Menu
close
Menu
close
Model Zoo
C++ Py

DX Model Zoo

Quantization Guide

  • Q-Lite (Standard): Standard INT8 quantization optimized for fast inference. Recommended as the default choice.
  • Q-Pro (Advanced): High-precision quantization with fine-tuning to maximize accuracy. (Note: Requires longer compilation time.)

Model Zoo & Test Environment

  • Total Models Available: 271
  • NPU Info: DX-M1 M.2 module
  • Host CPU: Intel(R) Core(TM) i5-14600K, 32G RAM
  • SDK Version: dx-com v2.3.0, dx-rt v3.3.0
  • Benchmark Cmd: run_model -m <MODEL_FILE> --use-ort -t 5 -b
  • Power Measurement: DX-M1 NPU chip only

* Note: Performance results may vary depending on the specific hardware configuration.

Image Classification (86)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
AlexNet ImageNet 224x224x3 0.72 61.10 BSD-3-Clause Top1 56.538 56.26 56.48 634 1,435.06
DeiTBase_384 ImageNet 384x384x3 58.06 86.86 Apache-2.0 Top1 83.094 78.53 - - - 15 22.38
DenseNet121 ImageNet 224x224x3 3.18 8.04 BSD-3-Clause Top1 74.434 73.77 - - - 59 180.85
DenseNet161 ImageNet 224x224x3 8.43 28.86 BSD-3-Clause Top1 77.108 77.08 - - - 24 79.32
DenseNet169 ImageNet 224x224x3 3.81 14.28 BSD-3-Clause Top1 75.584 75.51 - - - 41 136.93
DenseNet201 ImageNet 224x224x3 4.91 20.21 BSD-3-Clause Top1 76.884 75.87 - - - 28 98.16
EfficientFormer_L3 ImageNet 224x224x3 4.02 31.35 Apache-2.0 Top1 82.374 79.49 - - - 421 381.57
EfficientFormer_L7 ImageNet 224x224x3 10.36 82.14 Apache-2.0 Top1 83.306 82.17 - - - 139 148.18
EfficientNetB2 ImageNet 288x288x3 1.60 9.08 Apache-2.0 Top1 80.606 79.64 79.89 764 874.43
EfficientNetB3 ImageNet 300x300x3 2.58 12.19 BSD 3-Clause Top1 82.01 81.04 - - - 597 577.01
EfficientNetB4 ImageNet 380x380x3 5.96 19.28 BSD 3-Clause Top1 83.39 82.23 - - - 276 265.95
EfficientNetB5 ImageNet 456x456x3 13.38 30.30 BSD 3-Clause Top1 83.448 82.87 - - - 119 116.86
EfficientNetB6 ImageNet 528x528x3 24.34 42.93 BSD 3-Clause Top1 84.006 83.33 - - - 66 66.13
EfficientNetLite0 ImageNet 224x224x3 0.40 4.63 Apache-2.0 Top1 67.28 66.06 67.26 3,301 2,853.04
EfficientNetLite1 ImageNet 240x240x3 0.63 5.39 Apache-2.0 Top1 70.95 71.14 - - - 2,578 1,970.75
EfficientNetLite2 ImageNet 260x260x3 0.90 6.06 Apache-2.0 Top1 71.14 70.97 71.11 1,638 1,337.80
EfficientNetLite3 ImageNet 300x300x3 1.67 8.16 Apache-2.0 Top1 75.31 75.20 75.51 1,059 802.15
EfficientNetLite4 ImageNet 380x380x3 4.08 12.95 Apache-2.0 Top1 77.83 77.35 77.5 530 359.25
EfficientNetV2L ImageNet 480x480x3 60.99 118.26 Apache-2.0 Top1 85.794 85.33 - - - 79 32.05
EfficientNetV2S ImageNet 384x384x3 9.47 21.38 Apache-2.0 Top1 84.238 82.30 82.84 449 199.76
HarDNet39DS ImageNet 224x224x3 0.44 3.48 MIT Top1 72.08 71.46 71.68 2,073 2,607.10
HarDNet68 ImageNet 224x224x3 4.26 17.56 MIT Top1 76.474 76.32 76.39 629 433.49
InceptionV1 ImageNet 224x224x3 1.52 6.62 Apache-2.0 Top1 70.07 69.96 70.09 2,258 1,256.28
LeViT128 ImageNet 224x224x3 0.44 9.97 Apache-2.0 Top1 73.798 72.26 - - - 574 1,421.09
LeViT192 ImageNet 224x224x3 0.74 11.05 Apache-2.0 Top1 79.868 79.31 - - - 484 1,019.07
LeViT256 ImageNet 224x224x3 1.24 19.02 Apache-2.0 Top1 81.59 81.21 - - - 354 710.82
LeViT384 ImageNet 224x224x3 2.52 39.28 Apache-2.0 Top1 82.592 82.49 - - - 210 404.21
MnasNet0_5 ImageNet 224x224x3 0.11 2.21 Apache-2.0 Top1 67.75 65.04 - - - 6,784 7,277.97
MnasNet0_75 ImageNet 224x224x3 0.23 3.16 Apache-2.0 Top1 71.18 70.65 - - - 4,958 4,542.52
MnasNet1_0 ImageNet 224x224x3 0.33 4.36 Apache-2.0 Top1 73.468 73.06 - - - 4,318 3,713.45
MnasNet1_3 ImageNet 224x224x3 0.54 6.26 Apache-2.0 Top1 79.024 75.92 - - - 2,790 2,316.24
MobileNetV1 ImageNet 224x224x3 0.58 4.22 Apache-2.0 Top1 69.492 68.55 - - - 4,715 3,051.91
MobileNetV2 ImageNet 224x224x3 0.32 3.49 Apache-2.0 Top1 72.142 71.77 72.06 3,570 3,462.41
MobileNetV3Large ImageNet 224x224x3 0.23 5.47 Apache-2.0 Top1 75.256 73.55 73.94 3,080 3,863.54
OSNet0_25 ImageNet 224x224x3 0.14 0.71 MIT Top1 - 58.336 53.97 54.88 1,639 3,149.59
OSNet0_5 ImageNet 224x224x3 0.44 1.14 MIT Top1 - 69.446 58.53 63.93 1,543 2,061.63
RegNetX400MF ImageNet 224x224x3 0.42 5.48 Apache-2.0 Top1 74.884 74.16 74.49 1,374 2,349.58
RegNetX800MF ImageNet 224x224x3 0.81 7.24 Apache-2.0 Top1 77.522 76.96 77.29 1,025 1,391.44
RegNetX_16GF ImageNet 224x224x3 16.00 54.22 Apache-2.0 Top1 82.7 82.43 - - - 170 109.29
