-
1. Please check DbSchema Help / Output logs for errors.
2. Please include the DbSchema version, operating system and used database
8.2.7 build 200205
3. The steps to reproduce this issue
Use thi…
-
Thanks for sharing it. I run your code with pfld ported from tensorflow code and the landmark results in unstable although the face is not moving. Could you suggest the solution/method to fix it? your…
-
hi
可以讲一下怎么训练的landmark吗?采用什么算法和什么数据?
-
感谢大佬的分享!有个问题想不明白:
load_pfld_data.py中第36行,header = mx.recordio.IRHeader(0, label, i, 0),第三个参数 i 是怎么来的?
-
Hello Team,
Thanks a lot for the great work, but how to detect pose on android ?
I really need your help on that part
Thanks in advance
-
识别过程中,能否获取到辅助参数,例如姿态角、脸部置信度这些信息呢?
-
找不到‘./data/test_data/list.txt’文件
-
original paper:
> For instance,if disabling the geometry and data imbalance functionalities, our loss degenerates to a simple`2loss
your code:
```
mat_ratio = torch.mean(attribut…
-
在test_model.py 中,我找到如下代码
```
landmark_L1 = graph.get_tensor_by_name('landmark_L1:0')
landmark_L2 = graph.get_tensor_by_name('landmark_L2:0')
landmark_L3 = graph.get_tensor_by_name('landmark_L3:0')…
-
**line 298:**
#14 * 14 * 128
conv5_1 = slim.convolution2d(conv4_1, 512, [1, 1], stride=2, activation_fn=None,scope='conv5_1/expand')
**line 362:**
#14 * 14 * 128
conv6_…