Ivo Stranic

@satti007 Got it. Thank you! Is your main objective to store this data in hub format, or would you also want our App to support visualization of the data?

@dhiganthrao That's a part of it. We'd also need to build the code that returns the data as arrays, list or arrays, etc. (details TBD) for cases of single images...

Hi @St3V0Bay. Thx for following up on this thread! Adding custom compression is quite tricky, because even if it's implemented in Deep Lake OSS, it won't work in our visualizer...

Hi @FeryET. Thank you for checking in on this. We plan to release a conda distribution by December 24.

Hi @kordc. Unfortunately we were not able to get around to building a conda distribution. Is sounds like this is important for you to be able to use hub?

@kordc Sorry about not getting to this sooner. We'll ping you as soon as the conda distribution is ready.

@Manas2030

Hey @Manas2030 Just making sure that you're still working on this? How is progress so far? Do you need any assistance?

Hi @jaivanti The PyTorch components are complete, but the TensorFlow ones are not. Are you interested in those?

Hi @jaivanti Great. I think the best approach would be to implement the tutorial below using hub's TensorFlow integration. https://docs.activeloop.ai/hub-tutorials/training-an-object-detection-and-segmentation-model-in-pytorch Let me know if you have any questions.