Gpu taking longer to train


I dont know why my gpu taking 4 m longer sometimes over 6 min to train the model. while in the notebook provided and in the video jeremy took only 20 s to train. Is it this normal or what should i do to improve my gpu

my config is Nvidia gtx 1650 ti
cpu : amd ryzen 4600h

i have checked my gpu while the training is happening. the indeed is using the gpu and also checked twice using torch.cuda.is_available() and ```
torch.cuda.get_device_name(0) it is givinig me True and NVIdia gtx 1650 ti .

Any help would be appreciated

Are you on Linux or Windows WSL?

windows wsl , i dont think im using any wsl at present. I happened to use it for hugging face api in the video lecture 2 for deploying ml model.