Replace List Mesh to Tensor #17667
Open
Replace List Mesh to Tensor #17667
+189
−160
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jjiangTT
requested review from
cglagovichTT,
uaydonat,
johanna-rock-tt,
djordje-tt,
kpaigwar,
cfjchu,
ayerofieiev-tt,
dmakoviichuk-tt,
omilyutin-tt,
rfurko-tt,
TT-BrianLiu,
razorback3,
dongjin-na and
bbradelTT
as code owners
jjiangTT
force-pushed
the
jjiang/rm-list_mesh_to_tensor
branch
from
b4bdc28 to
1edac8f
Compare
omilyutin-tt
requested changes
Feb 6, 2025
djordje-tt
approved these changes
Feb 6, 2025
jjiangTT
force-pushed
the
jjiang/rm-list_mesh_to_tensor
branch
2 times, most recently
from
ba03eea to
9d1c27c
Compare
johanna-rock-tt
approved these changes
Feb 7, 2025
jjiangTT
force-pushed
the
jjiang/rm-list_mesh_to_tensor
branch
4 times, most recently
from
1a2d6ec to
3476824
Compare
jjiangTT
force-pushed
the
jjiang/rm-list_mesh_to_tensor
branch
3 times, most recently
from
2250279 to
8474bfa
Compare
bbradelTT
approved these changes
Feb 14, 2025
jjiangTT
force-pushed
the
jjiang/rm-list_mesh_to_tensor
branch
from
fc27d2c to
9dc5bb4
Compare
omilyutin-tt
approved these changes
Feb 14, 2025
|
The PR looks good! Make sure to kick off post commits and T3K unit tests to double check nothing is broken. |
jjiangTT
force-pushed
the
jjiang/rm-list_mesh_to_tensor
branch
3 times, most recently
from
7f12318 to
58b94cc
Compare
…types forward, fix bfloat16 type error
…or ttnn.multidevice, fix python hook name recursion error
module.def(
"get_mesh_device_core_grid",
[](MeshDevice& mesh_device) {
CoreCoord coords = mesh_device.compute_with_storage_grid_size();
new CoreGrid(coords.x, coords.y);
},
py::arg("mesh_device"));
…tnn/operations/core.py
…iner array generator
jjiangTT
force-pushed
the
jjiang/rm-list_mesh_to_tensor
branch
from
58b94cc to
1721aec
Compare
cfjchu
requested changes
Feb 14, 2025
| @@ -9,8 +9,10 @@ | |||
| import pathlib | |||
| import re | |||
| from types import ModuleType | |||
| from typing import List | |||
There was a problem hiding this comment.
Please remove all added lines here.
Comment on lines
+7
to
+9
| #include <pybind11/operators.h> | ||
| #include <pybind11/pybind11.h> | ||
| #include <pybind11/stl.h> |
There was a problem hiding this comment.
You shouldn't need these here. Please remove.
Comment on lines
+6
to
+29
| #include <pybind11/pytypes.h> | ||
| #include <cstdint> | ||
| #include <memory> | ||
| #include <utility> | ||
|
|
||
| #include "tt-metalium/assert.hpp" | ||
| #include "tt-metalium/bfloat16.hpp" | ||
| #include "tt-metalium/buffer.hpp" | ||
| #include "tt-metalium/core_coord.hpp" | ||
| #include "tt-metalium/overloaded.hpp" | ||
| #include "tt-metalium/core_coord.hpp" | ||
| #include "ttnn/common/constants.hpp" | ||
| #include "ttnn/distributed/api.hpp" | ||
| #include "ttnn/tensor/host_buffer/borrowed_buffer.hpp" | ||
| #include "ttnn/tensor/host_buffer/functions.hpp" | ||
| #include "ttnn/tensor/layout/page_config.hpp" | ||
| #include "ttnn/tensor/storage.hpp" | ||
| #include "ttnn/tensor/tensor_impl.hpp" | ||
| #include "ttnn/tensor/tensor_utils.hpp" | ||
| #include "ttnn/tensor/tensor.hpp" | ||
| #include "ttnn/types.hpp" | ||
| #include <tt-metalium/command_queue.hpp> | ||
| #include "pybind11/stl.h" | ||
|
|
||
| #include "umd/device/tt_xy_pair.h" |
There was a problem hiding this comment.
You've added a lot of new includes. Can you please make sure you're only including the ones you need?
|
Thanks, @jjiangTT! Just need to clean up some of the includes and then we can merge this in. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Ticket
Link to Github Issue
#15061
Problem description
ListMeshToTensor was a python class in distributed.py that didn't match the xtensor->torch.tensor convention and instead output a list[torch.tensor]. I've added the utility method ttnn.shardedtensor_to_tensorlist(ttnn.tensor)->list[torch.tensor] instead.
What's changed
-ListMeshToTensor removed, all usages replaced with ttnn.sharded_tensor_to_torch_tensor_list (tensor: ttnn.tensor) hook
-Added temporary python hook in operations/core.py to convert from ttnn.tensor lists to torch.tensor lists
Checklist