Code

Add TF and ONNX Workers#777

Code

Open

al-rigazzi wants to merge 71 commits intov1.0CrayLabs/SmartSim:v1.0from tf_worker_1.0CrayLabs/SmartSim:tf_worker_1.0Copy head branch name to clipboard

+7,440-240

Commits

Commits on Jun 11, 2024

• Initial MLI schemas and MessageHandler class (#607)

  Show description for d2fd6a7
  AlyssaCote

  authored
  d2fd6a7

  Copy full SHA for d2fd6a7

Commits on Jun 14, 2024

• Merge branch 'develop' into mli-feature

  ankona

  committed
  3c9915c

  Copy full SHA for 3c9915c

Commits on Jun 20, 2024

• ML Worker Manager MVP (#608)

  Show description for 38081da
  ankona

  authored
  38081da

  Copy full SHA for 38081da

Commits on Jun 25, 2024

• Remove device attribute from schemas (#619)

  Show description for ab900b8
  AlyssaCote

  authored
  ab900b8

  Copy full SHA for ab900b8

Commits on Jul 2, 2024

• Merge branch 'develop' into mli-feature

  ankona

  committed
  a9ffb14

  Copy full SHA for a9ffb14

• Merge branch 'develop' into mli-feature

  ankona

  committed
  ee2c110

  Copy full SHA for ee2c110

Commits on Jul 3, 2024

• Add model metadata to request schema (#624)

  Show description for 8a2f173
  AlyssaCote

  authored
  8a2f173

  Copy full SHA for 8a2f173

Commits on Jul 10, 2024

• Enable environment variable based configuration for ML Worker Manager (#621)

  Show description for 52abd32
  AlyssaCote

  authored
  52abd32

  Copy full SHA for 52abd32

Commits on Jul 15, 2024

• FLI-based Worker Manager (#622)

  Show description for eace71e
  al-rigazzi

  authored
  eace71e

  Copy full SHA for eace71e

Commits on Jul 17, 2024

• Add ability to specify hardware policies on dragon run requests (#631)

  Show description for 5fac3e2
  ankona

  authored
  5fac3e2

  Copy full SHA for 5fac3e2

• Revert "Add ability to specify hardware policies on dragon run requests" (#637)

  Show description for 0030a4a
  ankona

  authored
  0030a4a

  Copy full SHA for 0030a4a

Commits on Jul 18, 2024

• Merge latest develop into mli-feature (#640)

  Show description for b6c2f2b
  ankona

  authored
  b6c2f2b

  Copy full SHA for b6c2f2b

• Improve error handling in worker manager (#629)

  Show description for 272a1d7
  AlyssaCote

  authored
  272a1d7

  Copy full SHA for 272a1d7

• Schema performance improvements (#632)

  Show description for 7169f1c
  AlyssaCote

  authored
  7169f1c

  Copy full SHA for 7169f1c

Commits on Jul 19, 2024

• New develop merger (#645)

  Show description for 84101b3
  al-rigazzi

  authored
  84101b3

  Copy full SHA for 84101b3

Commits on Jul 26, 2024

• merging develop

  ankona

  committed
  e225c07

  Copy full SHA for e225c07

Commits on Jul 31, 2024

• Merge branch 'develop' into mli-feature

  ankona

  committed
  9f482b1

  Copy full SHA for 9f482b1

Commits on Aug 2, 2024

• Fix dragon installation issues (#652)

  Show description for 263e3c7
  ankona

  authored
  263e3c7

  Copy full SHA for 263e3c7

Commits on Aug 7, 2024

• Add FeatureStore descriptor to tensor & model keys (#633)

  Show description for 0453b8b
  ankona

  authored
  0453b8b

  Copy full SHA for 0453b8b

Commits on Aug 8, 2024

• Merge branch 'develop' into mli-feature

  ankona

  committed
  99ed41c

  Copy full SHA for 99ed41c

• Use torch.from_numpy instead of torch.tensor to reduce a copy (#661)

  Show description for 74d6e78
  AlyssaCote

  authored
  74d6e78

  Copy full SHA for 74d6e78

Commits on Aug 14, 2024

• MLI environment variables updated using new naming convention (#665)

  Show description for 391784c
  AlyssaCote

  authored
  391784c

  Copy full SHA for 391784c

Commits on Aug 20, 2024

• Remove pydantic dependency from MLI code (#667)

  Show description for f7ef49b
  AlyssaCote

  authored
  f7ef49b

  Copy full SHA for f7ef49b

Commits on Aug 26, 2024

• Enable specification of target hostname for a dragon task (#660)

  Show description for ef034d5
  ankona

  authored
  ef034d5

  Copy full SHA for ef034d5

• fix init reordering bug (#675)

  Show description for 6d5518b
  ankona

  authored
  6d5518b

  Copy full SHA for 6d5518b

Commits on Aug 28, 2024

• Queue-based Worker Manager (#647)

  Show description for 5d85995
  al-rigazzi

  authored
  5d85995

  Copy full SHA for 5d85995

Commits on Aug 29, 2024

• Refactor exception_handler to avoid unnecessary building and serialization of failure responses. (#687)

  Show description for 128598b
  AlyssaCote

  authored
  128598b

  Copy full SHA for 128598b

• SmartSim environment variables updated using new naming convention (#666)

  Show description for 8aa990c
  AlyssaCote

  authored
  8aa990c

  Copy full SHA for 8aa990c

Commits on Aug 30, 2024

• MLI file names conform to snake case (#689)

  Show description for f6d55d8
  AlyssaCote

  authored
  f6d55d8

  Copy full SHA for f6d55d8

Commits on Sep 10, 2024

• Add event broadcasting capability (#672)

  Show description for a3795e7
  ankona

  authored
  a3795e7

  Copy full SHA for a3795e7

• Update MLI docstrings part 1 (#692)

  Show description for 28bfd8f
  AlyssaCote

  authored
  28bfd8f

  Copy full SHA for 28bfd8f

• Update MLI docstrings part 2 (#699)

  Show description for 61ab71d
  AlyssaCote

  authored
  61ab71d

  Copy full SHA for 61ab71d

Commits on Sep 12, 2024

• Parametrize dragon install (#703)

  Show description for 3e9bffa
  ankona

  authored
  3e9bffa

  Copy full SHA for 3e9bffa

• Merge branch 'develop' into mli-feature

  ankona

  committed
  b4798da

  Copy full SHA for b4798da

Commits on Sep 18, 2024

• Ensure forks build latest public dragon in CI actions (#706)

  Show description for b0b1db6
  ankona

  authored
  b0b1db6

  Copy full SHA for b0b1db6

Commits on Sep 19, 2024

• Clean up error handling in MLI (#698)

  Show description for 0ebd5ab
  AlyssaCote

  authored
  0ebd5ab

  Copy full SHA for 0ebd5ab

• Version used to run benchmarks

  al-rigazzi

  committed
  7336031

  Copy full SHA for 7336031

• Merge branch 'mli-feature' into benchmark

  al-rigazzi

  committed
  77791d9

  Copy full SHA for 77791d9

• MLI helper methods (#709)

  Show description for d43f7c7
  AlyssaCote

  authored
  d43f7c7

  Copy full SHA for d43f7c7

Commits on Sep 20, 2024

• Bug fix (#715)