RegNetX_1_6GF ImageNet 224x224x3 1.62 9.17 Apache-2.0 Top1 79.684 79.19 - - - 718 722.80
RegNetX_1_6GF_3 ImageNet 224x224x3 1.62 9.17 Apache-2.0 Top1 77.06 76.86 - - - 718 716.80
RegNetX_32GF ImageNet 224x224x3 31.82 107.73 Apache-2.0 Top1 83.014 82.83 - - - 69 53.79
RegNetX_3_2GF ImageNet 224x224x3 3.20 15.27 Apache-2.0 Top1 81.186 80.63 - - - 532 461.49
RegNetX_8GF ImageNet 224x224x3 8.03 39.53 Apache-2.0 Top1 81.714 81.41 - - - 255 199.87
RegNetY200MF ImageNet 224x224x3 0.20 3.15 Apache-2.0 Top1 70.36 69.88 70.05 2,417 4,216.47
RegNetY400MF ImageNet 224x224x3 0.41 4.33 Apache-2.0 Top1 75.782 75.24 75.54 1,643 2,306.80
RegNetY800MF ImageNet 224x224x3 0.84 6.42 Apache-2.0 Top1 78.828 78.27 - - - 1,155 1,382.57
RegNetY_16GF ImageNet 384x384x3 46.92 83.53 Apache-2.0 Top1 86.014 85.62 - - - 37 34.88
RegNetY_1_6GF ImageNet 224x224x3 1.63 11.18 Apache-2.0 Top1 80.88 79.75 - - - 650 735.30
RegNetY_32GF ImageNet 384x384x3 95.07 144.97 BSD 3-Clause Top1 86.834 86.39 - - - 30 301.92
RegNetY_3_2GF ImageNet 224x224x3 3.21 19.40 Apache-2.0 Top1 81.97 81.27 - - - 342 400.36
RegNetY_8GF ImageNet 224x224x3 8.53 39.34 Apache-2.0 Top1 82.816 82.56 - - - 203 173.94
ResNet101 ImageNet 224x224x3 7.84 44.50 BSD-3-Clause Top1 81.898 81.47 81.65 639 281.81
ResNet152 ImageNet 224x224x3 11.57 60.12 BSD-3-Clause Top1 82.29 82.02 - - - 465 194.94
ResNet18 ImageNet 224x224x3 1.82 11.68 BSD-3-Clause Top1 69.754 69.57 69.64 2,663 1,204.87
ResNet18_BRECQ ImageNet 224x224x3 1.82 11.68 BSD-3-Clause Top1 70.992 70.64 - - - 2,659 1,216.91
ResNet34 ImageNet 224x224x3 3.67 21.79 BSD-3-Clause Top1 73.294 73.21 73.27 1,461 614.89
ResNet50 ImageNet 224x224x3 4.12 25.53 BSD-3-Clause Top1 80.854 80.54 80.69 1,067 515.02
ResNeXt101_64x4d ImageNet 224x224x3 15.53 83.35 BSD-3-Clause Top1 83.244 82.96 - - - 89 81.95
ResNeXt26_32x4d ImageNet 224x224x3 2.49 15.37 BSD-3-Clause Top1 75.852 75.60 75.66 857 643.45
ResNeXt50_32x4d ImageNet 224x224x3 4.27 24.99 BSD-3-Clause Top1 81.19 80.83 80.96 497 367.88
ResNeXt50_32x4d_imgclsmob ImageNet 224x224x3 4.27 24.99 BSD-3-Clause Top1 78.906 78.50 78.82 498 369.52
ShuffleNetV1_x1_0 ImageNet 224x224x3 0.15 2.42 Apache-2.0 Top1 65.314 65.63 - - - 857 2,149.92
ShuffleNetV2_x0_5 ImageNet 224x224x3 0.04 1.36 Apache-2.0 Top1 60.546 59.56 - - - 7,119 15,184.50
ShuffleNetV2_x1_0 ImageNet 224x224x3 0.15 2.27 Apache-2.0 Top1 69.348 68.72 - - - 4,560 6,595.10
ShuffleNetV2_x1_5 ImageNet 224x224x3 0.30 3.49 Apache-2.0 Top1 72.982 72.42 - - - 2,849 3,554.26
ShuffleNetV2_x2_0 ImageNet 224x224x3 0.59 7.38 Apache-2.0 Top1 76.224 75.67 - - - 1,914 2,172.27
SqueezeNet1_0 ImageNet 224x224x3 0.83 1.25 BSD-3-Clause Top1 58.088 57.05 - - - 2,177 1,753.69
SqueezeNet1_1 ImageNet 224x224x3 0.36 1.24 BSD-3-Clause Top1 58.18 57.26 - - - 4,553 4,562.94
SqueezeNet1_3 ImageNet 224x224x3 0.36 1.24 BSD-3-Clause Top1 60.682 59.70 - - - 4,566 4,653.14
VGG11 ImageNet 224x224x3 7.63 132.86 BSD-3-Clause Top1 69.034 68.84 68.96 287 266.28
VGG11BN ImageNet 224x224x3 7.63 132.86 BSD-3-Clause Top1 70.372 70.02 70.27 287 272.38
VGG13 ImageNet 224x224x3 11.34 133.05 BSD-3-Clause Top1 69.934 69.68 69.89 267 184.34
VGG13BN ImageNet 224x224x3 11.34 133.05 BSD-3-Clause Top1 71.556 71.40 - - - 267 208.96
VGG16 ImageNet 224x224x3 15.50 138.36 BSD-3-Clause Top1 71.582 71.38 - - - 252 142.25
VGG16BN ImageNet 224x224x3 15.50 138.36 BSD-3-Clause Top1 73.37 73.24 - - - 252 145.54
VGG19 ImageNet 224x224x3 19.67 143.67 BSD-3-Clause Top1 72.38 72.30 - - - 235 115.61
VGG19BN ImageNet 224x224x3 19.67 143.67 BSD-3-Clause Top1 74.238 74.07 74.22 235 140.23
ViT_Large_P32 ImageNet 224x224x3 15.57 306.54 BSD 3-Clause Top1 74.646 73.03 - - - 107 108.81
WideResNet101_2 ImageNet 224x224x3 22.80 126.82 BSD-3-Clause Top1 82.52 82.14 - - - 270 105.75
WideResNet50_2 ImageNet 224x224x3 11.43 68.85 BSD-3-Clause Top1 81.61 81.24 81.44 492 199.51
Ultralytics YOLO26-cls-l ImageNet 224x224x3 3.25 14.10 AGPL-3.0 Top1 79.034 78.32 - - - 871 614.73
Ultralytics YOLO26-cls-m ImageNet 224x224x3 2.58 11.62 AGPL-3.0 Top1 78.078 77.19 - - - 1,372 790.84
Ultralytics YOLO26-cls-n ImageNet 224x224x3 0.24 2.81 AGPL-3.0 Top1 71.394 67.87 - - - 3,581 5,833.08
Ultralytics YOLO26-cls-s ImageNet 224x224x3 0.82 6.72 AGPL-3.0 Top1 75.986 74.98 - - - 1,931 2,184.59
Ultralytics YOLO26-cls-x ImageNet 224x224x3 7.10 29.61 AGPL-3.0 Top1 79.902 79.36 - - - 465 279.73

Object Detection (116)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
damoyolo_tinynas_l20_m COCO 640x640x3 31.85 28.20 Apache-2.0 mAP 49.452 48.389 49.349 129 62.74
damoyolo_tinynas_l20_t COCO 640x640x3 9.13 8.50 Apache-2.0 mAP 42.543 41.36 42.477 163 193.77
damoyolo_tinynas_l25_s COCO 640x640x3 18.98 16.28 Apache-2.0 mAP 46.221 45.204 46.216 143 105.81
DamoYoloL COCO 640x640x3 50.07 42.06 Apache-2.0 mAP 50.36 49.427 49.161 101 40.02
DamoYoloM COCO 640x640x3 31.84 28.19 Apache-2.0 mAP 48.395 47.436 48.349 130 63.13
DamoYoloS COCO 640x640x3 18.96 16.27 Apache-2.0 mAP 46.524 45.055 46.014 144 108.77
DamoYoloT COCO 640x640x3 9.13 8.50 Apache-2.0 mAP 42.284 40.773 41.671 163 197.02
EfficientDet_D1 COCO 640x640x3 8.91 6.77 LGPL-3.0 mAP 30.227 28.862 - - - 85 142.54
EfficientDet_D2 COCO 768x768x3 15.38 8.35 LGPL-3.0 mAP 33.446 33.07 - - - 47 82.70
EfficientDet_D4 COCO 1024x1024x3 69.37 22.23 LGPL-3.0 mAP 42.706 40.341 - - - 11 19.16
NanoDet COCO 416x416x3 5.66 6.74 Apache-2.0 mAP 25.031 24.5 24.874 1,232 408.00
NanoDet_Plus COCO 416x416x3 0.80 1.19 Apache-2.0 mAP 13.086 12.571 - - - 1,330 1,190.56
NanoDet_Plus_15 COCO 416x416x3 1.54 2.46 Apache-2.0 mAP 14.469 14.418 - - - 908 760.89
NanoDet_RepVGGA COCO 640x640x3 21.44 10.79 Apache-2.0 mAP 29.33 28.973 29.187 435 113.21
NanoDet_RepVGGA12 COCO 640x640x3 14.19 4.87 Apache-2.0 mAP 29.695 27.392 - - - 368 181.22
SSDMV1 VOC2007Detection 300x300x3 1.55 9.46 Apache-2.0 mAP50 67.59 67.441 - - - 1,836 1,387.31
SSDMV2Lite VOC2007Detection 300x300x3 0.70 3.36 Apache-2.0 mAP50 68.704 68.429 68.744 1,640 1,664.59
SSDVGG16 VOC2007Detection 300x300x3 31.47 26.29 MIT mAP50 79.867 74.889 - - - 310 74.92
Ultralytics YOLO26-l COCO 640x640x3 46.42 24.85 AGPL-3.0 mAP 53.479 53.149 - - - 72 42.54
Ultralytics YOLO26-obb-l DOTAv1 1024x1024x3 124.25 25.67 AGPL-3.0 mAP 54.503 53.609 - - - 25 15.62
Ultralytics YOLO26-m COCO 640x640x3 36.57 20.45 AGPL-3.0 mAP 51.824 51.261 - - - 98 55.26
Ultralytics YOLO26-obb-m DOTAv1 1024x1024x3 98.70 21.27 AGPL-3.0 mAP 53.543 53.255 - - - 34 20.52
Ultralytics YOLO26-n COCO 640x640x3 3.35 2.45 AGPL-3.0 mAP 39.932 39.256 - - - 246 404.69
Ultralytics YOLO26-obb-n DOTAv1 1024x1024x3 8.96 2.51 AGPL-3.0 mAP 45.795 45.091 - - - 80 142.16
Ultralytics YOLO26-s COCO 640x640x3 11.63 9.54 AGPL-3.0 mAP 47.307 46.767 - - - 140 149.92
Ultralytics YOLO26-obb-s DOTAv1 1024x1024x3 31.52 9.82 AGPL-3.0 mAP 50.642 48.651 - - - 46 56.32
Ultralytics YOLO26-x COCO 640x640x3 101.72 55.77 AGPL-3.0 mAP 55.546 55.115 - - - 41 19.15
Ultralytics YOLO26-obb-x DOTAv1 1024x1024x3 272.11 57.62 AGPL-3.0 mAP 55.583 54.291 - - - 14 7.25
YoloV10B COCO 640x640x3 48.87 19.11 AGPL-3.0 mAP 52.111 50.711 51.453 114 41.17
YoloV10L COCO 640x640x3 63.65 24.41 AGPL-3.0 mAP 52.871 50.849 52.115 96 32.34
YoloV10M COCO 640x640x3 31.74 15.40 AGPL-3.0 mAP 50.855 49.164 50.059 125 60.15
YoloV10N COCO 640x640x3 4.02 2.34 AGPL-3.0 mAP 38.373 37.188 37.992 317 384.18
YoloV10N_PPU COCO 640x640x3 4.02 2.34 AGPL-3.0 mAP 38.331 36.934 - - - 117 269.50
YoloV10S COCO 640x640x3 12.04 7.29 AGPL-3.0 mAP 46.046 44.172 45.141 124 128.39
YoloV10X COCO 640x640x3 85.05 29.52 AGPL-3.0 mAP 54.029 53.214 53.399 57 22.21
Ultralytics YOLO11-l COCO 640x640x3 46.61 25.38 AGPL-3.0 mAP 52.32 52.082 52.05 84 43.94
Ultralytics YOLO11-m COCO 640x640x3 36.40 20.13 AGPL-3.0 mAP 50.519 50.343 50.12 124 57.95
Ultralytics YOLO11-n COCO 640x640x3 3.88 2.66 AGPL-3.0 mAP 38.635 38.259 38.447 341 415.48
Ultralytics YOLO11-n-ppu COCO 640x640x3 3.88 2.66 AGPL-3.0 mAP 38.634 37.971 - - - 119 292.96
Ultralytics YOLO11-s COCO 640x640x3 11.90 9.49 AGPL-3.0 mAP 45.847 45.592 45.755 203 159.57
Ultralytics YOLO11-x COCO 640x640x3 102.16 56.96 AGPL-3.0 mAP 53.56 53.399 53.319 46 19.46
YOLOV12N_PPU COCO 640x640x3 4.64 2.63 AGPL-3.0 mAP 39.907 36.478 - - - 39 129.54
YoloV3 COCO 640x640x3 81.13 61.92 GPL-3.0 mAP 46.654 46.586 46.432 110 26.68
YoloV3_416 COCO 416x416x3 33.09 61.92 GPL-3.0 mAP - 40.477 40.313 - - - 223 60.77
YOLOV3_416_PPU COCO 416x416x3 33.09 61.92 GPL-3.0 mAP 39.972 39.662 - - - 217 61.50
YOLOV3_608_PPU COCO 608x608x3 70.69 61.92 GPL-3.0 mAP 42.513 42.238 - - - 113 28.70
YoloV3_Gluon_416 COCO 416x3x416 33.06 61.92 Apache-2.0 mAP 33.359 33.149 - - - 74 56.15
YoloV3_Gluon_608 COCO 608x3x608 70.62 61.92 Apache-2.0 mAP 35.791 35.379 - - - 32 24.87
YoloV3_Tiny COCO 416x416x3 2.81 8.85 GPL-3.0 mAP - 17.608 17.016 - - - 931 683.17
YoloV4_Leaky_512 COCO 512x512x3 45.86 64.33 Unlicense mAP 47.215 45.767 - - - 166 43.11
YOLOV4_PPU COCO 512x512x3 51.04 64.33 Unlicense mAP 45.69 44.835 - - - 163 43.81
YoloV4Tiny_416 COCO 416x416x3 3.48 6.05 Unlicense mAP 20.657 20.37 - - - 663 620.16
Ultralytics YOLOv5-l COCO 640x640x3 57.10 46.53 AGPL-3.0 mAP 48.74 48.646 48.506 149 37.52
Ultralytics YOLOv5-l6-1280 COCO 1280x1280x3 233.00 76.73 AGPL-3.0 mAP 52.936 53.132 - - - 36 8.99
Ultralytics YOLOv5-l-640 COCO 640x640x3 59.17 47.80 AGPL-3.0 mAP 46.019 46.155 - - - 42 29.55
Ultralytics YOLOv5-m COCO 640x640x3 26.07 21.17 AGPL-3.0 mAP 45.082 44.777 44.821 241 80.33
Ultralytics YOLOv5-m6 COCO 640x640x3 26.07 21.27 AGPL-3.0 mAP 45.082 44.658 44.735 242 77.42
Ultralytics YOLOv5-m6-1280 COCO 1280x1280x3 106.41 35.70 AGPL-3.0 mAP 50.578 50.754 - - - 53 18.81
