Hugging Face's logo

STMicroelectronics
/
mobilenetv2

Image Classification
Model card Files
xet
Community
FBAGSTM commited on
Commit
d66dba3
·
verified ·
1 Parent(s): 4741174

Release AI-ModelZoo-4.0.0

Browse files
Files changed (1) hide show
  1. README.md +121 -124
README.md CHANGED
@@ -72,7 +72,7 @@ For an image resolution of NxM and P classes
72
 
73
  # Performances
74
 
75
- ## Metrics
76
 
77
  - Measures are done with default STM32Cube.AI configuration with enabled input / output allocated option.
78
  - `tfs` stands for "training from scratch", meaning that the model weights were randomly initialized before training.
@@ -80,71 +80,89 @@ For an image resolution of NxM and P classes
80
  - `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.
81
 
82
 
83
- ### Reference **NPU** memory footprint on food-101 and ImageNet dataset (see Accuracy for details on dataset)
84
- |Model | Dataset | Format | Resolution | Series | Internal RAM | External RAM | Weights Flash | STM32Cube.AI version | STEdgeAI Core version |
85
- |----------|------------------|--------|-------------|------------------|------------------|---------------------|---------------|----------------------|-------------------------|
86
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | food-101 | Int8 | 128x128x3 | STM32N6 | 240 | 0.0 | 680.92 | 10.2.0 | 2.2.0 |
87
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft_int8.tflite) | food-101 | Int8 | 224x224x3 | STM32N6 | 980 | 0.0 | 695.95 | 10.2.0 | 2.2.0 |
88
- | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_1.0_224_fft/mobilenet_v2_1.0_224_fft_int8.tflite) | food-101 | Int8 | 224x224x3 | STM32N6 | 2058 | 0.0 | 3070.61 | 10.2.0 | 2.2.0 |
89
- | [MobileNet v2 0.35 fft](ST_pretrainedmodel_public_dataset/person/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Person | Int8 | 128x128x3 | STM32N6 | 240 | 0.0 | 554.94 | 10.2.0 | 2.2.0 |
90
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_128/mobilenet_v2_0.35_128_int8.tflite) | ImageNet | Int8 | 128x128x3 | STM32N6 | 240 | 0.0 | 1806.61 | 10.2.0 | 2.2.0 |
91
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_224/mobilenet_v2_0.35_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32N6 | 980 | 0.0 | 1821.64 | 10.2.0 | 2.2.0 |
92
- | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32N6 | 2058 | 0.0 | 4196.3 | 10.2.0 | 2.2.0 |
93
- | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.4_224/mobilenet_v2_1.4_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32N6 | 2361 | 0.0 | 7285.86 | 10.2.0 | 2.2.0 |
94
-
95
-
96
- ### Reference **NPU** inference time on food-101 and ImageNet dataset (see Accuracy for details on dataset)
97
- | Model | Dataset | Format | Resolution | Board | Execution Engine | Inference time (ms) | Inf / sec | STM32Cube.AI version | STEdgeAI Core version |
98
- |--------|------------------|--------|-------------|------------------|------------------|---------------------|-----------|----------------------|-------------------------|
99
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | food-101 | Int8 | 128x128x3 | STM32N6570-DK | NPU/MCU | 3.34 | 299.4 | 10.2.0 | 2.2.0 |
100
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft_int8.tflite) | food-101 | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 6.03 | 165.83 | 10.2.0 | 2.2.0 |
101
- | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_1.0_224_fft/mobilenet_v2_1.0_224_fft_int8.tflite) | food-101 | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 17.31 | 57.77 | 10.2.0 | 2.2.0 |
102
- | [MobileNet v2 0.35 fft](ST_pretrainedmodel_public_dataset/person/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Person | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 2.95 | 338.98 | 10.2.0 | 2.2.0 |
103
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_128/mobilenet_v2_0.35_128_int8.tflite) | ImageNet | Int8 | 128x128x3 | STM32N6570-DK | NPU/MCU | 6.37 | 156.98 | 10.2.0 | 2.2.0 |
104
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_224/mobilenet_v2_0.35_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 9.06 | 110.37 | 10.2.0 | 2.2.0 |
105
- | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 20.3 | 49.26 | 10.2.0 | 2.2.0 |
106
- | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.4_224/mobilenet_v2_1.4_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 33.8 | 29.95 | 10.2.0 | 2.2.0 |
107
-
108
-
109
- ### Reference **MCU** memory footprint based on Flowers and ImageNet dataset (see Accuracy for details on dataset)
110
-
111
- | Model | Dataset | Format | Resolution | Series | Activation RAM | Runtime RAM | Weights Flash | Code Flash | Total RAM | Total Flash | STM32Cube.AI version |
112
- |--------------------------------------------------------------------------------------------------------------------------------------------------|----------|--------|-------------|---------|----------------|-------------|---------------|------------|-------------|-------------|----------------------|
113
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Flowers | Int8 | 128x128x3 | STM32H7 | 237.32 KiB | 30.15 KiB | 406.86 KiB | 107.4 KiB | 267.46 KiB | 514.26 KiB | 10.2.0 |
114
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft_int8.tflite) | Flowers | Int8 | 224x224x3 | STM32H7 | 832.64 KiB | 30.2 KiB | 406.86 KiB | 107.52 KiB | 862.84 KiB | 514.38 KiB | 10.2.0 |
115
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_128/mobilenet_v2_0.35_128_int8.tflite) | ImageNet | Int8 | 128x128x3 | STM32H7 | 237.32 KiB | 30.15 KiB | 1654.5 KiB KiB | 107.4 KiB | 267.47 KiB | 1762.79 KiB | 10.2.0 |
