Peng Lu

您好，您说的这种情况可能是模型预测关键点坐标出错或者 plot 可视化的时候坐标轴设置错误。top_down_img_demo_with_mmdet.py 得到的 pose_results 里面的关键点坐标值一般都是大于 0 的。您可以先检查一下推理得到的关键点坐标值数值范围。如果正常的话，可以再检查下 plot 的时候坐标轴的方向

Hi, thanks for testing the code on your images. Your task is a style-to-style transfer and the weight should be set to 3 according to [official code](https://github.com/msracver/Deep-Image-Analogy). Also, the resize_ratio...

Sorry I am busy these days so I cannot explain in detail. The two main differences are: 1. Deconvolution process: I follow the guide of this [issue](https://github.com/msracver/Deep-Image-Analogy/issues/30) and deconvolute the...

Yes. I ran the code on VM of GCP which has a 4-core CPU and a K80, and it takes around 30 mins. The patchmatch process on low level features...

The unittest for the missed lines, indeed the unittest for `webcam_executor.run`, involves multi-threading. It will cause `segmentation fault` in `pr_stage_test / build_cpu (3.7)` and fail while pass in other environments....

Hi, we retrain the CID model and get better performance. | Arch | Input Size | AP | AP50 | AP75 | AR | AR50 | | :-------------------------------------------- | :--------:...

It seems like a numpy issue. You can try to upgrade your numpy. For more discussions, please refer to https://github.com/open-mmlab/mmpose/issues/534.

这个 bug 可能是因为 FC 层 input, weight 维度不一致导致的。 MobileNetV3 `arch='small'` 的时候，输出的 feature map 维度数是96，`keypoint_heat.in_channels` 可以改成96试一下。 或者可以尝试 `arch='big'`，对应的 backbone 输出的 feature map 维度数是 160

> `from __future__ import annotations` may have compatibility issues on python 3.6. Please consider using a string like 'BufferManager' for type hints within the class itself. Solved. The corresponding unittest...

Hi there, I train the LitePose-S on coco with config `configs/body/2d_kpt_sview_rgb_img/associative_embedding/coco/litepose_S_coco_448x448.py` and pretrained model converted from `LitePose/0927/pretrained/coco-pretrained-S-mmpose.pth` using the script provided by @jin-s13, but got results slightly lower than reported...