-
Notifications
You must be signed in to change notification settings - Fork 16
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
Removed create_golden_Tensor, created "safe" custom Constructor #2136
Conversation
@vroubtsovTT I tried to make the vector a |
That may have been a sign that you were on the right track. Somebody wanted to copy your struct and couldn't because std::unique_ptr member fields aren't copyable -- this failed as it should. But a std::unique_ptr member will be move-copyable... you should try to make your struct move-copyable by adding
(and/or implementing such a move constructor if |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Explorer changes LGTM
Closes #1961 Closes #1759 ### Problem description `transpose` wasn't exposed to `ttir_builder`, preventing llama attention layer modelling ### Special Thanks To @vprajapati-tt for inadvertently fixing a blocking memory issue that was exposed by this bug (See #2136 )
- PR closes #2027 - Changes container for `GoldenTensor` in Python to be a shared_ptr, ensuring data is maintained and not copied. - Uses `GoldenTensor` custom constructor instead of `create_golden_tensor`. - Creates copy of supplied data_ptr (size of data must also be provided, example in `ttir_builder.py`) and stores within GoldenTensor object.
Closes #1961 Closes #1759 ### Problem description `transpose` wasn't exposed to `ttir_builder`, preventing llama attention layer modelling ### Special Thanks To @vprajapati-tt for inadvertently fixing a blocking memory issue that was exposed by this bug (See #2136 )
GoldenTensor
in Python to be a shared_ptr, ensuring data is maintained and not copied.GoldenTensor
custom constructor instead ofcreate_golden_tensor
.ttir_builder.py
) and stores within GoldenTensor object.