116
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_224/mobilenet_v2_0.35_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32H7 | 832.64 KiB | 30.2 KiB | 1654.5 KiB | 107.52 KiB | 862.84 KiB | 1762.9 KiB | 10.2.0 |
117
- | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32H7 | 1727.02 KiB | 30.2 KiB | 3458.97 KiB | 157.37 KiB | 1757.22 KiB | 3616.34 KiB | 10.2.0 |
118
- | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.4_224/mobilenet_v2_1.4_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32H7 | 2332.2 KiB | 30.2 KiB | 6015.34 KiB | 191.16 KiB | 2362.39 KiB | 6206.53 KiB | 10.2.0 |
119
-
120
- ### Reference **MCU** inference time based on Flowers and ImageNet dataset (see Accuracy for details on dataset)
121
-
122
- | Model | Dataset | Format | Resolution | Board | Execution Engine | Frequency | Inference time (ms) | STM32Cube.AI version |
123
- |--------------------------------------------------------------------------------------------------------------------------------------------------|----------|--------|-------------|------------------|------------------|-------------|---------------------|-----------------------|
124
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Flowers | Int8 | 128x128x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 96.52 ms | 10.2.0 |
125
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft_int8.tflite) | Flowers | Int8 | 224x224x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 297.38 ms | 10.2.0 |
126
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_128/mobilenet_v2_0.35_128_int8.tflite) | ImageNet | Int8 | 128x128x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 100.66 ms | 10.2.0 |
127
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_224/mobilenet_v2_0.35_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 301.58 ms | 10.2.0 |
128
- | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 1124.79 ms | 10.2.0 |
129
- | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.4_224/mobilenet_v2_1.4_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 2038.28 ms | 10.2.0 |
130
-
131
- ### Reference **MPU** inference time based on Flowers and ImageNet dataset (see Accuracy for details on dataset)
132
- | Model | Dataset | Format | Resolution | Quantization | Board | Execution Engine | Frequency | Inference time (ms) | %NPU | %GPU | %CPU | X-LINUX-AI version | Framework |
133
- |--------------------------------------------------------------------------------------------------------------------------------------------------|----------|--------|------------|---------------|-------------------|------------------|-----------|---------------------|-------|-------|------|--------------------|-----------------------|
134
- | [MobileNet v2 1.0_per_tensor](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8_per_tensor.tflite) | ImageNet | Int8 | 224x224x3 | per-tensor | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 12.39 ms | 81.42 | 18.58 |0 | v6.1.0 | OpenVX |
135
- | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8.tflite) | ImageNet | Int8 | 224x224x3 | per-channel ** | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 76.19 ms | 2.61 | 97.39 |0 | v6.1.0 | OpenVX |
136
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft_int8.tflite) | Flowers | Int8 | 224x224x3 | per-channel ** | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 25.5 ms | 4.31 | 95.69 |0 | v6.1.0 | OpenVX |
137
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Flowers | Int8 | 128x128x3 | per-channel ** | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 9.12 ms | 12.45 | 87.55 |0 | v6.1.0 | OpenVX |
138
- | [MobileNet v2 1.0_per_tensor](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8_per_tensor.tflite) | ImageNet | Int8 | 224x224x3 | per-tensor | STM32MP157F-DK2 | 2 CPU | 800 MHz | 331.69 ms | NA | NA |100 | v6.1.0 | TensorFlowLite 2.18.0 |
139
- | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8.tflite) | ImageNet | Int8 | 128x128x3 | per-channel | STM32MP157F-DK2 | 2 CPU | 800 MHz | 193.84 ms | NA | NA |100 | v6.1.0 | TensorFlowLite 2.18.0 |
140
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft_int8.tflite) | Flowers | Int8 | 224x224x3 | per-channel | STM32MP157F-DK2 | 2 CPU | 800 MHz | 54.15 ms | NA | NA |100 | v6.1.0 | TensorFlowLite 2.18.0 |
141
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Flowers | Int8 | 128x128x3 | per-channel | STM32MP157F-DK2 | 2 CPU | 800 MHz | 17.44 ms | NA | NA |100 | v6.1.0 | TensorFlowLite 2.18.0 |
142
- | [MobileNet v2 1.0_per_tensor](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8_per_tensor.tflite) | ImageNet | Int8 | 224x224x3 | per-tensor | STM32MP135F-DK2 | 1 CPU | 1000 MHz | 418.33 ms | NA | NA |100 | v6.1.0 | TensorFlowLite 2.18.0 |
143
- | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8.tflite) | ImageNet | Int8 | 128x128x3 | per-channel | STM32MP135F-DK2 | 1 CPU | 1000 MHz | 310.64 ms | NA | NA |100 | v6.1.0 | TensorFlowLite 2.18.0 |
144
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft_int8.tflite) | Flowers | Int8 | 224x224x3 | per-channel | STM32MP135F-DK2 | 1 CPU | 1000 MHz | 84.39 ms | NA | NA |100 | v6.1.0 | TensorFlowLite 2.18.0 |
145
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Flowers | Int8 | 128x128x3 | per-channel | STM32MP135F-DK2 | 1 CPU | 1000 MHz | 26.85 ms | NA | NA |100 | v6.1.0 | TensorFlowLite 2.18.0 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
146
 
