New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
how to generate KeyedJaggedTensor on GPU #1715
Comments
what do you mean by "compress"? They should be torch.tensors right? |
sorry, I have keys as a string list, values as a list tensor, lengths as a list tensor. |
@xiexbing I think you can just first move the tensors to GPU, and then initialize a kjt that way? if I remember correctly, that should work just fine? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hello,
I compressed the values, and lengths to reduce the PCIe cost. then I want to move the keys, compressed values and lengths to GPU, and then decompress and generate KJT on GPUs.
My question is: any examples on how to do this?
thanks.
The text was updated successfully, but these errors were encountered: