Open R4Ajeti opened 3 months ago
Hi, sounds like a great project. Can you please privately send us the email you tried to use here so we can investigate if there is indeed an issue? https://pages.awscloud.com/GLOBAL_PM_PA_amazon-sagemaker_20211116_7014z000000rjq2-registration.html
thanks
Hi, @MicheleMonclova
You have my email on this GitHub account is public, feel free to check it and investigate what is the issue about. I don't share anything to noone privately this is my only account for this type of experiments. Personally I love open source since I am not rich, and a lot of big companies support them. And since soon code is gonna be open source from automated labels to manual checking policy, so do the final model will be.
PS: I fixed the problem with kaggle GPU usage, the only problem was you need to verify your number to be able to use it. I just accidentally while surfing google I saw suggesting it and worked for me if anyone has same issue feel free to try it.
Happy coding, 💻
Describe the bug Here is the bug description: I am trying to train a model image to text recognition specifically in my own language (using manual label checking). On the first version, my dataset was relatively small so I trained on colab too many times with diff settings without any issue, and everything went very well but everything. I was so excited so now when I expanded my dataset to 46x, on Colab using free minutes it has gone more than a week and a half and finally half of the entire epochs are done. I searched Google and youtube for alternatives I found Kaggle but for some reason somehow I cannot see a GPU usage somewhere not sure if they have temporarily disabled GPU or what is happening there UI. Next alternative was aws sagemaker-studio lab that youtube guy suggested. I applied throw this link: https://studiolab.sagemaker.aws/requestAccount, after some minutes I got the response of ToS violation something like that and this Is the first time ever I see or clicked that link, a little bit confusing but who knows it is what it is.
I MUST NEED A POWERFUL GPU: because I won't stop after this project I am gonna try to move on text-to-text translation in my language using again manually check translation. Or can someone from the community suggest any working alternative because only alternative left to me is lightning ai where I applied but it says I am on the waiting list.
Happy coding, :)
To Reproduce Steps to reproduce the behavior:
Expected behavior A clear and concise description of what you expected to happen.
Screenshots
Desktop (please complete the following information):
Additional context Add any other context about the problem here.