147
  ** **To get the most out of MP25 NPU hardware acceleration, please use per-tensor quantization**
 
 
 
148
  ### Accuracy with Flowers dataset
149
 
150
 
@@ -152,18 +170,10 @@ Dataset details: [link](http://download.tensorflow.org/example_images/flower_pho
152
 
153
  | Model | Format | Resolution | Top 1 Accuracy |
154
  |-------|--------|------------|----------------|
155
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_tfs/mobilenet_v2_0.35_128_tfs.h5) | Float | 128x128x3 | 87.06 % |
156
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_tfs/mobilenet_v2_0.35_128_tfs_int8.tflite) | Int8 | 128x128x3 | 87.47 % |
157
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_tl/mobilenet_v2_0.35_128_tl.h5) | Float | 128x128x3 | 88.15 % |
158
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_tl/mobilenet_v2_0.35_128_tl_int8.tflite) | Int8 | 128x128x3 | 88.01 % |
159
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft.h5) | Float | 128x128x3 | 91.83 % |
160
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Int8 | 128x128x3 | 91.01 % |
161
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_tfs/mobilenet_v2_0.35_224_tfs.h5) | Float | 224x224x3 | 88.69 % |
162
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_tfs/mobilenet_v2_0.35_224_tfs_int8.tflite) | Int8 | 224x224x3 | 88.83 % |
163
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_tl/mobilenet_v2_0.35_224_tl.h5) | Float | 224x224x3 | 88.96 % |
164
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_tl/mobilenet_v2_0.35_224_tl_int8.tflite) | Int8 | 224x224x3 | 88.01 % |
165
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft.h5) | Float | 224x224x3 | 93.6 % |
166
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/flowers/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft_int8.tflite) | Int8 | 224x224x3 | 92.78 % |
167
 
168
 
169
  ### Accuracy with Plant-village dataset
@@ -173,18 +183,10 @@ Dataset details: [link](https://data.mendeley.com/datasets/tywbtsjrjv/1) , Licen
173
 
174
  | Model | Format | Resolution | Top 1 Accuracy |
175
  |-------|--------|------------|----------------|
176
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_128_tfs/mobilenet_v2_0.35_128_tfs.h5) | Float | 128x128x3 | 99.86 % |
177
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_128_tfs/mobilenet_v2_0.35_128_tfs_int8.tflite) | Int8 | 128x128x3 | 99.83 % |
178
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_128_tl/mobilenet_v2_0.35_128_tl.h5) | Float | 128x128x3 | 93.51 % |
179
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_128_tl/mobilenet_v2_0.35_128_tl_int8.tflite) | Int8 | 128x128x3 | 92.33 % |
180
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft.h5) | Float | 128x128x3 | 99.77 % |
181
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Int8 | 128x128x3 | 99.48 % |
182
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_224_tfs/mobilenet_v2_0.35_224_tfs.h5) | Float | 224x224x3 | 99.86 % |
183
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_224_tfs/mobilenet_v2_0.35_224_tfs_int8.tflite) | Int8 | 224x224x3 | 99.81 % |
184
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_224_tl/mobilenet_v2_0.35_224_tl.h5) | Float | 224x224x3 | 93.62 % |
185
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_224_tl/mobilenet_v2_0.35_224_tl_int8.tflite) | Int8 | 224x224x3 | 92.8 % |
186
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft.h5) | Float | 224x224x3 | 99.95 % |
187
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant-village/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft_int8.tflite) | Int8 | 224x224x3 | 99.68 % |
188
 
189
 
190
  ### Accuracy with Food-101 dataset
@@ -193,38 +195,29 @@ Dataset details: [link](https://data.vision.ee.ethz.ch/cvl/datasets_extra/food-1
193
 
