Hi @bnnadi123 thanks for reaching out! We used 4 gpus in parallel and jobs were usually started in the evening, ran overnight and were done in the morning. I can't tell you exactly how long they took (don't remember) but the cluster had a maximum run time of 24h so for sure less than that.
Originally posted by @dsenti in #11
@Thoriri
Hello Thoriri,
I tried tagging @dsenti for additional input on this. I also noticed that the paper has been accepted to IEEE EMBC.
I wanted to ask if there is still a possibility of the model weights being open-sourced. I have attempted to train the model from scratch using the same architecture and parameters, but I am encountering issues with overfitting and convergence. As a result, I have not been able to successfully reproduce the results reported in the paper.
I would greatly appreciate any guidance or updates you can provide.
Originally posted by @dsenti in #11
@Thoriri
Hello Thoriri,
I tried tagging @dsenti for additional input on this. I also noticed that the paper has been accepted to IEEE EMBC.
I wanted to ask if there is still a possibility of the model weights being open-sourced. I have attempted to train the model from scratch using the same architecture and parameters, but I am encountering issues with overfitting and convergence. As a result, I have not been able to successfully reproduce the results reported in the paper.
I would greatly appreciate any guidance or updates you can provide.