Update README.md
Browse files
README.md
CHANGED
|
@@ -75,7 +75,11 @@ For Yamnet-1024
|
|
| 75 |
# Performances
|
| 76 |
|
| 77 |
## Metrics
|
| 78 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 79 |
|
| 80 |
### Reference **NPU** memory footprint based on ESC-10 dataset
|
| 81 |
|Model | Dataset | Format | Resolution | Series | Internal RAM (KiB) | External RAM (KiB) | Weights Flash (KiB) | STM32Cube.AI version | STEdgeAI Core version |
|
|
|
|
| 75 |
# Performances
|
| 76 |
|
| 77 |
## Metrics
|
| 78 |
+
|
| 79 |
+
* Measures are done with default STEDGEAI configuration with enabled input / output allocated option.
|
| 80 |
+
|
| 81 |
+
* `tl` stands for "transfer learning", meaning that the model backbone weights were initialized from a pre-trained model, then only the last layer was unfrozen during the training.
|
| 82 |
+
|
| 83 |
|
| 84 |
### Reference **NPU** memory footprint based on ESC-10 dataset
|
| 85 |
|Model | Dataset | Format | Resolution | Series | Internal RAM (KiB) | External RAM (KiB) | Weights Flash (KiB) | STM32Cube.AI version | STEdgeAI Core version |
|