mirrors
/
PaddleClas
mirror of https://github.com/PaddlePaddle/PaddleClas.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
PaddleClas/docs/zh_CN/models/HRNet.md

74 lines
6.5 KiB
Markdown
Raw Normal View History

add menu and modify documentation
2021-11-25 15:40:40 +08:00
# HRNet 系列
-----
## 目录
init docs
2020-04-10 00:45:02 +08:00
add menu and modify documentation
2021-11-25 15:40:40 +08:00
* [1. 概述](#1)
* [2. 精度、FLOPS 和参数量](#2)
* [3. 基于 V100 GPU 的预测速度](#3)
* [4. 基于 T4 GPU 的预测速度](#4)
batch ad figs
2020-04-13 23:54:13 +08:00
add menu and modify documentation
2021-11-25 15:40:40 +08:00
<a name='1'></a>
## 1. 概述
HRNet 是 2019 年由微软亚洲研究院提出的一种全新的神经网络，不同于以往的卷积神经网络，该网络在网络深层仍然可以保持高分辨率，因此预测的关键点热图更准确，在空间上也更精确。此外，该网络在对分辨率敏感的其他视觉任务中，如检测、分割等，表现尤为优异。
该系列模型的 FLOPS、参数量以及 T4 GPU 上的预测耗时如下图所示。
batch ad figs
2020-04-13 23:54:13 +08:00
add T4 benchmark
2020-05-09 16:46:11 +08:00
![](../../images/models/T4_benchmark/t4.fp32.bs4.HRNet.flops.png)
batch ad figs
2020-04-13 23:54:13 +08:00
add T4 benchmark
2020-05-09 16:46:11 +08:00
![](../../images/models/T4_benchmark/t4.fp32.bs4.HRNet.params.png)
![](../../images/models/T4_benchmark/t4.fp32.bs4.HRNet.png)
batch ad figs
2020-04-13 23:54:13 +08:00
add fp16 pngs
2020-05-09 19:18:17 +08:00
![](../../images/models/T4_benchmark/t4.fp16.bs4.HRNet.png)
add menu and modify documentation
2021-11-25 15:40:40 +08:00
目前 PaddleClas 开源的这类模型的预训练模型一共有 7 个，其指标如图所示，其中 HRNet_W48_C 指标精度异常的原因可能是因为网络训练的正常波动。
init docs
2020-04-10 00:45:02 +08:00
add menu and modify documentation
2021-11-25 15:40:40 +08:00
<a name='2'></a>
## 2. 精度、FLOPS 和参数量
init docs
2020-04-10 00:45:02 +08:00
| Models | Top1 | Top5 | Reference<br>top1 | Reference<br>top5 | FLOPS<br>(G) | Parameters<br>(M) |
|:--:|:--:|:--:|:--:|:--:|:--:|:--:|
| HRNet_W18_C | 0.769 | 0.934 | 0.768 | 0.934 | 4.140 | 21.290 |
polish dygraph doc and deploy
2020-10-19 13:05:34 +08:00
| HRNet_W18_C_ssld | 0.816 | 0.958 | 0.768 | 0.934 | 4.140 | 21.290 |
init docs
2020-04-10 00:45:02 +08:00
| HRNet_W30_C | 0.780 | 0.940 | 0.782 | 0.942 | 16.230 | 37.710 |
| HRNet_W32_C | 0.783 | 0.942 | 0.785 | 0.942 | 17.860 | 41.230 |
| HRNet_W40_C | 0.788 | 0.945 | 0.789 | 0.945 | 25.410 | 57.550 |
| HRNet_W44_C | 0.790 | 0.945 | 0.789 | 0.944 | 29.790 | 67.060 |
| HRNet_W48_C | 0.790 | 0.944 | 0.793 | 0.945 | 34.580 | 77.470 |
polish dygraph doc and deploy
2020-10-19 13:05:34 +08:00
| HRNet_W48_C_ssld | 0.836 | 0.968 | 0.793 | 0.945 | 34.580 | 77.470 |
init docs
2020-04-10 00:45:02 +08:00
| HRNet_W64_C | 0.793 | 0.946 | 0.795 | 0.946 | 57.830 | 128.060 |
Add se_hrnet_w64_c_ssld_pretrained (#452) add se hrnet64_c
2020-12-06 17:52:34 +08:00
| SE_HRNet_W64_C_ssld | 0.847 | 0.973 | | | 57.830 | 128.970 |
init docs
2020-04-10 00:45:02 +08:00
add menu and modify documentation
2021-11-25 15:40:40 +08:00
<a name='3'></a>
