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

Stripped the git history to reduce the repo size #1686

Closed
wants to merge 1,467 commits into from

Conversation

arjunsuresh
Copy link
Contributor

@arjunsuresh arjunsuresh commented Apr 29, 2024

These files are removed from the git history

	Filename                  Git id
	---------------------------------------------------------------------
	model.pkl               | 0c33c4d8 (7.1 MB), 591a32a7 (5.8 MB), ...
	model_init.pb           | 83adb1d0 (9.3 MB), d2352df2 (16.0 MB), ...
	model_init.pbtxt        | 161fe94e (50.3 MB), 70cbedc0 (53.1 MB), ...
	sphereface_float.tflite | 8b6e841a (86.5 MB)

Before the change

arjun@hp-envy:~/inference$ du -sh
511M	.

After the change

arjun@hp-envy:~/inference$ du -sh
75M	.

nvpohanh and others added 30 commits June 13, 2022 15:21
Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: 77b22df
* New object detector resnext50 sample implementation

* Remove resnext50 model source code
Remove model source code and adapt scripts to work with torchscript file
Publish torchscript and onnx models in zenodo and add them to README

* Integrate new detection model with openimages

* Add script and instructions to export resnext model

* Add calibration list and scripts to download them

Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: 26951d7
Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: b284212
* Support v2.1 in submission checker

* Add ssd-resnext50 fields

Former-commit-id: 6fb638f
Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: 042f210
…ommons#1184)

* [Loadgen for LON] making non-const function for sut->Name() call

sut->Name() call requiires dynamic behavior through QDL in LON.

* Removing const from SUT Name()

* Removing const from Name() - missed ones

Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: ef663f8
Using python3 is easier here as python2 is not used anymore and python command gives error on macOS

Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: 7c3c697
* Update README (mlcommons#1167)

* Change python to python3

* Update name and model scores

* Rename resnext50 to retinanet

* Add new BERT calibration dataset (mlcommons#1172)

Co-authored-by: rameshchukka <rnaidu02@yahoo.com>

* LoadGen: Update to version number

* Update random seeds (mlcommons#1179)

* Fix regex (mlcommons#1178)

* Merge master into 2.1. Fix merge issues (mlcommons#1191)

* Update CLA bot (mlcommons#1180)

* removing obselete audit directory (mlcommons#1188)

* [Loadgen for LON] making non-const function for sut->Name() call (mlcommons#1184)

* [Loadgen for LON] making non-const function for sut->Name() call

sut->Name() call requiires dynamic behavior through QDL in LON.

* Removing const from SUT Name()

* Removing const from Name() - missed ones

Co-authored-by: rameshchukka <rnaidu02@yahoo.com>

* Update run_local.sh (mlcommons#1181)

Using python3 is easier here as python2 is not used anymore and python command gives error on macOS

Co-authored-by: rameshchukka <rnaidu02@yahoo.com>

Co-authored-by: Bruno Ferreira <bruno.mgf@gmail.com>
Co-authored-by: georgelyuan <53881988+georgelyuan@users.noreply.github.com>
Co-authored-by: Jinho Suh <83969361+nv-jinhosuh@users.noreply.github.com>
Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Co-authored-by: Arjun Suresh <arjunsuresh1987@gmail.com>

* Compliance patch (mlcommons#1195)

* Replace ssd-resnet by retinanet

* Fix compliance folder issue

* Replace ssd-resnet by retinanet (mlcommons#1194)

Co-authored-by: tjablin <tjablin@google.com>

* Update power-dev submodule hash to latest r2.1

* Remove const in QSL (mlcommons#1199)

* Update power-dev submodule

* [r2.1] Add bert-99 and bert-99.9 to the submission-checker model_mapping guess

* Flip order of conditions

* Remove retinanet from TEST04

retinanet NMS runtime is significantly larger than the old ssd benchmarks due to the larger number of classes in OpenImages. As such, the assumption that TEST04 makes that all samples should have roughly equal computational cost no longer holds for this workload.