194
  | Model | Format | Resolution | Top 1 Accuracy |
195
  |-------|--------|------------|----------------|
196
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_128_tfs/mobilenet_v2_0.35_128_tfs.h5) | Float | 128x128x3 | 64.22 % |
197
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_128_tfs/mobilenet_v2_0.35_128_tfs_int8.tflite) | Int8 | 128x128x3 | 63.41 % |
198
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_128_tl/mobilenet_v2_0.35_128_tl.h5) | Float | 128x128x3 | 44.74 % |
199
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_128_tl/mobilenet_v2_0.35_128_tl_int8.tflite) | Int8 | 128x128x3 | 42.01 % |
200
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft.h5) | Float | 128x128x3 | 64.21 % |
201
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Int8 | 128x128x3 | 63.41 % |
202
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_224_tfs/mobilenet_v2_0.35_224_tfs.h5) | Float | 224x224x3 | 72.31 % |
203
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_224_tfs/mobilenet_v2_0.35_224_tfs_int8.tflite) | Int8 | 224x224x3 | 72.05 % |
204
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_224_tl/mobilenet_v2_0.35_224_tl.h5) | Float | 224x224x3 | 49.01 % |
205
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_224_tl/mobilenet_v2_0.35_224_tl_int8.tflite) | Int8 | 224x224x3 | 47.26 % |
206
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft.h5) | Float | 224x224x3 | 73.74 % |
207
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_0.35_224_fft/mobilenet_v2_0.35_224_fft_int8.tflite) | Int8 | 224x224x3 | 73.16 % |
208
- | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_1.0_224_fft/mobilenet_v2_1.0_224_fft.h5) | Float | 224x224x3 | 77.77 % |
209
- | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food-101/mobilenet_v2_1.0_224_fft/mobilenet_v2_1.0_224_fft_int8.tflite) | Int8 | 224x224x3 | 77.09 % |
210
-
211
-
212
- ### Accuracy with person dataset
213
-
214
- The person dataset is derived from COCO-2014 and created using the script here (link). The dataset folder has 2 sub-folders — person and notperson containing images of the respective types
215
  Dataset details: [link](https://cocodataset.org/) , License [Creative Commons Attribution 4.0](https://creativecommons.org/licenses/by/4.0/legalcode), Quotation[[3]](#3) , Number of classes: 2 , Number of images: 84810
216
 
217
  | Model | Format | Resolution | Top 1 Accuracy |
218
  |------------|--------|-----------|----------------|
219
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/person/mobilenet_v2_0.35_128_tfs/mobilenet_v2_0.35_128_tfs.h5) | Float | 128x128x3 | 92.56 % |
220
- | [MobileNet v2 0.35 tfs](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/person/mobilenet_v2_0.35_128_tfs/mobilenet_v2_0.35_128_tfs_int8.tflite) | Int8 | 128x128x3 | 92.44 % |
221
- | [MobileNet v2 0.35 tl ](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/person/mobilenet_v2_0.35_128_tl/mobilenet_v2_0.35_128_tl.h5) | Float | 128x128x3 | 92.28 % |
222
- | [MobileNet v2 0.35 tl](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/person/mobilenet_v2_0.35_128_tl/mobilenet_v2_0.35_128_tl_int8.tflite) | Int8 | 128x128x3 | 91.63 % |
223
- | [MobileNet v2 0.35 fft ](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/person/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft.h5) | Float | 128x128x3 | 95.37 % |
224
- | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/person/mobilenet_v2_0.35_128_fft/mobilenet_v2_0.35_128_fft_int8.tflite) | Int8 | 128x128x3 | 94.95 % |
225
 
226
 
227
- ### Accuracy with ImageNet
228
 
229
  Dataset details: [link](https://www.image-net.org), Quotation[[4]](#4).
230
  Number of classes: 1000.
@@ -233,15 +226,19 @@ For the sake of simplicity, the accuracy reported here was estimated on the 5000
233
 
234
  | Model | Format | Resolution | Top 1 Accuracy |
235
  |----------|--------|------------|----------------|
236
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_128/mobilenet_v2_0.35_128.h5) | Float | 128x128x3 | 46.96 % |
237
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_128/mobilenet_v2_0.35_128_int8.tflite) | Int8 | 128x128x3 | 43.94 % |
238
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_224/mobilenet_v2_0.35_224.h5) | Float | 224x224x3 | 56.44 % |
239
- | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_0.35_224/mobilenet_v2_0.35_224_int8.tflite) | Int8 | 224x224x3 | 54.7 % |
240
- | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224.h5) | Float | 224x224x3 | 68.87 % |
241
- | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8.tflite) | Int8 | 224x224x3 | 67.97 % |
242
- | [MobileNet v2 1.0_per_tensor](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.0_224/mobilenet_v2_1.0_224_int8_per_tensor.tflite) | Int8 | 224x224x3 | 64.53 % |
243
- | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.4_224/mobilenet_v2_1.4_224.h5) | Float | 224x224x3 | 71.97 % |
244
- | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/ImageNet/mobilenet_v2_1.4_224/mobilenet_v2_1.4_224_int8.tflite) | Int8 | 224x224x3 | 71.46 % |
 
 
 
 
245
 
246
 
247
  ## Retraining and Integration in a simple example:
@@ -262,4 +259,4 @@ L. Bossard, M. Guillaumin, and L. Van Gool, "Food-101 -- Mining Discriminative C
262
 
263
  <a id="4">[4]</a>
264
  Olga Russakovsky*, Jia Deng*, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, Alexander C. Berg and Li Fei-Fei.
265
- (* = equal contribution) ImageNet Large Scale Visual Recognition Challenge.
 
72
 
73
  # Performances
74
 
75
+ ## Metricss
76
 
77
  - Measures are done with default STM32Cube.AI configuration with enabled input / output allocated option.
78
  - `tfs` stands for "training from scratch", meaning that the model weights were randomly initialized before training.
 
