Can I install NVTabular from source? [QST]
What is your question? I've installed NVTabular for 3 times today, and each time I encountered a different bug. It's quite torturing.
The latest error was the closest to success, however it reported the following error.

It seems like a version issue of dependency cudf, and I also met other different dependencies issues in the previous failed installations. However, I cannot control the dependencies verion since I installed NVTabular from conda. So, can I install NVTabular from source?
I've tried the official merlin-pytorch image, and met another bug. Please see: https://github.com/rapidsai/dask-cuda/issues/974
@zxgx sorry for the inconvenience. Are you trying to run this notebook: https://github.com/NVIDIA-Merlin/NVTabular/blob/main/examples/scaling-criteo/01-Download-Convert.ipynb ?
@zxgx can you please give more information regarding your env? where do you want to install NVTabular? in a conda env? are you running notebooks in a cloud instance? what's your HW? can you pls share your driver version, nvidia-smi output, and nvcc --version output. thanks.
@zxgx do you still have a question?