-
1. How to get pre-trained model on mpii dataset and fine-tune it on my own dataset?
-
Hello @terasakisatoshi
Thank you for your great work .
Your demo look great. But is there any difference with original pose proposal network paper ?
If you could add the evaluation result of MPI…
-
I'm using the pytorch branch, and I'd like to fine-tune your pre-trained models on my data, but there appears to be an issue:
```
python train.py --loadModel ../../models/sppe/duc_se.pth --dataset…
hoxi3 updated
5 years ago
-
The prediction of YOLOv3 usually a rectangle and I see that you use transforms method to resize the image before put it into HRNet. It stretching the image and make pose estimation network performs ba…
-
**Your Operating system and DeepLabCut version**
Windows 10 with an anaconda env
Please complete the following information about your system:
OS: Windows 10
DeepLabCut Version: 2.0.7
Browse…
-
Attempts at loading a model from the prototxt file, caffemodel file, and proto file have failed in a manner very similar to [Issue #22](https://github.com/chanyn/3Dpose_ssl/issues/22)
## **Expected…
-
@bearpaw thank you for your wonderful work,
I have trained my network. While, when I validated my best model in different test batch size, the validation acc suffered from different results, as I th…
-
可否更新一下,谢谢
-
In README it said that when choosing training hyper-params, the default values are the ones used to train the models in the paper.
However I found there are some differences between code default se…
-
When the model on MPII has been trained by one GPU, I want to use test.py to evaluation the result. The error about some pkl files occurs while I didn't find any such files.
![Screenshot from 2019-…