Skip to content
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

Did fastfold support bfloat16 inference? #165

Open
semal opened this issue Mar 24, 2023 · 5 comments
Open

Did fastfold support bfloat16 inference? #165

semal opened this issue Mar 24, 2023 · 5 comments
Labels
enhancement New feature or request

Comments

@semal
Copy link

semal commented Mar 24, 2023

image

fastfold/utils/geometry/vector.py seems got an unexpected data type.
@Gy-Lu
Copy link
Contributor

Gy-Lu commented Mar 25, 2023

Thanks for your question, this feature is still under development.

@Gy-Lu Gy-Lu added the enhancement New feature or request label Mar 25, 2023
@semal
Copy link
Author

semal commented Mar 27, 2023

image

more information about the data type change.

@semal
Copy link
Author

semal commented Mar 27, 2023

This error comes from xx = 1 - 2* (y ** 2 + z** 2), dtype of 1 should give a specific datatype. Another error comes from w *= inv_normwhich is a inplace operation, maybe should change to w = w * inv_norm

@Ian-pku
Copy link

Ian-pku commented Dec 10, 2023

hi! have you achieved inference with bf16? which part of the source code did you modified?

@Lix1993
Copy link

Lix1993 commented Jul 11, 2024

Readme says: We can inference a sequence of length 10000 in bf16 with 61GB memory on a Nvidia A100(80GB).

How can I infer with bf16?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
Development

No branches or pull requests

4 participants