Ultralytics YOLOv5-m6-6-1280 COCO 1280x1280x3 106.41 36.11 AGPL-3.0 mAP 51.079 50.754 - - - 53 18.55
Ultralytics YOLOv5-m-640 COCO 640x640x3 26.59 21.79 AGPL-3.0 mAP 42.482 42.54 - - - 56 56.33
Ultralytics YOLOv5-m-WoSpp-640 COCO 640x640x3 26.83 22.68 AGPL-3.0 mAP 42.997 42.354 - - - 57 58.77
Ultralytics YOLOv5-n COCO 640x640x3 2.71 1.87 AGPL-3.0 mAP 28.081 27.126 27.646 365 621.60
Ultralytics YOLOv5-n6-1280 COCO 1280x1280x3 11.05 3.65 AGPL-3.0 mAP 35.812 35.329 - - - 86 150.61
Ultralytics YOLOv5-n-61-1280 COCO 1280x1280x3 11.05 3.65 AGPL-3.0 mAP 35.813 35.328 - - - 85 144.94
Ultralytics YOLOv5-s COCO 640x640x3 9.10 7.23 AGPL-3.0 mAP 37.451 37.001 37.158 334 224.76
Ultralytics YOLOv5-s6-1280 COCO 1280x1280x3 37.16 12.61 AGPL-3.0 mAP 44.49 44.076 - - - 77 56.05
Ultralytics YOLOv5-s-61-1280 COCO 1280x1280x3 37.16 13.02 AGPL-3.0 mAP 44.49 44.076 - - - 77 56.04
Ultralytics YOLOv5-s-320 COCO 320x320x3 2.35 7.27 AGPL-3.0 mAP 30.515 30.234 - - - 1,476 879.80
Ultralytics YOLOv5-s-640 COCO 640x640x3 9.02 7.46 AGPL-3.0 mAP 35.652 35.312 - - - 98 152.72
Ultralytics YOLOv5-s-BboxDecoding-640 COCO 640x640x3 8.95 7.46 AGPL-3.0 mAP 35.303 34.92 - - - 98 156.57
Ultralytics YOLOv5-s-C3tr-640 COCO 640x640x3 9.51 7.37 AGPL-3.0 mAP 37.351 36.793 - - - 282 210.59
Ultralytics YOLOv5-s-ppu COCO 640x640x3 9.10 7.23 AGPL-3.0 mAP 37.03 36.644 - - - 416 231.33
Ultralytics YOLOv5-s-WoSpp-640 COCO 640x640x3 9.08 7.86 AGPL-3.0 mAP 34.488 34.079 - - - 98 145.98
Ultralytics YOLOv5-x6-1280 COCO 1280x1280x3 434.52 141.14 AGPL-3.0 mAP 54.758 54.368 - - - 17 4.54
Ultralytics YOLOv5-x-640 COCO 640x640x3 106.52 86.71 AGPL-3.0 mAP 50.507 50.334 - - - 72 18.95
Ultralytics YOLOv5-xs-WoSpp-512 COCO 512x512x3 5.81 7.86 AGPL-3.0 mAP 32.825 32.317 - - - 170 241.26
YoloV6L_640 COCO 640x640x3 77.96 59.61 GPL-3.0 mAP 52.252 52.05 - - - 96 26.75
YoloV6M_640 COCO 640x640x3 43.19 34.86 GPL-3.0 mAP 49.566 49.09 - - - 129 53.33
YoloV6N COCO 640x640x3 5.79 4.65 GPL-3.0 mAP 36.838 35.86 - - - 600 365.18
YoloV6N0_1_0 COCO 640x640x3 5.64 4.32 GPL-3.0 mAP 34.722 32.061 33.572 550 385.33
YoloV6N0_2_1 COCO 640x640x3 5.64 4.32 GPL-3.0 mAP 35.416 35.123 35.278 626 364.52
YoloV6N_NmsCore_640 COCO 640x640x3 5.64 4.35 GPL-3.0 mAP 35.502 35.263 - - - 626 363.79
YoloV7 COCO 640x640x3 55.28 36.92 GPL-3.0 mAP 50.86 50.967 50.823 128 38.83
YOLOV7_PPU COCO 640x640x3 55.28 36.92 GPL-3.0 mAP 50.841 50.68 - - - 123 39.51
YoloV7_W6 COCO 1280x1280x3 187.39 70.39 GPL-3.0 mAP 54.37 54.285 - - - 43 11.50
YoloV7_W6_wo_decoding COCO 1280x1280x3 187.11 70.39 GPL-3.0 mAP 54.37 54.285 - - - 43 11.46
YoloV7_wo_decoding COCO 640x640x3 55.21 36.91 GPL-3.0 mAP 51.069 50.965 - - - 128 39.22
YoloV7_X COCO 640x640x3 98.83 71.33 GPL-3.0 mAP 52.867 52.631 - - - 76 21.14
YoloV7D6_1280 COCO 1280x1280x3 365.38 133.76 GPL-3.0 mAP 56.097 56.002 - - - 21 5.68
YoloV7E6 COCO 1280x1280x3 269.21 97.20 GPL-3.0 mAP 55.216 55.648 55.472 25 7.58
YoloV7E6E_1280 COCO 1280x1280x3 439.22 151.69 GPL-3.0 mAP 56.549 56.423 - - - 13 4.39
YoloV7Tiny COCO 640x640x3 7.01 6.24 GPL-3.0 mAP 37.289 36.988 37.098 332 249.45
YOLOV7X_PPU COCO 640x640x3 98.83 71.33 GPL-3.0 mAP 52.544 52.276 - - - 74 21.07
Ultralytics YOLOv8-l COCO 640x640x3 85.13 43.69 AGPL-3.0 mAP 52.573 51.688 51.681 104 25.39
Ultralytics YOLOv8-m COCO 640x640x3 41.13 25.91 AGPL-3.0 mAP 50.111 49.237 49.225 162 49.72
Ultralytics YOLOv8-n COCO 640x640x3 4.89 3.18 AGPL-3.0 mAP 37.316 36.308 36.628 456 367.39
Ultralytics YOLOv8-n-ppu COCO 640x640x3 4.89 3.18 AGPL-3.0 mAP 36.694 36.115 - - - 134 266.29
Ultralytics YOLOv8-s COCO 640x640x3 15.24 11.18 AGPL-3.0 mAP 44.803 44.049 44.119 366 134.33
Ultralytics YOLOv8-s-decoding COCO 640x640x3 15.24 11.16 AGPL-3.0 mAP 44.212 44.108 - - - 365 137.05
Ultralytics YOLOv8-s-ppu COCO 640x640x3 15.24 11.18 AGPL-3.0 mAP 44.212 43.707 - - - 123 121.35
Ultralytics YOLOv8-x COCO 640x640x3 132.08 68.23 AGPL-3.0 mAP 53.635 52.84 - - - 57 15.40
YoloV9_GELAN_C COCO 640x640x3 53.92 25.31 GPL-3.0 mAP 52.161 51.761 - - - 101 38.13
YoloV9C COCO 640x640x3 53.92 25.31 GPL-3.0 mAP 52.16 41.027 45.475 101 38.72
YoloV9M COCO 640x640x3 40.41 20.00 GPL-3.0 mAP 50.393 50.134 - - - 147 51.31
YoloV9S COCO 640x640x3 14.50 7.13 GPL-3.0 mAP 45.99 45.197 44.895 315 133.83
YoloV9T COCO 640x640x3 4.56 2.03 GPL-3.0 mAP 37.705 36.429 36.745 395 339.20
YOLOV9T_PPU COCO 640x640x3 4.56 2.03 GPL-3.0 mAP 37.707 36.421 - - - 126 246.76
YoloXL_640 COCO 640x640x3 80.77 54.17 Apache-2.0 mAP 49.681 49.37 - - - 113 26.18
YoloXLLeaky COCO 640x640x3 78.01 54.17 Apache-2.0 mAP 48.623 48.515 - - - 113 25.71
YoloXM_640 COCO 640x640x3 38.74 25.30 Apache-2.0 mAP 46.721 46.503 - - - 189 52.76
YoloXS COCO 640x640x3 14.41 8.96 Apache-2.0 mAP 40.29 40.051 40.107 412 143.52
YOLOXS_PPU COCO 640x640x3 14.41 8.96 Apache-2.0 mAP 40.336 40.097 - - - 276 135.79
YoloXSLeaky COCO 640x640x3 13.49 8.96 Apache-2.0 mAP 38.298 37.965 38.028 413 140.77
YoloXSWideLeaky COCO 640x640x3 29.89 20.12 Apache-2.0 mAP 42.636 42.326 42.471 227 65.61
YoloXTiny COCO 416x416x3 3.55 5.05 Apache-2.0 mAP 32.605 32.315 32.465 972 520.28
YoloXX_640 COCO 640x640x3 145.24 99.02 Apache-2.0 mAP 51.107 50.887 - - - 56 13.77

