Unable to get running on cuda and windows 11

[clearsitedesigns]

So first and foremost, I'm very familiar with this script. I have everything setup and running for Cuda, tested it and validated cudas running. This works on my mac m1 no issue. I tried to get it running today on my PC with a 4070. I am just running sample orca ingestion, and I let run over night knowing it shouldn't take that long, but it never finished. It created the tokenizing and chunking and then never actually ingested.

is there something up with the ingestion process that might be a bug in the current version? I would have expected ingestion to take around 30 mins or less. With 40k chunks, based on my previous experience. I wonder if there is an issue in the way langchain ingests documents through its document loader.

Any ideas what to look for?