  Show description for 5ec287c
  AlyssaCote

  authored
  5ec287c

  Copy full SHA for 5ec287c

Commits on Sep 25, 2024

• Committing pre-merge

  al-rigazzi

  committed
  ce02433

  Copy full SHA for ce02433

• Merge branch 'mli-feature' of https://github.com/CrayLabs/SmartSim into benchmark

  al-rigazzi

  committed
  5217a0a

  Copy full SHA for 5217a0a

• Working after merge

  al-rigazzi

  committed
  c1c4604

  Copy full SHA for c1c4604

• Improve TF Worker throughput

  al-rigazzi

  committed
  3a5a8ce

  Copy full SHA for 3a5a8ce

Commits on Sep 26, 2024

• Revert changes to throughput scripts

  al-rigazzi

  committed
  48caac2

  Copy full SHA for 48caac2

• Style, lint, changelog

  al-rigazzi

  committed
  f79f53c

  Copy full SHA for f79f53c

• Add tests for TF worker

  al-rigazzi

  committed
  f569b9c

  Copy full SHA for f569b9c

• Improve standalone_worker_manager.py

  al-rigazzi

  committed
  50ce6ad

  Copy full SHA for 50ce6ad

• Remove commented sections

  al-rigazzi

  committed
  d49634f

  Copy full SHA for d49634f

• Style

  al-rigazzi

  committed
  a8d501b

  Copy full SHA for a8d501b

• Switch to Channel.make_process_local

  al-rigazzi

  committed
  79b4954

  Copy full SHA for 79b4954

• Add timeout and exc handling to WM response

  al-rigazzi

  committed
  152d434

  Copy full SHA for 152d434

• More commented lines to remove

  al-rigazzi

  committed
  14b20f1

  Copy full SHA for 14b20f1

• Remove debug information

  al-rigazzi

  committed
  dbbdcd9

  Copy full SHA for dbbdcd9

Commits on Sep 27, 2024

• Fix issue when keras layer has "resource" field

  al-rigazzi

  committed
  b9bdf99

  Copy full SHA for b9bdf99

Commits on Oct 2, 2024

• Address first comments.

  al-rigazzi

  committed
  3f571f4

  Copy full SHA for 3f571f4

Commits on Oct 3, 2024

• Add ONNX worker

  al-rigazzi

  committed
  dc7307a

  Copy full SHA for dc7307a

Commits on Oct 4, 2024

• Add onnx mock app

  al-rigazzi

  committed
  dccc07d

  Copy full SHA for dccc07d

• Style

  al-rigazzi

  committed
  a590a36

  Copy full SHA for a590a36

• Add optional compile step for Torch model

  al-rigazzi

  committed
  bf324d2

  Copy full SHA for bf324d2

Commits on Oct 10, 2024

• Add integration of dragon-based event broadcasting (#710)

  Show description for ca01cb1
  ankona

  authored
  ca01cb1

  Copy full SHA for ca01cb1

Commits on Oct 11, 2024

• Refine try-catch in onnx worker

  al-rigazzi

  committed
  dc31b75

  Copy full SHA for dc31b75

• Merge branch 'mli-feature' of https://github.com/CrayLabs/SmartSim into tf_worker

  al-rigazzi

  committed
  5afcfcf

  Copy full SHA for 5afcfcf

• Use new ProtoClient in apps

  al-rigazzi

  committed
  4e6ddff

  Copy full SHA for 4e6ddff

• Style

  al-rigazzi

  committed
  9947013

  Copy full SHA for 9947013

• Mypy

  al-rigazzi

  committed
  23de37f

  Copy full SHA for 23de37f

• Style

  al-rigazzi

  committed
  10d59c8

  Copy full SHA for 10d59c8

Commits on Oct 14, 2024

• Fix tests

  al-rigazzi

  committed
  ce5a306

  Copy full SHA for ce5a306

• Fix tests

  al-rigazzi

  committed
  f23c267

  Copy full SHA for f23c267

Commits on Oct 31, 2024

• Merge branch 'v1.0' of https://github.com/CrayLabs/SmartSim into tf_worker_1.0

  al-rigazzi

  committed
  6a08b82

  Copy full SHA for 6a08b82

• Complete post-merge operations

  al-rigazzi

  committed
  da92472

  Copy full SHA for da92472