Releases · mlcommons/cm4mlops

What's Changed

fixed several URLs (all tests passed) by @gfursin in #342
Changes for supporting model and dataset download to host - Mixtral by @anandhu-eng in #346
Support custom git clone branch in docker by @anandhu-eng in #343
Merge from go, fixes #337 by @arjunsuresh in #348
Includes cuda version to run suffix by @anandhu-eng in #354
Fixes for const in script by @arjunsuresh in #355
Fix docker image naming SCC24, extended CM script tests by @arjunsuresh in #356
Fixes for MLPerf Inference Github Actions by @arjunsuresh in #362
Fix typo in gh action by @arjunsuresh in #363
Fix CUDA num_devices by @arjunsuresh in #365
Support cleaning of Nvidia SDXL model by @arjunsuresh in #366
Improvements to Nvidia MLPerf interface by @arjunsuresh in #367
Fixes to pull changes for Nvidia implementation by @arjunsuresh in #369
Improvements to MLPerf inference final report generation by @arjunsuresh in #371
Support get-platform-details for mlperf-inference by @arjunsuresh in #373
Support system_info.txt in MLPerf inference submission generation by @arjunsuresh in #374
Cleanups for mlperf inference get-platform-details by @arjunsuresh in #375
Improvements to get-platform-details by @arjunsuresh in #376
Improvements for reproducing AMD implementation by @anandhu-eng in #379
Improvements to amd LLAMA2 70B command generation - Added server scenario by @anandhu-eng in #383
Build wheels and release them into PYPI by @anandhu-eng in #385
Do not pass mlperf_conf for inference-src >= 4.1.1 by @arjunsuresh in #404
Fix version check for mlperf-inference-src by @arjunsuresh in #405
Added no-compilation-warning variation for loadgen by @arjunsuresh in #406
Support 8G Nvidia GPUs for MLPerf Inference by @arjunsuresh in #411
Fix bug on benchmark-program exit check by @arjunsuresh in #412
Improve the benchmark-program-mlperf run command by @arjunsuresh in #413
CM4MLOps snapshot with MLPerf inference: 20241024 by @gfursin in #415
fix batch size duplication issue by @anandhu-eng in #416
Update cm repo branch - docker by @anandhu-eng in #422
Fixes for Nvidia MLPerf inference SS and MS by @arjunsuresh in #423
Improvements for Nvidia MLPerf inference by @arjunsuresh in #428
added compressed_tools module by @anandhu-eng in #430
Updated logic for mounting non cache folder by @anandhu-eng in #427
Fixes for latest MLPerf inference submission checker changes by @arjunsuresh in #431
Fixes for Latest MLPerf inference changes by @arjunsuresh in #432
Fixes for latest MLPerf inference changes by @arjunsuresh in #433
Submission generation fixes by @anandhu-eng in #424
Support custom path for saving platform details by @anandhu-eng in #418
Add getting started to cm4mlops docs by @anandhu-eng in #435
Merge from Mlperf inference by @anandhu-eng in #436
Fixes for docker detached mode by @arjunsuresh in #438
Fixes for get-platform-details by @arjunsuresh in #441
Testing CM Test automation by @arjunsuresh in #442
Added github action for individual CM tests by @arjunsuresh in #443
Fix Individual CM script test by @arjunsuresh in #445
Fix individual CM script test by @arjunsuresh in #446
Fix gh action for individual CM sript tests by @arjunsuresh in #447
Initial PR - gh actions for submission generation for non CM based benchmarks by @anandhu-eng in #440
Fix GH action for individual CM script testing by @arjunsuresh in #449
Enable docker run for individual CM script tests by @arjunsuresh in #450
Fixes huggingface downloader by @arjunsuresh in #452
capture framework version from cm_sut_info.json by @anandhu-eng in #451
Sync: Mlperf inference by @arjunsuresh in #444
Support docker_base_image and docker_cm_repo for CM tests by @arjunsuresh in #453
Improved docker meta for cm test script by @arjunsuresh in #454
Support test_input_index and test_input_id to customise CM test script by @arjunsuresh in #458
Improvements to version-detect in get-generic-sys-util by @arjunsuresh in #460
Use pkg-config deps for get-generic-sys-util by @arjunsuresh in #464
Fix pstree version detect on macos by @arjunsuresh in #466
Fix Nvidia mlperf inference retinanet | onnx version by @arjunsuresh in #468
Support detached mode for nvidia-mlperf-inference-gptj by @arjunsuresh in #469
Cleanups for get-generic-sys-util by @arjunsuresh in #470
Fix tmp-run-env.out name by @arjunsuresh in #471
Add Version RE for g++-11 by @arjunsuresh in #472
Use berkeley link for imagenet-aux by default by @arjunsuresh in #473
Code cleanup by @anandhu-eng in #475
Fixes for Sdxl MLPerf inference by @arjunsuresh in #481
Added google dns to mlperf-inference docker by @arjunsuresh in #484
enable submission generation gh action test globally by @anandhu-eng in #487
Enables docker run in inference submission generation by @anandhu-eng in #486
Added docker detatched option by @anandhu-eng in #477
Fixes #261 partially by @Oseltamivir in #426
Enabled docker run - gh action submission generation by @anandhu-eng in #489
Fixes for MLPerf inference, intel conda URL by @arjunsuresh in #491
Dont use '-dt' for Nvidia ml-model-gptj by @arjunsuresh in #492
Update starting weights filename for SDXL MLPerf inference by @arjunsuresh in #494
Implements #455: Copy local repo to docker instead of git clone by @Oseltamivir in #467
Fixes for Nvidia MLPerf inference gptj,sdxl by @arjunsuresh in #495
Cleanups to MLPerf inference preprocess script by @arjunsuresh in #496
Support sample-ids for coco2014 accuracy script by @arjunsuresh in #497
Support download to host - amd llama2 by @anandhu-eng in #480
Added a retry for git clone failure by @arjunsuresh in #499
Use custom version for dev branch of inference-src by @arjunsuresh in #500
Fix path error shown to user by @anandhu-eng in #502
Fixes for the MLPerf inference nightly test failures by @arjunsuresh in #506
pip install cm4mlops - handle systems where sudo is absent by @anandhu-eng in #504
logic update for detect sudo by @anandhu-eng in #508
Fix Nvidia MLPerf inference gptj model name suffix by @arjunsuresh in #509
Skip sys-utils-install when no...