Uploading files directly from the local file system or from google drive in google colab is faster in the training model?
Hi! This probably will not answer your question completely. If the dataset is small I would move it to my local file system in colab that is ‘/content/’, this is because in colab there is a cap on number of times you can load the data from drive to bypass it it is recommended to put the data in local file as there is no cap on number times you load the data from local system.
1 Like
Thanks for your reply
I have another question, please
Run time in google colab during the training process give me a Busy run time
What should I do! (i purchased a pro version of colab)
Well I am not an expert on this if given it was a free version I would have said maybe you exhausted your daily quota for GPU but given the fact you are on a pro version I am not sure of if that would be the reason, from the screenshot it looks the model is still training?
One thing you can do is to check the number of active sessions and kill the sessions which are not needed. In the free version you can use two sessions simultaneously. Hope it helps.