Semantic Segmentation (8)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
BiSeNetV1 Cityscapes 1024x2048x3 118.98 13.27 Apache-2.0 mIoU 75.367 75.056 - - - 20 14.92
BiSeNetV2 Cityscapes 1024x2048x3 99.14 3.35 Apache-2.0 mIoU 74.951 74.616 74.771 29 19.17
DeepLabV3MobilenetV2 VOCSegmentation 512x512x3 5.99 5.10 Apache-2.0 mIoU 70.051 68.91 - - - 331 271.53
DeepLabV3PlusMobilenet VOCSegmentation 512x512x3 26.62 5.80 Apache-2.0 mIoU 70.806 67.637 68.879 258 74.37
DeepLabV3PlusMobileNetV2 VOCSegmentation 512x512x3 16.99 5.21 Apache-2.0 mIoU 70.809 70.157 - - - 269 110.83
SegFormer_b0_512x1024 CityScapes 512x1024x3 20.79 3.72 Apache-2.0 mIoU 70.542 67.974 - - - 383 1,396.54
SegFormer_b0_512x1024_H CityScapes 512x1024x3 21.45 3.72 NVIDIA License mIoU 71.712 69.084 - - - 382 1,393.13
UNet_MobileNetV2 OxfordPet 256x3x256 4.70 10.08 Apache-2.0 mIoU 77.111 75.706 - - - 248 927.95

