Here, we give the full list of publicly pre-trained models supported by the Hailo Model Zoo.
- Benchmark Networks are marked with
- Networks available in TAPPAS are marked with
- Benchmark and TAPPAS networks run in performance mode
- All models were compiled using Hailo Dataflow Compiler v3.29.0
| Network Name | mIoU | HW Accuracy | FPS (Batch Size=1) | FPS (Batch Size=8) | Input Resolution (HxWxC) | Params (M) | OPS (G) | Pretrained | Source | Compiled |
|---|---|---|---|---|---|---|---|---|---|---|
| fcn8_resnet_v1_18 | 69.41 | 69.26 | 18 | 20 | 1024x1920x3 | 11.20 | 142.82 | download | link | rgbx |
| segformer_b0_bn | 69.81 | 68.18 | 11 | 19 | 512x1024x3 | 3.72 | 35.76 | download | link | rgbx |
stdc1  |
74.55 | 73.95 | 16 | 24 | 1024x1920x3 | 8.27 | 126.47 | download | link | rgbx |
| Network Name | mIoU | HW Accuracy | FPS (Batch Size=1) | FPS (Batch Size=8) | Input Resolution (HxWxC) | Params (M) | OPS (G) | Pretrained | Source | Compiled |
|---|---|---|---|---|---|---|---|---|---|---|
| unet_mobilenet_v2 | 77.32 | 77.02 | 146 | 245 | 256x256x3 | 10.08 | 28.88 | download | link | rgbx |
| Network Name | mIoU | HW Accuracy | FPS (Batch Size=1) | FPS (Batch Size=8) | Input Resolution (HxWxC) | Params (M) | OPS (G) | Pretrained | Source | Compiled |
|---|---|---|---|---|---|---|---|---|---|---|
| deeplab_v3_mobilenet_v2 | 76.04 | 74.58 | 43 | 57 | 513x513x3 | 2.10 | 17.65 | download | link | rgbx |
| deeplab_v3_mobilenet_v2_wo_dilation | 71.46 | 71.05 | 86 | 119 | 513x513x3 | 2.10 | 3.21 | download | link | rgbx |