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

Fix ttnn.from_torch for 0D/1D tensors with tile layout #16484

Merged
merged 7 commits into from
Jan 10, 2025

Conversation

sminakov-tt
Copy link
Contributor

@sminakov-tt sminakov-tt commented Jan 7, 2025

Ticket

Problem description

Part 3 of the original PR for ttnn.from_torch support for 0D/1D tensors with tile layout which got reverted previously

What's changed

Variety of Shape fixes throughout the codebase

Checklist

ayerofieiev-tt pushed a commit that referenced this pull request Jan 9, 2025
### Ticket

### Problem description
There is a double move occurring inside invoke_composite / launch_op.
The issue is caused by passing a mutable lambda, which changes the
argument into launch_op, which in turn calls the lambda multiple times.
This change is a prerequisite for
#16484

### What's changed
Remove `mutable` modifier for the passed lambda, removed incorrect
usages of `std::forward`.

### Checklist
- [x] [Post commit CI
passes](https://github.com/tenstorrent/tt-metal/actions/runs/12685459880)
- [x] [T3K frequent CI
passes](https://github.com/tenstorrent/tt-metal/actions/runs/12685464588)
- [x] [T3K unit tests CI
passes](https://github.com/tenstorrent/tt-metal/actions/runs/12687593651)
- [x] [Model regression CI testing
passes](https://github.com/tenstorrent/tt-metal/actions/runs/12685476479)
- [x] [Device performance regression CI testing
passes](https://github.com/tenstorrent/tt-metal/actions/runs/12685472338)
- [x] New/Existing tests provide coverage for changes
@ayerofieiev-tt ayerofieiev-tt merged commit b2912fe into main Jan 10, 2025
226 checks passed
@ayerofieiev-tt ayerofieiev-tt deleted the sminakov/from_torch-p3 branch January 10, 2025 09:04
sminakov-tt added a commit that referenced this pull request Jan 11, 2025
### Ticket

### Problem description
Nightly stable diffusion tests are currently failing, after
#16484

### What's changed
Fixed calculation of `output_tensor_end` in one of the branches of
layout conversion

### Checklist
- [x] [Post commit CI
passes](https://github.com/tenstorrent/tt-metal/actions/runs/12718154798)
- [x] [Device performance regression CI testing
passes](https://github.com/tenstorrent/tt-metal/actions/runs/12718161215)
- [x] [Nightly model and ttnn
tests](https://github.com/tenstorrent/tt-metal/actions/runs/12718140006)
- [x] New/Existing tests provide coverage for changes
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants