Update README.md
Browse files
README.md
CHANGED
|
@@ -66,7 +66,10 @@ For an image resolution of NxM and P classes
|
|
| 66 |
|
| 67 |
## Metrics
|
| 68 |
|
| 69 |
-
Measures are done with default STM32Cube.AI configuration with enabled input / output allocated option.
|
|
|
|
|
|
|
|
|
|
| 70 |
|
| 71 |
|
| 72 |
### Reference **NPU** memory footprint on food-101 and ImageNet dataset (see Accuracy for details on dataset)
|
|
@@ -166,7 +169,7 @@ Dataset details: [link](http://download.tensorflow.org/example_images/flower_pho
|
|
| 166 |
### Accuracy with Plant-village dataset
|
| 167 |
|
| 168 |
|
| 169 |
-
Dataset details: [link](https://data.mendeley.com/datasets/tywbtsjrjv/1)
|
| 170 |
|
| 171 |
| Model | Format | Resolution | Top 1 Accuracy |
|
| 172 |
|-------|--------|------------|----------------|
|
|
@@ -187,7 +190,7 @@ Dataset details: [link](https://data.mendeley.com/datasets/tywbtsjrjv/1) , Licen
|
|
| 187 |
### Accuracy with Food-101 dataset
|
| 188 |
|
| 189 |
|
| 190 |
-
Dataset details: [link](https://data.vision.ee.ethz.ch/cvl/datasets_extra/food-101/)
|
| 191 |
|
| 192 |
| Model | Format | Resolution | Top 1 Accuracy |
|
| 193 |
|-------|--------|------------|----------------|
|
|
@@ -209,7 +212,7 @@ Dataset details: [link](https://data.vision.ee.ethz.ch/cvl/datasets_extra/food-1
|
|
| 209 |
|
| 210 |
### Accuracy with ImageNet dataset
|
| 211 |
|
| 212 |
-
Dataset details: [link](https://www.image-net.org),
|
| 213 |
Number of classes: 1000.
|
| 214 |
To perform the quantization, we calibrated the activations with a random subset of the training set.
|
| 215 |
For the sake of simplicity, the accuracy reported here was estimated on the 50000 labelled images of the validation set.
|
|
|
|
| 66 |
|
| 67 |
## Metrics
|
| 68 |
|
| 69 |
+
- Measures are done with default STM32Cube.AI configuration with enabled input / output allocated option.
|
| 70 |
+
- `tfs` stands for "training from scratch", meaning that the model weights were randomly initialized before training.
|
| 71 |
+
- `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.
|
| 72 |
+
- `fft` stands for "full fine-tuning", meaning that the full model weights were initialized from a transfer learning pre-trained model, and all the layers were unfrozen during the training.
|
| 73 |
|
| 74 |
|
| 75 |
### Reference **NPU** memory footprint on food-101 and ImageNet dataset (see Accuracy for details on dataset)
|
|
|
|
| 169 |
### Accuracy with Plant-village dataset
|
| 170 |
|
| 171 |
|
| 172 |
+
Dataset details: [link](https://data.mendeley.com/datasets/tywbtsjrjv/1), License [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/), Quotation[[2]](#2) , Number of classes: 39, Number of images: 61 486
|
| 173 |
|
| 174 |
| Model | Format | Resolution | Top 1 Accuracy |
|
| 175 |
|-------|--------|------------|----------------|
|
|
|
|
| 190 |
### Accuracy with Food-101 dataset
|
| 191 |
|
| 192 |
|
| 193 |
+
Dataset details: [link](https://data.vision.ee.ethz.ch/cvl/datasets_extra/food-101/), Number of classes: 101 , Number of images: 101 000
|
| 194 |
|
| 195 |
| Model | Format | Resolution | Top 1 Accuracy |
|
| 196 |
|-------|--------|------------|----------------|
|
|
|
|
| 212 |
|
| 213 |
### Accuracy with ImageNet dataset
|
| 214 |
|
| 215 |
+
Dataset details: [link](https://www.image-net.org), Quotation[[4]](#4)
|
| 216 |
Number of classes: 1000.
|
| 217 |
To perform the quantization, we calibrated the activations with a random subset of the training set.
|
| 218 |
For the sake of simplicity, the accuracy reported here was estimated on the 50000 labelled images of the validation set.
|