80
  - `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.
81
 
82
 
83
+ ### Reference **NPU** memory footprint on food101 and imagenet dataset (see Accuracy for details on dataset)
84
+ |Model | Dataset | Format | Resolution | Series | Internal RAM (KiB) | External RAM (KiB)| Weights Flash (KiB)| STEdgeAI Core version |
85
+ |----------|------------------|--------|-------------|------------------|------------------|---------------------|---------------|-------------------------|
86
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | food101 | Int8 | 128x128x3 | STM32N6 | 240 | 0.0 | 530.59 | 3.0.0 |
87
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_qdq_w4_53.32%_w8_46.68%_a8_100%_acc_64.61.onnx) | food101 | Int8/Int4 | 128x128x3 | STM32N6 | 240 | 0.0 | 396.44 | 3.0.0 |
88
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_int8.tflite) | food101 | Int8 | 224x224x3 | STM32N6 | 931 | 0.0 | 557.44 | 3.0.0 |
89
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_qdq_w4_53.32%_w8_46.68%_a8_100%_acc_74.86.onnx) | food101 | Int8/Int4 | 224x224x3 | STM32N6 | 1127 | 0.0 | 423.28 | 3.0.0 |
90
+ | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a100_224_fft/mobilenetv2_a100_224_fft_int8.tflite) | food101 | Int8 | 224x224x3 | STM32N6 | 2058 | 0.0 | 2686.42 | 3.0.0 |
91
+ | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a100_224_fft/mobilenetv2_a100_224_fft_qdq_w4_30.91%_w8_69.09%_a8_100%_acc_80.06.onnx) | food101 | Int8/Int4 | 224x224x3 | STM32N6 | 2058 | 0.0 | 2336.39 | 3.0.0 |
92
+ | [MobileNet v2 0.35 fft](ST_pretrainedmodel_public_dataset/coco_person/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Person | Int8 | 128x128x3 | STM32N6 | 240 | 0.0 | 404.55 | 3.0.0 |
93
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_128/mobilenetv2_a035_128_int8.tflite) | imagenet | Int8 | 128x128x3 | STM32N6 | 240 | 0.0 | 1656.28 | 3.0.0 |
94
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_128/mobilenetv2_a035_128_qdq_w4_85.64%_w8_14.36%_a8_100%_acc_43.53.onnx) | imagenet | Int8/Int4 | 128x128x3 | STM32N6 | 240 | 0.0 | 962.22 | 3.0.0 |
95
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_224/mobilenetv2_a035_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32N6 | 931 | 0.0 | 1683.13 | 3.0.0 |
96
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_224/mobilenetv2_a035_224_qdq_w4_85.64%_w8_14.36%_a8_100%_acc_56.25.onnx) | imagenet | Int8/Int4 | 224x224x3 | STM32N6 | 1127 | 0.0 | 989.06 | 3.0.0 |
97
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32N6 | 2058 | 0.0 | 3812.11 | 3.0.0 |
98
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_qdq_w4_48.7%_w8_51.3%_a8_100%_acc_69.54.onnx) | imagenet | Int8/Int4 | 224x224x3 | STM32N6 | 2058 | 0.0 | 2988.05 | 3.0.0 |
99
+ | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a140_224/mobilenetv2_a140_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32N6 | 2361 | 0.0 | 6746.7 | 3.0.0 |
100
+ | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a140_224/mobilenetv2_a140_224_qdq_w4_42.82%_w8_57.18%_a8_100%_acc_73.12.onnx) | imagenet | Int8/Int4 | 224x224x3 | STM32N6 | 2361 | 0.0 | 5480.25 | 3.0.0 |
101
+
102
+
103
+ ### Reference **NPU** inference time on food101 and imagenet dataset (see Accuracy for details on dataset)
104
+ | Model | Dataset | Format | Resolution | Board | Execution Engine | Inference time (ms) | Inf / sec | STEdgeAI Core version |
105
+ |--------|------------------|--------|-------------|------------------|------------------|---------------------|-----------|-------------------------|
106
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | food101 | Int8 | 128x128x3 | STM32N6570-DK | NPU/MCU | 2.82 | 354.6 | 3.0.0 |
107
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_qdq_w4_53.32%_w8_46.68%_a8_100%_acc_64.61.onnx) | food101 | Int8/Int4 | 128x128x3 | STM32N6570-DK | NPU/MCU | 2.65 | 377.36 | 3.0.0 |
108
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_int8.tflite) | food101 | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 5.67 | 176.36 | 3.0.0 |
109
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_qdq_w4_53.32%_w8_46.68%_a8_100%_acc_74.86.onnx) | food101 | Int8/Int4 | 224x224x3 | STM32N6570-DK | NPU/MCU | 5.43 | 184.16 | 3.0.0 |
110
+ | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a100_224_fft/mobilenetv2_a100_224_fft_int8.tflite) | food101 | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 17.44 | 57.34 | 3.0.0 |
111
+ | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a100_224_fft/mobilenetv2_a100_224_fft_qdq_w4_30.91%_w8_69.09%_a8_100%_acc_80.06.onnx) | food101 | Int8/Int4 | 224x224x3 | STM32N6570-DK | NPU/MCU | 16.43 | 60.86 | 3.0.0 |
112
+ | [MobileNet v2 0.35 fft](ST_pretrainedmodel_public_dataset/coco_person/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Person | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 2.47 | 404.86 | 3.0.0 |
113
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_128/mobilenetv2_a035_128_int8.tflite) | imagenet | Int8 | 128x128x3 | STM32N6570-DK | NPU/MCU | 5.83 | 171.53 | 3.0.0 |
114
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_128/mobilenetv2_a035_128_qdq_w4_85.64%_w8_14.36%_a8_100%_acc_43.53.onnx) | imagenet | Int8/Int4 | 128x128x3 | STM32N6570-DK | NPU/MCU | 4.05 | 246.91 | 3.0.0 |
115
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_224/mobilenetv2_a035_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 8.68 | 115.2 | 3.0.0 |
116
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_224/mobilenetv2_a035_224_qdq_w4_85.64%_w8_14.36%_a8_100%_acc_56.25.onnx) | imagenet | Int8/Int4 | 224x224x3 | STM32N6570-DK | NPU/MCU | 6.83 | 146.4 | 3.0.0 |
117
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 20.45 | 48.9 | 3.0.0 |
118
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_qdq_w4_48.7%_w8_51.3%_a8_100%_acc_69.54.onnx) | imagenet | Int8/Int4 | 224x224x3 | STM32N6570-DK | NPU/MCU | 18.21 | 54.91 | 3.0.0 |
119
+ | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a140_224/mobilenetv2_a140_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32N6570-DK | NPU/MCU | 34.74 | 28.79 | 3.0.0 |
120
+ | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a140_224/mobilenetv2_a140_224_qdq_w4_42.82%_w8_57.18%_a8_100%_acc_73.12.onnx) | imagenet | Int8/Int4 | 224x224x3 | STM32N6570-DK | NPU/MCU | 31.94 | 31.3 | 3.0.0 |
121
+
122