* Update TEST04 readme

* Update final report script (mlcommons#1211)

* Support for network division and multi-nodes (mlcommons#1214)

* Compute symmetric difference correctly

* Ignore .gitkeep

* 2.0 -> 2.1, Filter Power

* Four digit system IDs

* Fix an unnecessary '.' at beginning of notes if hw_notes is empty (mlcommons#1224)

* Fix an unnecessary '.' at beginning of notes if hw_notes is empty

* Add Part3 of Compliance TEST01 to Submission Checker

* Update submission-checker.py

Fix the required_delta for higher accuracy models

* Fix delta calculation for compliance TEST01 part 3

* Run pyformat on submission related scripts

* Tuples (mlcommons#1234)

* Support tuples for accelerator count

* Support ssd-large in open

Co-authored-by: Pablo Gonzalez <pablo.gonzalez@factored.ai>
Co-authored-by: nvitramble <84538536+nvitramble@users.noreply.github.com>
Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Co-authored-by: Core.Halt <46298335+corehalt@users.noreply.github.com>
Co-authored-by: Bruno Ferreira <bruno.mgf@gmail.com>
Co-authored-by: georgelyuan <53881988+georgelyuan@users.noreply.github.com>
Co-authored-by: Jinho Suh <83969361+nv-jinhosuh@users.noreply.github.com>
Co-authored-by: Arjun Suresh <arjunsuresh1987@gmail.com>
Co-authored-by: Po-Han Huang <pohanh@nvidia.com>
Co-authored-by: Michael Goin <michael@neuralmagic.com>
Former-commit-id: 22579b4
* Passing bert data and model paths from env

* Added an option to skip accuracy verification for bert

* Added an option to override log path for bert

Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: 8c59377
Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: 161b745
* Adds support for multi-threaded preprocessing of images

* Added cuda support for bert

* Fixes issue with mlcommons#410

* Fixes issue with mlcommons#410

* Fixes issue with mlcommons#1249 for image-classification, avoid torch usage for onnxruntime backend

* Fixes issue with mlcommons#1249

* Update file names in Readme

* Removed unnecessary comment

* Fix for mlcommons#1264, disable highest level of optimizations for onnxruntime backend on aarch64

Co-authored-by: pgmpablo157321 <pablo.gonzalez@factored.ai>
Former-commit-id: 270360d
mlcommons#1254 has changed the names for the external files, resulting in an error while executing the reference BERT.
This is a simple fix to complement those changes.

Former-commit-id: 83e2cc9
Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: 37d3778
Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: 7937922
Co-authored-by: rameshchukka <rnaidu02@yahoo.com>
Former-commit-id: 986ade9
* Added Github test for loadgen

* Fix git branch

Former-commit-id: 0b18682
pgmpablo157321 and others added 23 commits February 7, 2024 09:27
Co-authored-by: Miro <mirhodak@amd.com>
Former-commit-id: d7dba08
* Update README.md

No longer need custom fork as the relevant changes are in the inference repository

* Update dataset.py

---------

Co-authored-by: Miro <mirhodak@amd.com>
Former-commit-id: b0777f0
Co-authored-by: Miro <mirhodak@amd.com>
Former-commit-id: 3190d09
…and dlrmv2 models (mlcommons#1604)

* Update README.md

Add CM commands to download Stable diffusion models

* Update README.md

* Update README.md

Former-commit-id: 840435a
* Turn equal issue mode off for Llama2 TEST06

* Add TEST06 to the output dir

Former-commit-id: 817dd96
* Fix submission checker and TEST06 for Llama2

* Remove redundant line

* Move test_dir check

Former-commit-id: 0ed5190
…UNet) (mlcommons#1624)

Currently 3D-UNet is the only workload using equal-issue mode on Offline scenario. 
Recent code change on LLM equal-issue mode caused 3D-UNet accuracy run to run more than 1 queries, causing the accuracy log to bloat and fail the accuracy checking script.
This change fixes the problem described above.

Former-commit-id: f06b920
* Hotfix: DLRMv2 Audit Test01 fallback failure 

DLRMv2 Audit TEST01 may go to fallback route and the accuracy check script (accuracy-dlrm.py) didn't expect this to happen. It always expects entire sample set to be in the accuracy log while Audit TEST01 would generate subset only.

This fixes the Audit TEST01 failure described above.

* typo fix

Former-commit-id: 486a629
Former-commit-id: e39003a
@arjunsuresh arjunsuresh requested a review from a team as a code owner April 29, 2024 21:48
Copy link

github-actions bot commented Apr 29, 2024

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

@arjunsuresh
Copy link
Contributor Author

closing this as this PR cannot take a sync from master branch

@github-actions github-actions bot locked and limited conversation to collaborators May 18, 2024
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.