Benchmark DALI decoding for pytorch and tensorflow

I totally missed that [DALI added audio decoding ops](https://docs.nvidia.com/deeplearning/dali/user-guide/docs/supported_ops.html#nvidia.dali.ops.AudioDecoder). As far as I can see, the ops seems to be based on sndfile and do not support GPU decoding. I wonder if this would still improve single threaded ETL speed on pytorch or tensorflow.

pinging DALI devs @mzient @szalpal as they might have benchmarked the op against [tfio](https://www.tensorflow.org/io/api_docs/python/tfio/audio/decode_wav) or [torchaudio](https://pytorch.org/audio/) which are also sndfile based.