+
123
+ ### Reference **MCU** memory footprint based on Flowers and imagenet dataset (see Accuracy for details on dataset)
124
+
125
+ | Model | Dataset | Format | Resolution | Series | Activation RAM | Runtime RAM | Weights Flash | Code Flash | Total RAM | Total Flash | STEdgeAI Core version |
126
+ |--------|----------|--------|-------------|---------|-------------|---------------|------------|-------------|-------------|----------------|-----------------------|
127
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Flowers | Int8 | 128x128x3 | STM32H7 | 237.32 KiB | 3.77 KiB | 406.86 KiB | 64.3 KiB | 241.09 KiB | 471.16 KiB | 3.0.0 |
128
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_int8.tflite) | Flowers | Int8 | 224x224x3 | STM32H7 | 699.32 KiB | 3.77 KiB | 406.86 KiB | 64.69 KiB | 703.09 KiB | 471.55 KiB | 3.0.0 |
129
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_128/mobilenetv2_a035_128_int8.tflite) | imagenet | Int8 | 128x128x3 | STM32H7 | 237.32 KiB | 3.36 KiB | 1654.5 KiB KiB | 65.25 KiB | 240.68 KiB | 1719.75 KiB | 3.0.0 |
130
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_224/mobilenetv2_a035_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32H7 | 699.32 KiB | 3.36 KiB | 1654.5 KiB | 65.68 KiB | 702.68 KiB | 1720.18 KiB | 3.0.0 |
131
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32H7 | 1433.13 KiB | 3.36 KiB | 3458.97 KiB | 104.92 KiB | 1436.49 KiB | 3563.89 KiB | 3.0.0 |
132
+ | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a140_224/mobilenetv2_a140_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32H7 | 2143.27 KiB | 3.36 KiB | 6015.34 KiB | 132.17 KiB | 2146.63 KiB | 6147.51 KiB | 3.0.0 |
133
+
134
+ ### Reference **MCU** inference time based on Flowers and imagenet dataset (see Accuracy for details on dataset)
135
+
136
+ | Model | Dataset | Format | Resolution | Board | Execution Engine | Frequency | Inference time (ms) | STEdgeAI Core version |
137
+ |---------------------------------|----------|--------|-------------|------------------|------------------|-------------|---------------------|-----------------------|
138
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Flowers | Int8 | 128x128x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 100.09 ms | 3.0.0 |
139
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_int8.tflite) | Flowers | Int8 | 224x224x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 308.57 ms | 3.0.0 |
140
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_128/mobilenetv2_a035_128_int8.tflite) | imagenet | Int8 | 128x128x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 113.43 ms | 3.0.0 |
141
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_224/mobilenetv2_a035_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 321.76 ms | 3.0.0 |
142
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 1118.27 ms | 3.0.0 |
143
+ | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a140_224/mobilenetv2_a140_224_int8.tflite) | imagenet | Int8 | 224x224x3 | STM32H747I-DISCO | 1 CPU | 400 MHz | 2035.56 ms | 3.0.0 |
144
+
145
+ ### Reference **MPU** inference time based on Flowers and imagenet dataset (see Accuracy for details on dataset)
146
+
147
+ | Model | Dataset | Format | Resolution | Quantization | Board | Execution Engine | Frequency | Inference time (ms) | %NPU | %GPU | %CPU | X-LINUX-AI version | Framework |
148
+ |--------------------------------------------------------------------------------------------------------------------------------------------------|----------|--------|------------|----------------|-------------------|------------------|-----------|---------------------|------|-------|------|--------------------|-----------------------|
149
+ | [MobileNet v2 1.0_per_tensor](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8_per_tensor.tflite) | imagenet | Int8 | 224x224x3 | per-tensor | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 12.15 | 81.71| 18.29 | 0 | v6.1.0 | OpenVX |
150
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8.tflite) | imagenet | Int8 | 224x224x3 | per-channel ** | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 75.91 | 2.77 | 97.23 | 0 | v6.1.0 | OpenVX |
151
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_int8.tflite) | Flowers | Int8 | 224x224x3 | per-channel ** | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 25.30 | 3.89 | 96.11 | 0 | v6.1.0 | OpenVX |
152
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Flowers | Int8 | 128x128x3 | per-channel ** | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 8.97 | 11.73| 88.27 | 0 | v6.1.0 | OpenVX |
153
+ | [MobileNet v2 1.0_per_tensor](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8_per_tensor.tflite) | imagenet | Int8 | 224x224x3 | per-tensor | STM32MP157F-DK2 | 2 CPU | 800 MHz | 346.87 | NA | NA | 100 | v6.1.0 | TensorFlowLite 2.18.0 |
154
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8.tflite) | imagenet | Int8 | 224x224x3 | per-channel | STM32MP157F-DK2 | 2 CPU | 800 MHz | 206.64 | NA | NA | 100 | v6.1.0 | TensorFlowLite 2.18.0 |
155
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_int8.tflite) | Flowers | Int8 | 224x224x3 | per-channel | STM32MP157F-DK2 | 2 CPU | 800 MHz | 51.33 | NA | NA | 100 | v6.1.0 | TensorFlowLite 2.18.0 |
156
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Flowers | Int8 | 128x128x3 | per-channel | STM32MP157F-DK2 | 2 CPU | 800 MHz | 16.27 | NA | NA | 100 | v6.1.0 | TensorFlowLite 2.18.0 |
157
+ | [MobileNet v2 1.0_per_tensor](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8_per_tensor.tflite) | imagenet | Int8 | 224x224x3 | per-tensor | STM32MP135F-DK2 | 1 CPU | 1000 MHz | 434.12 | NA | NA | 100 | v6.1.0 | TensorFlowLite 2.18.0 |
158
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8.tflite) | imagenet | Int8 | 224x224x3 | per-channel | STM32MP135F-DK2 | 1 CPU | 1000 MHz | 316.76 | NA | NA | 100 | v6.1.0 | TensorFlowLite 2.18.0 |
159
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_int8.tflite) | Flowers | Int8 | 224x224x3 | per-channel | STM32MP135F-DK2 | 1 CPU | 1000 MHz | 81.91 | NA | NA | 100 | v6.1.0 | TensorFlowLite 2.18.0 |
160
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Flowers | Int8 | 128x128x3 | per-channel | STM32MP135F-DK2 | 1 CPU | 1000 MHz | 25.75 | NA | NA | 100 | v6.1.0 | TensorFlowLite 2.18.0 |
161
 
