Skip to content

HAILO15H_classification.rst

Latest commit

 

History

History
970 lines (523 loc) · 32 KB

HAILO15H_classification.rst

File metadata and controls

970 lines (523 loc) · 32 KB

Public Models

  • System host: Intel® Core™ i5-9400 CPU @ 2.90GHz
  • Hailo Dataflow Compiler Version v5.3.0
  • Measurement conditions: Measuring from the SoC, room temperature

Classification


Link Legend


Key / Icon Description
⭐ Networks used by Hailo-apps.
S Source – Link to the model's open-source repository.
PT Pretrained – Download the pretrained model file (ZIP format).
HEF, NV12, RGBX Compiled Models – Links to models in various formats: - HEF: RGB format - NV12: NV12 format - RGBX: RGBX format
PR Profiler Report – Download the model's performance profiling report.

Imagenet


Network Name float Accuracy (top1) Hardware Accuracy (top1) FPS (Batch Size=1) FPS (Batch Size=8) Links Input Resolution (HxWxC) Params (M) OPS (G)
cas_vit_m 81.2 81.1 81.2 207 384x384x3 12.42 10.89
cas_vit_s 79.9 79.8 112 295 384x384x3 5.5 5.4
cas_vit_t 81.9 81.6 63.3 135 384x384x3 21.76 20.85
davit_tiny 82.7 82.3 20.9 47.2 224x224x3 28.36 9.1
deit_base 80.9 79.8 52.5 141 224x224x3 80.26 35.22
deit_small 78.2 77.5 127 427 224x224x3 20.52 9.4
deit_tiny 69.1 68.6 148 513 224x224x3 5.3 2.57
efficientformer_l1 79.1 76.5 91.5 180 224x224x3 12.3 2.6
efficientnet_l 80.5 79.3 126 234 300x300x3 10.55 19.4
efficientnet_lite0 75.0 73.8 2215 2215 224x224x3 4.63 0.78
efficientnet_lite1 76.7 76.3 962 962 240x240x3 5.39 1.22
efficientnet_lite2 77.5 76.7 208 498 260x260x3 6.06 1.74
efficientnet_lite3 79.3 78.7 166 371 280x280x3 8.16 2.8
efficientnet_lite4 80.8 80.1 138 331 300x300x3 12.95 5.10
efficientnet_m 78.9 78.5 550 550 240x240x3 6.87 7.32
efficientnet_s 77.6 76.9 903 903 224x224x3 5.41 4.72
fastvit_sa12⭐ 79.8 76.6 284 915 224x224x3 11.99 3.59
hardnet39ds⭐ 73.4 73.0 569 1655 224x224x3 3.48 0.86
hardnet68⭐ 75.5 75.2 221 564 224x224x3 17.56 8.5
inception_v1 69.7 69.5 1307 1307 224x224x3 6.62 3
levit128 78.4 76.5 220 905 224x224x3 9.2 0.8
levit192 79.7 77.5 231 999 224x224x3 10.9 1.3
levit256 81.4 79.3 175 692 224x224x3 18.9 2.3
levit384 82.3 79.2 114 431 224x224x3 39.1 4.7
mobilenet_v1 71.0 70.4 4157 4157 224x224x3 4.22 1.14
mobilenet_v2_1.0 71.8 71.0 3454 3454 224x224x3 3.49 0.62
mobilenet_v2_1.4 74.2 73.2 1817 1817 224x224x3 6.09 1.18
mobilenet_v3 72.2 71.8 3281 3281 224x224x3 4.07 2
nextvit_base 83.2 83.3 48.5 207 224x224x3 44.8 16.6
nextvit_small 82.5 82.6 64.1 279 224x224x3 31.7 11.6
poolformer_s12 74.0 73.1 75.0 285 224x224x3 11.93 3.67
regnetx_1.6gf 77.0 76.6 2740 2740 224x224x3 9.17 3.22
regnetx_800mf 75.2 74.9 4776 4776 224x224x3 7.24 1.6
repghost_1_0x 73.0 72.3 353 1262 224x224x3 4.1 0.28
repghost_2_0x 77.2 77.0 210 781 224x224x3 9.8 1.04
repvgg_a1 74.4 72.3 2018 2018 224x224x3 12.79 4.7
repvgg_a2 76.5 74.4 300 644 224x224x3 25.5 10.2
resmlp12_relu 75.3 74.9 150 688 224x224x3 15.77 6.04
resnet_v1_18⭐ 71.3 70.7 2708 2708 224x224x3 11.68 3.64
resnet_v1_34⭐ 72.7 72.2 364 1043 224x224x3 21.79 7.34
resnet_v1_50⭐ 75.2 74.7 332 1046 224x224x3 25.53 6.98
resnext26_32x4d 76.2 75.9 861 861 224x224x3 15.37 4.96
resnext50_32x4d 79.3 78.4 253 641 224x224x3 24.99 8.48
squeezenet_v1.1 59.8 59.3 4308 4308 224x224x3 1.24 0.78
swin_small 83.1 80.0 27.3 89.4 224x224x3 50 17.6
swin_tiny 81.3 79.4 49.9 148 224x224x3 29 9.1
vit_base 84.5 83.6 57.3 177 224x224x3 86.5 35.188
vit_base_bn 80.0 79.1 72.2 209 224x224x3 86.5 35.188
vit_large 83.2 82.5 18.9 51.6 224x224x3 304.2 123.4
vit_small 81.5 80.5 117 416 224x224x3 21.12 8.62
vit_small_bn 78.1 77.3 166 664 224x224x3 21.12 8.62
vit_tiny 75.5 74.7 148 513 224x224x3 5.73 2.2
vit_tiny_bn 69.0 67.4 334 1499 224x224x3 5.73 2.2
⚡