This is a fully quantized version (asymmetrical int8) of the MicroNet VWW-2 model developed by Arm, from the MicroNets paper. It is trained on the 'Visual Wake Words' dataset, more information can be found here: https://arxiv.org/pdf/1906.05721.pdf.
Apache-2.0
The class labels associated with this model can be created by running the script get_class_labels.sh.
| Platform |
Optimized |
| Cortex-A |
✖️ |
| Cortex-M |
✔️ |
| Mali GPU |
✔️ |
| Ethos U |
✔️ |
- ✔️ - Will run on this platform.
- ✖️ - Will not run on this platform.
Dataset: Visual Wake Words
| Metric |
Value |
| Accuracy |
0.768 |
| Optimization |
Value |
| Quantization |
INT8 |
| Input Node Name |
Shape |
Description |
| input |
(1, 50, 50, 1) |
A 50x50 input image. |
| Output Node Name |
Shape |
Description |
| Identity |
(1, 2) |
Per-class confidence across the two classes (0=no person present, 1=person present). |