Hugging Face's logo

STMicroelectronics
/
yolov8n

Object Detection
Model card Files
xet
Community
FBAGSTM commited on
Commit
4038fcc
·
verified ·
1 Parent(s): 5b65f73

Update ST Model Zoo

Browse files
Files changed (1) hide show
  1. README.md +27 -34
README.md CHANGED
@@ -1,10 +1,3 @@
1
- ---
2
- license: other
3
- license_name: sla0044
4
- license_link: >-
5
- https://github.com/STMicroelectronics/stm32aimodelzoo/object_detection/yolov8n/LICENSE.md
6
- pipeline_tag: object-detection
7
- ---
8
  # Yolov8n object detection quantized
9
 
10
  ## **Use case** : `Object detection`
@@ -56,32 +49,33 @@ With an image resolution of NxM and K classes to detect:
56
 
57
  ## Metrics
58
 
59
- Measures are done with default STM32Cube.AI configuration with enabled input / output allocated option.
 
 
 
 
 
60
 
61
 
62
  ### Reference **NPU** memory footprint based on COCO Person dataset (see Accuracy for details on dataset)
63
- |Model | Dataset | Format | Resolution | Series | Internal RAM | External RAM | Weights Flash | STM32Cube.AI version | STEdgeAI Core version |
64
- |----------|------------------|--------|-------------|------------------|------------------|---------------------|-------|----------------------|-------------------------|
65
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_192_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 192x192x3 | STM32N6 | 697.5 | 0.0 | 2965.61 | 10.0.0 | 2.0.0 |
66
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_256_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 256x256x3 | STM32N6 | 1626 | 0.0 | 2970.13 | 10.0.0 | 2.0.0 |
67
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_320_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 320x320x3 | STM32N6 | 2162.5 | 0.0 | 2975.99 | 10.0.0 | 2.0.0 |
68
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_416_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 416x416x3 | STM32N6 | 2704 | 0.0 | 2987.52 | 10.0.0 | 2.0.0 |
69
-
70
-
71
  ### Reference **NPU** inference time based on COCO Person dataset (see Accuracy for details on dataset)
72
- | Model | Dataset | Format | Resolution | Board | Execution Engine | Inference time (ms) | Inf / sec | STM32Cube.AI version | STEdgeAI Core version |
73
- |--------|------------------|--------|-------------|------------------|------------------|---------------------|-------|----------------------|-------------------------|
74
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_192_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 192x192x3 | STM32N6570-DK | NPU/MCU | 18.91 | 52.89 | 10.0.0 | 2.0.0 |
75
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_256_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 256x256x3 | STM32N6570-DK | NPU/MCU | 28.6 | 34.97 | 10.0.0 | 2.0.0 |
76
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_320_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 320x320x3 | STM32N6570-DK | NPU/MCU | 38.32 | 26.09 | 10.0.0 | 2.0.0 |
77
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_416_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 416x416x3 | STM32N6570-DK | NPU/MCU | 63.03 | 15.86 | 10.0.0 | 2.0.0 |
78
-
79
-
80
  ### Reference **MPU** inference time based on COCO Person dataset (see Accuracy for details on dataset)
81
  Model | Format | Resolution | Quantization | Board | Execution Engine | Frequency | Inference time (ms) | %NPU | %GPU | %CPU | X-LINUX-AI version | Framework |
82
  |-----------|--------|------------|---------------|-------------------|------------------|-----------|---------------------|-------|-------|------|--------------------|-----------------------|
83
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_256_quant_pc_uf_pose_coco-st.tflite) | Int8 | 256x256x3 | per-channel** | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 102.8 ms | 11.70 | 88.30 |0 | v5.0.0 | OpenVX |
84
- | [YOLOv8n per tensor](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_256_quant_pt_uf_pose_coco-st.tflite) | Int8 | 256x256x3 | per-tensor | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 17.57 ms | 86.79 | 13.21 |0 | v5.0.0 | OpenVX |
85
 
86
  ** **To get the most out of MP25 NPU hardware acceleration, please use per-tensor quantization**
87
 
@@ -93,19 +87,19 @@ Dataset details: [link](https://cocodataset.org/#download) , License [CC BY 4.0]
93
 
94
  | Model | Format | Resolution | AP* |
95
  |-------|--------|------------|----------------|
96
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_192_quant_pc_uf_od_coco-person.tflite) | Int8 | 192x192x3 | 56.90 % |
97
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_256_quant_pc_uf_od_coco-person.tflite) | Int8 | 256x256x3 | 62.60 % |
98
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_320_quant_pc_uf_od_coco-person.tflite) | Int8 | 320x320x3 | 66.20 % |
99
- | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_416_quant_pc_uf_od_coco-person.tflite) | Int8 | 416x416x3 | 68.90 % |
100
 
101
- \* EVAL_IOU = 0.4, NMS_THRESH = 0.5, SCORE_THRESH =0.001
102
 
103
  ## Integration in a simple example and other services support:
104
 
105
  Please refer to the stm32ai-modelzoo-services GitHub [here](https://github.com/STMicroelectronics/stm32ai-modelzoo-services).
106
  The models are stored in the Ultralytics repository. You can find them at the following link: [Ultralytics YOLOv8-STEdgeAI Models](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/).
107
 
108
- Please refer to the [Ultralytics documentation](https://docs.ultralytics.com/tasks/pose/#train) to retrain the models.
109
 
110
  # References
111
 
@@ -132,5 +126,4 @@ Please refer to the [Ultralytics documentation](https://docs.ultralytics.com/tas
132
  timestamp = {Mon, 13 Aug 2018 16:48:13 +0200},
133
  biburl = {https://dblp.org/rec/bib/journals/corr/LinMBHPRDZ14},
134
  bibsource = {dblp computer science bibliography, https://dblp.org}
135
- }
136
-
 
