Georgy Ayzel
Georgy Ayzel
Dear @giszters, I suppose that you are right and it could be related to preprocessing routines that are pretty complex, i.e., scaling, taking log, etc. Good luck!
Hi @IncubatorShokuhou, thank you for your interest in `rainymotion`! `rainymotion` models require inputs in the form of 3D numpy array (frames, dim_x, dim_y) of radar data for previous timesteps where...
Hi @RubenImhoff, Thanks for pointing that out. I will definitely do that in the next release (sad frog). At the moment, `rainymotion` explicitly requires OpenCV of version 3.4.* (in .yml...
Hi @DBIGSHUAI, you could access hdf5 data as follows: ```python import h5py data = h5py.File("RYDL.hdf5", mode="r") ``` After that, you could use the verification example provided in the rainymotion library...
Hi Ruben, Could you, please, send me an email with your code and data? ayzel AT uni-potsdam.de I will try to reproduce the problem. At first glance, the problem is...
By the way, we have substantially improved the Dense group of 'rainymotion' models -- at the moment they are both efficient and fast. You can try them too. And thank...
Hi @DBIGSHUAI, Could you describe what you mean in detail? For example, you could visualize model structure using standard built-in keras functionality, i.e. https://keras.io/api/utils/model_plotting_utils/. There are also independent libraries such...
Hi @DBIGSHUAI, I am sorry, but I did not get your question. You do have precipitation scans from your radar, right? If it is outside Germany (that I guess is...