been selected.
Make sure to select the ODR and the FSR for the selected sensors.
The blue highlighted region on the visualized signal displayed during “Event” data recording does not cover the appropriate signal.
For Event data to work properly, there needs to be periods of rest at the beginning and end of the recording, with the actual “event” occurring in between. For example, if we are trying to collect a “C” gesture, we need to do the following: 1. Determine how long our “C” gesture will take. 2. Set the event length to be ~1 second longer than our length from step 1. 3. During collection, make sure the device is at rest during each start/stop operation – this guarantees that we capture the full event within the recording range (i.e. the time between “start” and “stop”). 4. Check the visualized signal to see whether the event appears clearly within the window and the appropriate portion is captured.
After collecting data, I cannot find it on the training page.
On the training page, different datasets with the same sensor configurations and same label names are grouped. Please click the drop-down icon to see all datasets.
Why is the performance of my models unsatisfactory? What can I do to improve model accuracy?
1. Go to the Details Page by clicking the “Details” icon in the Models Page, check Learning Curve to see whether you need more data. 2. Go to the Sensitivity Analysis page by clicking the “Sensitivity Analysis” icon and try adjusting weights and re-compiling the library. 3. For more information, please refer to the Qeexo AutoML User Guide.
Unable to flash Data Collection Application to the target hardware.
This could be caused by various situations. 1. Make sure your target hardware is connected and matches with the current project. 2. If your device in the Arduino Nano 33 BLE Sense, try pressing the reset button twice before starting to flash. 3. Try connecting the device via another port of your laptop. 4. Check if your computer is connected to the internet through a proxy. If so, temporarily disabling the proxy may fix the issue.
Unable to flash Data Collection Application or machine learning models to the Arduino Nano 33 BLE Sense.
Pressing the reset button twice before starting to flash should resolve this issue.
The correct Qeexo Device Manager is added to Chrome Extensions, but I still get the “Chrome Extension Mismatch” error.
This may happen if previous version(s) of the Qeexo Device Manager Chrome Extension still exists. Go to “Window-Extensions” to delete it.
Connecting target device to a Macbook which has USB-C port through a USB-hub device doesn’t work.
The typical current consumption for normal mode is 33mA and it goes down to 29mA when it is in flash mode. However, this current consumption is not always stable when we use a USB-hub device. Sometimes, the current doesn’t come back to 33mA even though device is in normal mode. This issue never happens when we use “USB-C to USB” converter instead of using USB-hub. Based on this, we think USB-hub may not be able to drive enough power for our target device operation.
See Our Popular Blogs
-
PRESS Sound Recognition with Qeexo AutoML
Introduction Sound Recognition is a technology based on traditional pattern recognition theories and signal analysis methods which is widely used in speech recognition, music recognition and many other research areas such as acoustical oceanography [1]. Generally, microphones are regarded as sufficient sensing modalities as input to machine learning methods within these fields. Microphones are capable […]
Zhongyu Ouyang and Dr. Geoffrey Newman 21 September 2020 -
PRESS Qeexo AutoML Enables Machine Learning on Arm Cortex-M0 and Cortex-M0+
First company to build an automated ML platform for the Arm Cortex-M0 and Cortex-M0+ processor MOUNTAIN VIEW, Calif. (PRWEB) September 09, 2020 Qeexo, developer of an automated machine learning (ML) platform that accelerates the deployment of tinyML at the edge, today announces that its Qeexo AutoML platform now supports machine learning on Arm® Cortex®-M0 and […]
Qeexo, Co. 09 September 2020 -
PRESS Inference Settings: Instance Length and Classification Interval
Qeexo AutoML enables machine learning application developers to customize inference settings based on their use-case. These parameters are critical for achieving the best live performance of models on the embedded target. In this article, we will discuss the two parameters associated with the inference settings; instance length and classification interval. Instance Length Instance length is […]
Xun (Jared) Liu, Dr. Rajen Bhatt, and Dr. Geoffrey Newman