162
  ** **To get the most out of MP25 NPU hardware acceleration, please use per-tensor quantization**
163
+
164
+ ** **Note:** On STM32MP2 devices, per-channel quantized models are internally converted to per-tensor quantization by the compiler using an entropy-based method. This may introduce a slight loss in accuracy compared to the original per-channel models.
165
+
166
  ### Accuracy with Flowers dataset
167
 
168
 
 
170
 
171
  | Model | Format | Resolution | Top 1 Accuracy |
172
  |-------|--------|------------|----------------|
173
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft.keras) | Float | 128x128x3 | 91.83 % |
174
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Int8 | 128x128x3 | 91.01 % |
175
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft.keras) | Float | 224x224x3 | 93.6 % |
176
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/tf_flowers/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_int8.tflite) | Int8 | 224x224x3 | 92.78 % |
 
 
 
 
 
 
 
 
177
 
178
 
179
  ### Accuracy with Plant-village dataset
 
183
 
184
  | Model | Format | Resolution | Top 1 Accuracy |
185
  |-------|--------|------------|----------------|
186
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant_leaf_diseases/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft.keras) | Float | 128x128x3 | 99.77 % |
187
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant_leaf_diseases/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Int8 | 128x128x3 | 99.48 % |
188
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant_leaf_diseases/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft.keras) | Float | 224x224x3 | 99.95 % |
189
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/plant_leaf_diseases/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_int8.tflite) | Int8 | 224x224x3 | 99.68 % |
 
 
 
 
 
 
 