Instance Segmentation (14)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
YOLACT_RegNetX_1_6gf COCO 512x512x3 63.00 18.02 MIT mAP 19.426 22.026 - - - 62 32.63
YOLACT_RegNetX_800mf COCO 512x512x3 58.68 16.23 MIT mAP 18.165 20.803 - - - 68 36.66
Ultralytics YOLO26-seg-l COCO 640x640x3 78.53 28.01 AGPL-3.0 mAP 39.665 44.557 - - - 54 25.69
Ultralytics YOLO26-seg-m COCO 640x640x3 68.69 23.61 AGPL-3.0 mAP 43.5 43.081 - - - 67 29.26
Ultralytics YOLO26-seg-n COCO 640x640x3 5.68 2.77 AGPL-3.0 mAP 30.921 33.019 - - - 189 260.73
Ultralytics YOLO26-seg-s COCO 640x640x3 19.88 10.44 AGPL-3.0 mAP 35.615 39.16 - - - 107 89.45
Ultralytics YOLO26-seg-x COCO 640x640x3 173.48 62.86 AGPL-3.0 mAP 40.619 46.039 - - - 29 11.34
Ultralytics YOLOv5-seg-l COCO 640x640x3 76.77 47.89 AGPL-3.0 mAP 39.293 39.34 39.301 110 27.68
Ultralytics YOLOv5-seg-m COCO 640x640x3 37.29 21.97 AGPL-3.0 mAP 36.061 36.122 - - - 152 53.31
Ultralytics YOLOv5-seg-n COCO 640x640x3 4.11 1.99 AGPL-3.0 mAP 22.866 22.353 22.691 202 390.36
Ultralytics YOLOv5-seg-s COCO 640x640x3 14.23 7.61 AGPL-3.0 mAP 31.079 30.832 30.967 187 140.55
Ultralytics YOLOv8-seg-m COCO 640x640x3 57.03 27.29 AGPL-3.0 mAP 39.683 39.649 - - - 117 34.39
Ultralytics YOLOv8-seg-n COCO 640x640x3 6.95 3.40 AGPL-3.0 mAP 29.775 29.835 29.81 264 250.51
Ultralytics YOLOv8-seg-s COCO 640x640x3 22.43 11.84 AGPL-3.0 mAP 36.044 35.999 35.996 225 87.14

