rnyak

I get same error `RuntimeError: cuDF failure at: ....: Cannot open output file` when the subdirectory doesn't exist but the error msg is not informative.. it is hard to tell...

@EvenOldridge @viswa-nvidia what Benedikt refers here that in the scaling-criteo [03-Training-with-TF.ipynb](https://github.com/NVIDIA-Merlin/NVTabular/blob/main/examples/scaling-criteo/03-Training-with-TF.ipynb) notebook, we define numerical layers for numerical features, but they are never fed in the model architecture, therefore it...

@liguo88 which NVT pipeline are you running? is it from Transformers4Rec repo examples? can you please provide a simple reproducible example with a small synthetic dataset you can generate?

@jershi425 when I train the model in the `Training-with-HugeCTR` nb, I get the following during model training: ``` [22d23h55m60s][HUGECTR][INFO]: Iter: 1900 Time(100 iters): 0.262071s Loss: 0.523336 lr:0.001000 [22d23h55m60s][HUGECTR][INFO]: Rank0: Dump...

@jershi425 can you also explain these lines? why `range(1, 31)` ? where is `31` coming from? thanks. ``` rp_lst = [0] cur = 0 for i in range(1, 31): if...

> @rnyak The `set_mempolicy: Operation not permitted` issue can be solved after you add `--privileged` when you start your docker container. The range(1, 31) comes from the fact that we...

> @rnyak @jershi425 what is the plan for this PR? @albert17 we already have [multi-hot training example](https://github.com/NVIDIA-Merlin/NVTabular/blob/main/examples/getting-started-movielens/03-Training-with-HugeCTR.ipynbhttps://github.com/NVIDIA-Merlin/NVTabular/blob/main/examples/getting-started-movielens/03-Training-with-HugeCTR.ipynb) with HugeCTR based on this PR. It should work without issues. What is...

@rjzamora will test it and let you know. Thanks.

this might help for implementation: https://developer.nvidia.com/blog/gauss-rank-transformation-is-100x-faster-with-rapids-and-cupy/

@radekosmulski I tested the following it works for me. I am closing this bug, you can reopen if you need to. I am using `merlin-pytorch:22.07` container. ``` products = cudf.DataFrame(...