Hi! I’ve found the perfect version of the Imagenet dataset on huggingface, which is something I’m immensely grateful for and very excited about. However, every time I try to load it in a docker jupyter space running on A1OG with 1TB of persistent storage, the kernel quits unexpectedly at any point in the loading. It quit at 80% two times so far, which also makes me worried about the costs I’m incurring with absolutely nothing to show for it. Could I get some advice on what to do? This is the dataset (timm/imagenet-22k-wds · Datasets at Hugging Face). Can I even run this on hugging face? If not, could you recommend some other reliable virtual computing service?
Any advice or help is very much appreciated.
Thanks!
Tian
