Open inoue0426 opened 2 years ago
Using rcellminer data with original feature data, we can run DrugCell on Binder. https://mybinder.org/v2/gh/cannin/graph_neural_network_drug_response/6b109d5c31c760237d51dfb4bd5cbd1a44db5aae Definitely our data is huge and it's impossible to run on Binder. So if you want to use huge data, you must use your own machine.
Maybe switch to Google Colab?
When running DrugCell on the binder, will happen kernel restarting. This seems like OOM or something happens. DrugCell uses PyTorch and this is a huge library, so it can happen. Not sure how to resolve rn, but need to check what happens and then solve this.