Lingxi Xie

**Solution: looking into the `info` variable and checking the existence of `info[...]['episode']` -- if it exists, then the entire game has ended and we can output `info[...]['episode']['r']` as the reward.**...

@JarveeLee Our CenterNet and another one are similar but different ideas. Their CenterNet assumed that each object can be represented as a point and an anchor needs to be put...

> thank you for the code, i have a problom, how can i control the flops? for example, i want a model with input size 224*224, flops 60~80M, can you...

> 您可以看看这个全流程demo，https://github.com/HaxyMoly/Pangu-Weather-ReadyToGo 感谢答复，不过这个链接好像点不开。

按理说1060（6GB显存）是跑不通的，会不会是CPU在工作？结果正确吗？跑一次需要多久（如果除去载入时间接近1分钟，那就可能是CPU在运行）？

Hi @drcdr , thank you for your interests in our work and so many good questions. I will try to answer a few of them and Xin will later put...

Thanks for this question! I think multi-GPU works just like single-GPU. Since our search on CIFAR takes a few hours, we did not consider multi-GPU training. However, during our recent...

No, only data format conversion is required. The ONNX file contains both normalization and de-normalization, which is different from the pseudocode (in the pseudocode, the Model() functions only contain normalization...

Hi, To train good weather forecasting models, both the global data (rather than a region of 15 degs x 20 degs or so) is needed. This is because the future...

Hi, the thickness and vorticity can be calculated by calling the Python library `metpy`. You can refer to the `calc` and `units` parts in `metpy`. For radius, you can import...