Albert Wolant

Hello @kvablack thank you for reporting this issue. Do you have some standalone script to reproduce it reliably. It would help a lot as I had no luck in reproducing...

Hello @kvablack thank you for creating the issue. The idea looks very interesting. Could you share some of your performance results? Also, could you tell us more about the pipeline...

Hell @jadechoghari thanks for the question. I think you are looking for `enable_timestamps` argument of [our video reader](https://docs.nvidia.com/deeplearning/dali/user-guide/docs/operations/nvidia.dali.fn.readers.video.html#nvidia-dali-fn-readers-video) It works together with `file_list` argument. You can create a list and...

Sure, I understand now. I am afraid that currently there is now way to get arbitrary frames by timestamp. I will mark this issue as a possible enhancement to the...

Hi, thanks for the question. Thanks to the contribution from IBM we have a [conda recipe](https://github.com/NVIDIA/DALI/tree/master/conda) ready. Conda packages are maintained and distributed by IBM. For details look [here](https://docs.nvidia.com/deeplearning/sdk/dali-developer-guide/docs/installation.html#pre-built-packages-in-watson-machine-learing-community-edition).

Hello @quanvuong thanks for reporting the issue. Could you tell more about your setup? It looks like you are using multiple GPUs here? What are the other parameters? What are...

Hello @sali1997s Thanks for the question. Currently, something like this not supported unfortunately. This enhancement is in our TODO list for JAX integration. Could you tell more about your use...

Thanks for the reproduction. This is definitely a feature that could be added to DALI JAX support to enhance in functionality. For the first version of this integrating layer we...

Hello @treasan thanks for creating the issue. To better understand the requirement I wanted to ask do your use case expect the samples to have the same number of frames...