-
Notifications
You must be signed in to change notification settings - Fork 713
Add support for MobileNet v2 in Ethos-U55 and arm baremetal runner #3056
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
Conversation
robell
commented
Apr 16, 2024
- Builds on TOSA support for mv2
- Currently has accuracy issues for on-target execution.
- Adding support mv2 from examples/models (torchvision mv2)
- Adds ability to use portable Q/DQ ops to enable float signature quantized models (to be improved with int signature when possible to enable integer images rather than int -> float -> quant flow.
- Adds permute for ethos-u backend to enable NCHW to NHWC (which we'll aim to replace with a compiler pass on entry/exit of the subgraph, or the NHWC support when available.
- small fix to enable qantized ops in runtime for the IOQDQ pass, and build/register those kernels/ops in bare metal build
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/3056
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 6f4ce19 with merge base cc12d9b ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
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.
This is great. Thanks @robell.
|
@digantdesai has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
2 similar comments
|
@digantdesai has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
@digantdesai has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
f346e0f to
c2442c4
Compare
|
Sigh.. can you rebase please? |
5996cbf to
0790d93
Compare
|
@digantdesai has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
1 similar comment
|
@digantdesai has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
Signed-off-by: Rob Elliott <robert.elliott@arm.com>
Signed-off-by: Rob Elliott <robert.elliott@arm.com>
Currently only enabled in the aot_ compile flow as this requires AoT kernel registration. Signed-off-by: Rob Elliott <robert.elliott@arm.com> Change-Id: I0c41b3a34613c900fc3f350f4f83b41378ac0add
… issues until further fixes * enable further input types and layouts * add optional temp support for reordering NCHW * include portable quantized ops for IOQDQ fallback Optional reordering will be removed once NHWC support is introduced. Signed-off-by: Rob Elliott <robert.elliott@arm.com> Change-Id: Iba2d4954fa14bc937584e59cb2ce82f4cdc1205d
The templated function with this change has an ambigious overload, making it unambiguous by casting up to the templated type in this quantized function
Signed-off-by: Rob Elliott <robert.elliott@arm.com>
Signed-off-by: Rob Elliott <robert.elliott@arm.com>
Signed-off-by: Rob Elliott <robert.elliott@arm.com>
Signed-off-by: Rob Elliott <robert.elliott@arm.com>
Signed-off-by: Rob Elliott <robert.elliott@arm.com>
|
@digantdesai or @mergennachin could we import this now, thanks to the fix in #3449 this now passes when rebased. |
|
@digantdesai has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
@digantdesai merged this pull request in 623f0e0. |