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.

Sign up for GitHub

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

Jump to bottom

Investigate low accuracy in unet-brain #762

Open
ayerofieiev-tt opened this issue Feb 12, 2025 · 4 comments
Open

Investigate low accuracy in unet-brain #762

ayerofieiev-tt opened this issue Feb 12, 2025 · 4 comments
Assignees
@amalbasaTT
Labels
accuracy

Comments

@ayerofieiev-tt
Copy link
Member

No description provided.

@ayerofieiev-tt
Copy link
Member Author 

      1.  Fatal error at ttnn_prefix_conv_17 = conv_wrapper(ttnn_to_device_104, ttnn_from_torch_77, None, 1, 128, 64, [128, 128], [3, 3], [1, 1], [1, 1], [1, 1], 1, device, False, None, ):
	- Always | FATAL    | Statically allocated circular buffers in program 247 clash with L1 buffers on core range [(x=0,y=0) - (x=7,y=7)]. L1 buffer allocated at 285312 and static circular buffer region ends at 296160
	- Input ttnn_to_device_104 spec:
			- [1, 1, 16384, 128]
			- dtype: DataType::BFLOAT16
			- layout: Layout::TILE
	- Input ttnn_from_torch_77 spec:
			- [64, 128, 3, 3]
			- dtype: DataType::BFLOAT16
			- layout: Layout::ROW_MAJOR

@ayerofieiev-tt
Copy link
Member Author

@amalbasaTT can you fire a separate issue for this Out of memory issue for the specific Conv input variation?

@ayerofieiev-tt
Copy link
Member Author

CC @pavlepopovic (to keep in the loop early)

@amalbasaTT
Copy link

tenstorrent/tt-metal#17869

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@amalbasaTT amalbasaTT

Labels
accuracy
Projects
PyTorch 2.0 TT-NN Compiler
Status: In Progress
Milestone
No milestone
Development

No branches or pull requests
2 participants
@ayerofieiev-tt @amalbasaTT