 
190
 
191
 
192
  ### Accuracy with Food-101 dataset
 
195
 
196
  | Model | Format | Resolution | Top 1 Accuracy |
197
  |-------|--------|------------|----------------|
198
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft.keras) | Float | 128x128x3 | 65.88 % |
199
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Int8 | 128x128x3 | 65 % |
200
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_qdq_w4_53.32%_w8_46.68%_a8_100%_acc_64.61.onnx) | Int8/Int4 | 128x128x3 | 64.61 % |
201
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft.keras) | Float | 224x224x3 | 76.47 % |
202
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_int8.tflite) | Int8 | 224x224x3 | 75.4 % |
203
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a035_224_fft/mobilenetv2_a035_224_fft_qdq_w4_53.32%_w8_46.68%_a8_100%_acc_74.86.onnx) | Int8/Int4 | 224x224x3 | 74.86 % |
204
+ | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a100_224_fft/mobilenetv2_a100_224_fft.keras) | Float | 224x224x3 | 82.13 % |
205
+ | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a100_224_fft/mobilenetv2_a100_224_fft_int8.tflite) | Int8 | 224x224x3 | 81.6 % |
206
+ | [MobileNet v2 1.0 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a100_224_fft/mobilenetv2_a100_224_fft_qdq_w4_30.91%_w8_69.09%_a8_100%_acc_80.06.onnx) | Int8/Int4 | 224x224x3 | 80.06 % |
207
+
208
+
209
+ ### Accuracy with coco_person dataset
210
+
211
+ The coco_person dataset is derived from COCO-2014 and created using the script here (link). The dataset folder has 2 sub-folders — person and not person containing images of the respective types
 
 
 
 
 
212
  Dataset details: [link](https://cocodataset.org/) , License [Creative Commons Attribution 4.0](https://creativecommons.org/licenses/by/4.0/legalcode), Quotation[[3]](#3) , Number of classes: 2 , Number of images: 84810
213
 
214
  | Model | Format | Resolution | Top 1 Accuracy |
215
  |------------|--------|-----------|----------------|
216
+ | [MobileNet v2 0.35 fft ](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/coco_person/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft.keras) | Float | 128x128x3 | 95.37 % |
217
+ | [MobileNet v2 0.35 fft](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/coco_person/mobilenetv2_a035_128_fft/mobilenetv2_a035_128_fft_int8.tflite) | Int8 | 128x128x3 | 94.95 % |
 
 
 
 
218
 
219
 
220
+ ### Accuracy with imagenet
221
 
222
  Dataset details: [link](https://www.image-net.org), Quotation[[4]](#4).
223
  Number of classes: 1000.
 
226
 
227
  | Model | Format | Resolution | Top 1 Accuracy |
228
  |----------|--------|------------|----------------|
229
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_128/mobilenetv2_a035_128.keras) | Float | 128x128x3 | 46.96 % |
230
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_128/mobilenetv2_a035_128_int8.tflite) | Int8 | 128x128x3 | 43.94 % |
231
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_128/mobilenetv2_a035_128_qdq_w4_85.64%_w8_14.36%_a8_100%_acc_43.53.onnx) | Int8/Int4 | 128x128x3 | 43.53 % |
232
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_224/mobilenetv2_a035_224.keras) | Float | 224x224x3 | 58.13 % |
233
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_224/mobilenetv2_a035_224_int8.tflite) | Int8 | 224x224x3 | 56.77 % |
234
+ | [MobileNet v2 0.35](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a035_224/mobilenetv2_a035_224_qdq_w4_85.64%_w8_14.36%_a8_100%_acc_56.25.onnx) | Int8/Int4 | 224x224x3 | 56.25 % |
235
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224.keras) | Float | 224x224x3 | 70.37 % |
236
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8.tflite) | Int8 | 224x224x3 | 69.75 % |
237
+ | [MobileNet v2 1.0](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_qdq_w4_48.7%_w8_51.3%_a8_100%_acc_69.54.onnx) | Int8/Int4 | 224x224x3 | 69.54 % |
238
+ | [MobileNet v2 1.0_per_tensor](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a100_224/mobilenetv2_a100_224_int8_per_tensor.tflite) | Int8 | 224x224x3 | 65.84 % |
239
+ | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a140_224/mobilenetv2_a140_224.keras) | Float | 224x224x3 | 73.74 % |
240
+ | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a140_224/mobilenetv2_a140_224_int8.tflite) | Int8 | 224x224x3 | 73.45 % |
241
+ | [MobileNet v2 1.4](https://github.com/STMicroelectronics/stm32ai-modelzoo/tree/main/image_classification/mobilenetv2/Public_pretrainedmodel_public_dataset/imagenet/mobilenetv2_a140_224/mobilenetv2_a140_224_qdq_w4_42.82%_w8_57.18%_a8_100%_acc_73.12.onnx) | Int8/Int4 | 224x224x3 | 73.12 % |
242
 
243
 
244
  ## Retraining and Integration in a simple example:
 
259
 
260
  <a id="4">[4]</a>
261
  Olga Russakovsky*, Jia Deng*, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, Alexander C. Berg and Li Fei-Fei.
262
+ (* = equal contribution) imagenet Large Scale Visual Recognition Challenge.