More GPU for processing data

Someone told me this :

Detection of covid 19 using chest X-ray images using graph convolution networks
Dataset-33500 covid chest X-ray images
Training and testing accuracy calculated in 16gb MacBook Pro with Gpu which has taken around 8-10hrs to get it executed for 100 n 300 epochs
Now same code has to run for 500 epochs so looking for high performance machine so that execution can be finished in 1 hr

Using Tensorflow n keras.
So to train and test accuracy it’s taking lot of time.
If she gets a High performance machine may be execution time will be less ?

What other options are available ? Online GPUs ? colab ?

Hi @anjanesh, Training time depends on the various parameters like the size of the Dataset used for training, Batch_size, Model complexity etc…

High-performance machines can significantly reduce the training time. Thank You.

So should using something like paperspace ?

Hi @anjanesh, Yes, you can use any cloud platform, but if you have access to higher end GPU support you can decrease the training time. Thank You.