Zero Shot Instance Segmentation (1)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
FastSAM_S COCO 1024x1024x3 57.15 11.84 AGPL-3.0 AR10 12.95 12.84 - - - 89 36.30

Pose Estimation (10)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
CenterPose_RegNetX_1_6GF_FPN COCOPose 512x512x3 42.33 12.37 MIT mAP 24.164 23.837 - - - 143 52.41
CenterPose_RegNetX_800MF COCOPose 640x640x3 32.56 14.28 MIT mAP 29.72 29.037 - - - 79 47.84
CenterPose_RepVGG_A0 COCOPose 416x416x3 14.21 11.71 MIT mAP 16.727 16.271 - - - 319 138.33
Ultralytics YOLO26-pose-l COCOPose 640x640x3 48.89 26.00 AGPL-3.0 mAP 66.268 63.825 - - - 70 40.00
Ultralytics YOLO26-pose-m COCOPose 640x640x3 39.05 21.61 AGPL-3.0 mAP 65.358 64.278 - - - 95 50.38
Ultralytics YOLO26-pose-n COCOPose 640x640x3 4.41 2.99 AGPL-3.0 mAP 52.774 51.849 - - - 235 315.58
Ultralytics YOLO26-pose-s COCOPose 640x640x3 13.25 10.43 AGPL-3.0 mAP 59.513 56.255 - - - 135 131.46
Ultralytics YOLO26-pose-x COCOPose 640x640x3 105.69 57.63 AGPL-3.0 mAP 67.959 68.145 - - - 40 18.42
Ultralytics YOLOv8-pose-m COCOPose 640x640x3 42.18 26.49 AGPL-3.0 mAP 63.196 61.582 62.088 158 48.41
Ultralytics YOLOv8-pose-s COCOPose 640x640x3 16.05 11.66 AGPL-3.0 mAP 58.342 57.199 57.923 346 130.12

Image De-noising (6)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
DnCNN_15 BSD68 512x512x1 145.79 0.56 Apache-2.0 PSNR
SSIM		 31.709
0.8905		 31.2214
0.8856		 - - - 48 13.28
DnCNN_25 BSD68 512x512x1 145.79 0.56 Apache-2.0 PSNR
SSIM		 29.1919
0.8276		 28.5335
0.8160		 - - - 47 13.14
DnCNN_50 BSD68 512x512x1 145.79 0.56 Apache-2.0 PSNR
SSIM		 26.1882
0.7184		 24.7169
0.6708		 - - - 47 12.77
DnCNN_Color CBSD68 512x512x3 175.49 0.67 MIT PSNR
SSIM		 33.825
0.929		 31.4781
0.8609		 - - - 42 11.07
DnCNN_Gray BSD68 512x512x1 174.89 0.67 MIT PSNR
SSIM		 31.409
0.882		 30.9141
0.8750		 - - - 42 12.24
DnCNN_Gray_V2 BSD68 512x512x1 174.89 0.67 MIT PSNR
SSIM		 25.688
0.710		 19.7828
0.3416		 - - - 42 11.29

Depth Estimation (2)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
FastDepth NYU 224x224x3 547.19 1.38 MIT RMSE 0.604 0.608 - - - 724 1,930.27
SCDepthV3 NYU 256x320x3 5.36 14.32 GPL-3.0 RMSE 0.881 0.929 - - - 330 305.07

