Remove nvidia and dask channels

[vyasr]

Now that we have dropped support for CUDA 11 we no longer require the nvidia channel.
With the changes in rapidsai/rapids-dask-dependency#85, RAPIDS now only uses released versions of dask, so we no longer need the dask channel either.
Contributes to rapidsai/build-planning#184

[jakirkham]

Thanks Vyas! 🙏

Do we want to update this doc as well?

cugraph/python/cugraph-service/README.md

Line 13 in d49fb01

conda install -c rapidsai-nightly -c rapidsai -c conda-forge -c nvidia cugraph-service-server

Also see -c nvidia in this Dockerfile. Though don't have a good handle on how this is being used. So think it would be ok to punt and raise an issue for follow up

cugraph/benchmarks/shared/build_cugraph_ucx/cugraph_ucx.dockerfile

Lines 49 to 55 in d49fb01

	RUN gpuci_mamba_retry install -y -c pytorch -c rapidsai-nightly -c rapidsai -c conda-forge -c nvidia \
	cudatoolkit=$CUDATOOLKIT_VER \
	cugraph=$RAPIDS_VER \
	pytorch=$PYTORCH_VER \
	python=$PYTHON_VER \
	setuptools \
	tqdm

[vyasr]

Yeah let's save that for a follow-up where we do docs more thoroughly. I did it in the cuxfilter PR but in hindsight it would be better to keep the PRs a bit more focused so that we can do similar updates to all repos across RAPIDS simultaneously in sets of PRs.

[jakirkham]

/merge