We have evaluated three different aspects of the models:
- Learning/(re)training time
- Storage requirement
- Performance (accuracy)
In terms of training time, in a desktop (Intel Xenon CPU E5-1650 v3 @3.5 GHz and 32 GB RAM) with GPU support, LSTM and CNN 1D on the ECG data took more than one hour, and MobileNet v1 on the acne dataset took less than 1 hour.
The storage requirement of a model is an essential consideration in resource-constrained IoT devices. The following screenshot presents the storage requirements for the three models we tested for the two use cases:
As shown, a saved model of LSTM took 234 MB of storage, CNN 1D took 8.5 MB, and MobileNet v1 (CNN) took 16.3 MB. In terms of storage requirements, all of the models, except the current version of LSTM, are fine to be deployed in many resource-constrained IoT devices, including Raspberry Pi 3 or smartphones...
