Upcoming 10'000 samples training dataset

#116
by viserjor - opened
Image and Signal Processing • ISP org

Dead participants
The competition is reaching to an end (June 30th) and there's a tight competition in the top three positions. To allow testing more complex machine learning model architectures, I'm currently running a new training dataset with 10'000 samples for the scenario A that you can combine with the two existing ones to make a dataset with 12'500 samples. The simulations are taking some time but I expect having them ready by the 23rd or 24th of June . Keep an eye on the huggingface dataset portal (https://huggingface.co/datasets/isp-uv-es/rtm_emulation/tree/main/scenarioA/train)

Image and Signal Processing • ISP org

Dear participants
You can already find the 10'000 training samples for the scenario A in the following path: https://huggingface.co/datasets/isp-uv-es/rtm_emulation/tree/main/scenarioA/train
You can use this dataset, alone or in combination with the current datasets, as you wish:

  • training more complex deep-NN architectures
  • valitation/testing
  • active learning

Sign up or log in to comment