Face Detection (15)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
RetinaFace_MobileNet025 WiderFace 640x640x3 1.02 0.42 MIT AP(Easy)
AP(Med)
AP(Hard)		 89.266
83.048
54.068		 88.216
81.525
52.227		 - - - 425 838.47
SCRFD10G WiderFace 640x640x3 13.41 4.23 MIT AP(Easy)
AP(Med)
AP(Hard)		 95.469
94.021
82.674		 95.4
93.979
82.635		 95.448 341 139.83
SCRFD2_5G WiderFace 640x640x3 3.46 0.82 MIT AP(Easy)
AP(Med)
AP(Hard)		 93.888
92.042
77.000		 93.854
92.041
76.867		 - - - 445 416.20
SCRFD500M WiderFace 640x640x3 0.77 0.63 MIT AP(Easy)
AP(Med)
AP(Hard)		 91.080
88.467
69.375		 90.768
88.29
69.001		 90.941 561 1,042.18
ULFGFD_RFB_320 WiderFace 240x320x3 0.19 0.28 MIT AP(Easy)
AP(Med)
AP(Hard)		 73.690
60.070
30.614		 73.28
59.723
30.34		 - - - 4,760 7,540.98
ULFGFD_RFB_320_WO_PP WiderFace 240x320x3 0.19 0.28 MIT AP(Easy)
AP(Med)
AP(Hard)		 73.639
60.038
30.599		 73.214
59.683
30.322		 - - - 4,735 7,785.88
ULFGFD_RFB_640 WiderFace 480x640x3 0.77 0.28 MIT AP(Easy)
AP(Med)
AP(Hard)		 80.459
74.865
44.845		 80.325
74.518
44.429		 - - - 860 1,725.53
ULFGFD_Slim_320 WiderFace 240x320x3 0.17 0.26 MIT AP(Easy)
AP(Med)
AP(Hard)		 70.649
54.521
25.639		 69.97
54.055
25.266		 - - - 5,120 9,115.46
ULFGFD_Slim_320_WO_PP WiderFace 240x320x3 0.17 0.26 MIT AP(Easy)
AP(Med)
AP(Hard)		 70.539
54.453
25.610		 69.856
53.986
25.235		 - - - 5,198 8,833.02
Ultralytics YOLOv5-face-m WiderFace 640x640x3 25.84 21.04 AGPL-3.0 AP(Easy)
AP(Med)
AP(Hard)		 95.507
94.027
85.649		 95.576
94.097
86.296		 95.675 232 75.41
Ultralytics YOLOv5-face-s WiderFace 640x640x3 8.53 7.06 AGPL-3.0 AP(Easy)
AP(Med)
AP(Hard)		 94.570
92.940
83.698		 94.631
93.036
84.5		 94.736 423 236.32
YOLOv7_Face WiderFace 640x640x3 54.63 36.56 GPL-3.0 AP(Easy)
AP(Med)
AP(Hard)		 96.925
95.689
88.337		 96.974
95.699
88.266		 97.008 131 39.49
YOLOv7_W6_Face WiderFace 960x960x3 100.22 69.90 GPL-3.0 AP(Easy)
AP(Med)
AP(Hard)		 96.410
95.091
88.610		 96.452
95.161
88.726		 96.473 78 21.08
YOLOv7_W6_TTA_Face WiderFace 1280x1280x3 178.16 69.90 GPL-3.0 AP(Easy)
AP(Med)
AP(Hard)		 95.890
94.929
89.952		 96.025
95.071
90.359		 - - - 44 11.91
YOLOv7s_Face WiderFace 640x640x3 9.35 4.27 GPL-3.0 AP(Easy)
AP(Med)
AP(Hard)		 94.860
93.300
85.304		 94.886
93.26
85.252		 94.985 342 198.29

Face Attribute Recognition (1)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
FaceAttrResnetV1_18 CelebA 218x178x3 1.50 11.74 MIT Average Accuracy 0.912 0.911 - - - 2,143 1,383.00

Face Recognition (4)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
ArcFace_IResNet100_MS1M LFW 112x112x3 12.52 65.23 MIT Accuracy 0.987 0.867 - - - 228 149.91
ArcFace_IResNet50_MS1M LFW 112x112x3 6.40 43.59 MIT Accuracy 0.984 0.727 - - - 284 258.32
ArcFace_MobileFaceNet LFW 112x112x3 0.45 2.04 MIT Accuracy 0.939 0.936 - - - 2,210 3,206.07
ArcFace_R50 LFW 112x112x3 6.39 31.03 MIT Accuracy 0.97 0.754 - - - 587 336.96

Face Landmark Detection (2)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
TDDFA_V2_MobileNet05 AFLW20003D 120x120x3 0.07 0.86 MIT NME 3.499 3.532 - - - 15,798 24,849.00
TDDFA_V2_MobileNetV1 AFLW20003D 120x120x3 0.23 3.29 MIT NME 3.449 3.541 - - - 8,110 8,880.91

Person Attribute (2)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
DeepMAR_ResNet18 PETA 224x224x3 1.82 11.19 No License Average Accuracy 82.563 0.825 - - - 2,729 1,246.33
DeepMAR_ResNet50 PETA 224x224x3 4.12 23.55 No License Average Accuracy 84.39 0.844 - - - 1,116 510.87

Hand Landmark (1)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
MediaPipeHandsLite HandLandmark 224x224x3 0.16 1.01 Apache-2.0 MNAE 0.14 0.149 - - - 5,724 5,713.71

Low Light Enhancement (1)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
ZeroDCE LOL 400x600x3 19.75 79.42 CC BY-NC 4.0 PSNR
SSIM		 15.001
0.481		 15.0893
0.4863		 - - - 26 56.65

Super Resolution (2)

Class Name Dataset Input
Resolution		 Operations
(GFLOPs)		 Parameters
(M)		 License Metric Source Original (FP32) Quantized (INT8) Sample
Apps
Q-Lite Q-Pro Performance
Accuracy ONNX Accuracy DXNN JSON Accuracy DXNN JSON FPS FPS/Watt
ESPCN_x3 BSD100 17x17x1 0.01 0.02 BSD 3-Clause PSNR
SSIM		 23.873
0.997		 34.8389
0.999752		 - - - 18,944 95,182.60
ESPCN_x4 BSD100 17x17x1 0.01 0.03 BSD 3-Clause PSNR
SSIM		 20.357
0.993		 33.9080
0.999710		 - - - 19,026 91,660.80

Generated on 2026-04-21 07:06:17

DX Model Zoo - Powered by DeepX DevOps

DEEPX HQ
5F, 20, Pangyoyeok-ro 241beon-gil, Bundang-gu,
Seongnam-si, Gyeonggi-do, Republic of Korea
US Office
3003 North First Street #316, San Jose, CA 95135
Taiwan Office
Rm. 950, 9F., No. 208, Sec. 3, Nanjing E. Rd.,
Zhongshan Dist., Taipei City 104, Taiwan (R.O.C.)
Contact US
E-mail: tech_support@deepx.ai
Copyrightⓒ 2025 DEEPX Co., Ltd. All Rights Reserved.