How to upload a custom dataset to a contrasitive learning SIMCLR implementation

Link: simclr/ at main · larsh0103/simclr · GitHub

Hello, I implemented the github code (given in Link) and worked well. I want to implement the same code using my timeseries dataset present on my desktop. I am unable to stack the samples and read through dataloader . Please help

Can anyone please help in this.

You would need to explain the issue a bit more, what you’ve tried so far, where you are currently stuck etc. Right now the question sounds more like a request for someone to implement a fully working solution for you.

Looking at the link, I have given in my question, {simclr/ at main · larsh0103/simclr · GitHub} I have tried to upload my dataset by stacking 2 tensors in the “get_item” method as is done in the link( for cifar10 dataset. But when iterating through dataloader it gives me the error that i have described in “ValueError: not enough values to unpack (expected 2, got 1)”. Now able to understand how should i return my 2 stacked samples (with data augmentations done) from get_item method so that it is readable by training dataloader.

Are you still running into the same issue from the other topic?
I’ve followed up with a code snippet, which works fine and which you could use to compare your code to mine to narrow down where the difference (and thus error) is coming from.

Corrected. Thanks a lot for the help.