 
 
 
 
 
 
 
1
  # Yolov8n object detection quantized
2
 
3
  ## **Use case** : `Object detection`
 
49
 
50
  ## Metrics
51
 
52
+ Measures are done with default STM32Cube.AI configuration with enabled input / output allocated option.
53
+ > [!CAUTION]
54
+ > All YOLOv8 hyperlinks in the tables below link to an external GitHub folder, which is subject to its own license terms:
55
+ https://github.com/stm32-hotspot/ultralytics/blob/main/LICENSE
56
+ Please also check the folder's README.md file for detailed information about its use and content:
57
+ https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/README.md
58
 
59
 
60
  ### Reference **NPU** memory footprint based on COCO Person dataset (see Accuracy for details on dataset)
61
+ | Model | Dataset | Format | Resolution | Series | Internal RAM | External RAM | Weights Flash | STM32Cube.AI version | STEdgeAI Core version |
62
+ |---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-------------|----------|--------------|----------|----------------|----------------|-----------------|------------------------|-------------------------|
63
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_192_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 192x192x3 | STM32N6 | 261 | 0 | 2936.52 | 10.2.0 | 2.2.0 |
64
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_256_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 256x256x3 | STM32N6 | 624 | 0 | 2941.09 | 10.2.0 | 2.2.0 |
65
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_320_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 320x320x3 | STM32N6 | 839.06 | 0 | 2947.02 | 10.2.0 | 2.2.0 |
66
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_416_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 416x416x3 | STM32N6 | 2242.84 | 0 | 2958.34 | 10.2.0 | 2.2.0 |
 
 
67
  ### Reference **NPU** inference time based on COCO Person dataset (see Accuracy for details on dataset)
68
+ | Model | Dataset | Format | Resolution | Board | Execution Engine | Inference time (ms) | Inf / sec | STM32Cube.AI version | STEdgeAI Core version |
69
+ |---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-------------|----------|--------------|---------------|--------------------|-----------------------|-------------|------------------------|-------------------------|
70
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_192_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 192x192x3 | STM32N6570-DK | NPU/MCU | 16.88 | 59.24 | 10.2.0 | 2.2.0 |
71
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_256_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 256x256x3 | STM32N6570-DK | NPU/MCU | 24.94 | 40.1 | 10.2.0 | 2.2.0 |
72
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_320_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 320x320x3 | STM32N6570-DK | NPU/MCU | 31.75 | 31.5 | 10.2.0 | 2.2.0 |
73
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_416_quant_pc_uf_od_coco-person.tflite) | COCO-Person | Int8 | 416x416x3 | STM32N6570-DK | NPU/MCU | 53.11 | 18.83 | 10.2.0 | 2.2.0 |
 
 
74
  ### Reference **MPU** inference time based on COCO Person dataset (see Accuracy for details on dataset)
75
  Model | Format | Resolution | Quantization | Board | Execution Engine | Frequency | Inference time (ms) | %NPU | %GPU | %CPU | X-LINUX-AI version | Framework |
76
  |-----------|--------|------------|---------------|-------------------|------------------|-----------|---------------------|-------|-------|------|--------------------|-----------------------|
77
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_256_quant_pc_uf_pose_coco-st.tflite) | Int8 | 256x256x3 | per-channel** | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 102.8 ms | 11.70 | 88.30 |0 | v6.1.0 | OpenVX |
78
+ | [YOLOv8n per tensor](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_256_quant_pt_uf_pose_coco-st.tflite) | Int8 | 256x256x3 | per-tensor | STM32MP257F-DK2 | NPU/GPU | 800 MHz | 17.57 ms | 86.79 | 13.21 |0 | v6.1.0 | OpenVX |
79
 
80
  ** **To get the most out of MP25 NPU hardware acceleration, please use per-tensor quantization**
81
 
 
87
 
88
  | Model | Format | Resolution | AP* |
89
  |-------|--------|------------|----------------|
90
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_192_quant_pc_uf_od_coco-person.tflite) | Int8 | 192x192x3 | 53.50 % |
91
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_256_quant_pc_uf_od_coco-person.tflite) | Int8 | 256x256x3 | 58.40 % |
92
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_320_quant_pc_uf_od_coco-person.tflite) | Int8 | 320x320x3 | 61.80 % |
93
+ | [YOLOv8n per channel](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/object_detection/yolov8n_416_quant_pc_uf_od_coco-person.tflite) | Int8 | 416x416x3 | 64.80 % |
94
 
95
+ \* EVAL_IOU = 0.5, NMS_THRESH = 0.5, SCORE_THRESH = 0.001, MAX_DETECTIONS = 100
96
 
97
  ## Integration in a simple example and other services support:
98
 
99
  Please refer to the stm32ai-modelzoo-services GitHub [here](https://github.com/STMicroelectronics/stm32ai-modelzoo-services).
100
  The models are stored in the Ultralytics repository. You can find them at the following link: [Ultralytics YOLOv8-STEdgeAI Models](https://github.com/stm32-hotspot/ultralytics/blob/main/examples/YOLOv8-STEdgeAI/stedgeai_models/).
101
 
102
+ Please refer to the [Ultralytics documentation](https://docs.ultralytics.com/tasks/detect/#train) to retrain the models.
103
 
104
  # References
105
 
 
126
  timestamp = {Mon, 13 Aug 2018 16:48:13 +0200},
127
  biburl = {https://dblp.org/rec/bib/journals/corr/LinMBHPRDZ14},
128
  bibsource = {dblp computer science bibliography, https://dblp.org}
129
+ }