## 3. 基于 V100 GPU 的预测速度
init docs
2020-04-10 00:45:02 +08:00
add speed in V100 and mobile
2021-12-13 20:42:03 +08:00
| Models | Crop Size | Resize Short Size | FP32<br/>Batch Size=1<br/>(ms) | FP32<br/>Batch Size=4<br/>(ms) | FP32<br/>Batch Size=8<br/>(ms) |
|-------------|-----------|-------------------|-------------------|-------------------|-------------------|
| HRNet_W18_C | 224 | 256 | 6.66 | 8.94 | 11.95 |
| HRNet_W18_C_ssld | 224 | 256 | 6.66 | 8.92 | 11.93 |
| HRNet_W30_C | 224 | 256 | 8.61 | 11.40 | 15.23 |
| HRNet_W32_C | 224 | 256 | 8.54 | 11.58 | 15.57 |
| HRNet_W40_C | 224 | 256 | 9.83 | 15.02 | 20.92 |
| HRNet_W44_C | 224 | 256 | 10.62 | 16.18 | 25.92 |
| HRNet_W48_C | 224 | 256 | 11.07 | 17.06 | 27.28 |
| HRNet_W48_C_ssld | 224 | 256 | 11.09 | 17.04 | 27.28 |
| HRNet_W64_C | 224 | 256 | 13.82 | 21.15 | 35.51 |
add T4 benchmark
2020-05-09 16:46:11 +08:00
add menu and modify documentation
2021-11-25 15:40:40 +08:00
<a name='4'></a>
## 4. 基于 T4 GPU 的预测速度
add T4 benchmark
2020-05-09 16:46:11 +08:00
fix bs
2020-05-09 17:01:18 +08:00
| Models | Crop Size | Resize Short Size | FP16<br>Batch Size=1<br>(ms) | FP16<br>Batch Size=4<br>(ms) | FP16<br>Batch Size=8<br>(ms) | FP32<br>Batch Size=1<br>(ms) | FP32<br>Batch Size=4<br>(ms) | FP32<br>Batch Size=8<br>(ms) |
add T4 benchmark
2020-05-09 16:46:11 +08:00
|-------------|-----------|-------------------|------------------------------|------------------------------|------------------------------|------------------------------|------------------------------|------------------------------|
| HRNet_W18_C | 224 | 256 | 6.79093 | 11.50986 | 17.67244 | 7.40636 | 13.29752 | 23.33445 |
polish dygraph doc and deploy
2020-10-19 13:05:34 +08:00
| HRNet_W18_C_ssld | 224 | 256 | 6.79093 | 11.50986 | 17.67244 | 7.40636 | 13.29752 | 23.33445 |
add T4 benchmark
2020-05-09 16:46:11 +08:00
| HRNet_W30_C | 224 | 256 | 8.98077 | 14.08082 | 21.23527 | 9.57594 | 17.35485 | 32.6933 |
| HRNet_W32_C | 224 | 256 | 8.82415 | 14.21462 | 21.19804 | 9.49807 | 17.72921 | 32.96305 |
| HRNet_W40_C | 224 | 256 | 11.4229 | 19.1595 | 30.47984 | 12.12202 | 25.68184 | 48.90623 |
| HRNet_W44_C | 224 | 256 | 12.25778 | 22.75456 | 32.61275 | 13.19858 | 32.25202 | 59.09871 |
| HRNet_W48_C | 224 | 256 | 12.65015 | 23.12886 | 33.37859 | 13.70761 | 34.43572 | 63.01219 |
polish dygraph doc and deploy
2020-10-19 13:05:34 +08:00
| HRNet_W48_C_ssld | 224 | 256 | 12.65015 | 23.12886 | 33.37859 | 13.70761 | 34.43572 | 63.01219 |
add T4 benchmark
2020-05-09 16:46:11 +08:00
| HRNet_W64_C | 224 | 256 | 15.10428 | 27.68901 | 40.4198 | 17.57527 | 47.9533 | 97.11228 |
Add se_hrnet_w64_c_ssld_pretrained (#452) add se hrnet64_c
2020-12-06 17:52:34 +08:00
| SE_HRNet_W64_C_ssld | 224 | 256 | 32.33651 | 69.31189 | 116.07245 | 31.69770 | 94.99546 | 174.45766 |