Code

Abra merge test#2870

Code

Merged

cli99 merged 73 commits intoabramosaicml/composer:abrafrom abra-merge-testmosaicml/composer:abra-merge-testCopy head branch name to clipboardJan 17, 2024

+4,866-1,820

Commits

Commits on Nov 30, 2023

• Bump torch to 2.1.1 version (#2717)

  j316chuck

  authored
  1dccb14

Commits on Dec 1, 2023

• Add more info when run doesnt complete (#2751)

  aspfohl

  authored
  b11f7b6

Commits on Dec 4, 2023

• Lower sequence generation length on code gen to be dependent on max canonical solution length (#2682)

  Show description for 5d20db1
  bmosaicml

  authored
  5d20db1

Commits on Dec 7, 2023

• Remove flatten params (#2761)

  Show description for 8957550
  mvpatel2000

  authored
  8957550

• fix lint (#2767)

  mvpatel2000

  authored
  1a8a664

• lint (#2768)

  mvpatel2000

  authored
  e87c06d

• Use time.tokens for speedmonitor instead of dataset length (#2762)

  Show description for 7f55b7a
  mvpatel2000

  authored
  7f55b7a

Commits on Dec 8, 2023

• remove exception (#2759)

  mvpatel2000

  authored
  cb8f937

Commits on Dec 9, 2023

• time to clean up time parsing 😉 (#2770)

  Show description for f097fd7
  aspfohl

  authored
  f097fd7

Commits on Dec 11, 2023

• Upgrade RunConfig compute specification (#2772)

  Show description for 236b738
  aspfohl

  authored
  236b738

• Use async logging in MLflowLogger (#2693)

  Show description for 39d6df4
  chenmoneygithub

  authored
  39d6df4

• Fix FSDP _param_init_fn to not reinit parameters multiple times (#2765)

  dakinggg

  authored
  c04405e

• Gate FSDP param init test on torch 2.1 (#2774)

  dakinggg

  authored
  bc50049

Commits on Dec 12, 2023

• Parallelize OCI multipart download (#2750)

  coryMosaicML

  authored
  aad8901

• [UCVolumes] Add support for list API (#2769)

  panchalhp-db

  authored
  f497e60

• Add the memory timeline profiling support through the PyTorch profiler. (#2771)

  Show description for a7cad7c
  

  3 peopleauthored
  a7cad7c

Commits on Dec 13, 2023

• Improve torch memory profiling arguments processing (#2777)

  Show description for db3d187
  cli99

  authored
  db3d187

• Add platform AWS and bump aws ofi nccl version (#2776)

  willgleich

  authored
  0d61164

Commits on Dec 14, 2023

• Extend checkpoint loading to accept a validation function (#2726)

  irenedea

  authored
  776d172

• Fix checkpoint validation tests for torch 1.13 (#2779)

  Show description for 09f4580
  irenedea

  authored
  09f4580

• Bump version to 0.17.2 (#2780)

  Show description for 7e0e40a
  mvpatel2000

  authored
  7e0e40a

Commits on Dec 15, 2023

• bump transformers version (#2781)

  dakinggg

  authored
  45bb135

Commits on Dec 18, 2023

• Bump sphinxext-opengraph from 0.9.0 to 0.9.1 (#2784)

  Show description for 84059b6
  dependabot[bot]

  authored
  84059b6

• Bump coverage[toml] from 7.3.0 to 7.3.3 (#2783)

  Show description for 15324c7
  dependabot[bot]

  authored
  15324c7

• Update torch requirement from <2.1.2,>=1.13.1 to >=1.13.1,<2.1.3 (#2785)

  Show description for b8363bb
  dependabot[bot]

  authored
  b8363bb

Commits on Dec 19, 2023

• [UCVolumes] Rely on databricks-sdk auth for the right requirements (#2789)

  panchalhp-db

  authored
  af8797d

• Enable system metrics in mosaic mlflow logger (#2775)

  Show description for 420cb07
  

  3 peopleauthored
  420cb07

Commits on Dec 20, 2023

• Update parse_uri (#2787)

  irenedea

  authored
  f24f43c

• default-no-memory-timeline (#2790)

  cli99

  authored
  96df92d

• Add eot token to ICL generate kwargs (#2782)

  Show description for a8a261b
  bmosaicml

  authored
  a8a261b

Commits on Dec 21, 2023

• Add nightly image for torch 2.2.0 12-20-23 (#2791)

  j316chuck

  authored
  ff145d3

• Add torch nightly 12-13 (#2792)

  j316chuck

  authored
  a3ea7a4

Commits on Dec 26, 2023

• Add process group as arg to FSDP (#2794)

  Show description for 2aa50e7
  mvpatel2000

  authored
  2aa50e7

Commits on Dec 28, 2023

• Bump coverage[toml] from 7.3.3 to 7.3.4 (#2798)

  Show description for 910223e
  dependabot[bot]

  authored
  910223e

Commits on Jan 2, 2024

• Fix load_ignore_keys with rng (#2803)

  Show description for db424e5
  mvpatel2000

  authored
  db424e5

• Bump ipykernel from 6.26.0 to 6.28.0 (#2806)

  Show description for 070095e
  dependabot[bot]

  authored
  070095e

• Bump junitparser from 3.1.0 to 3.1.1 (#2805)

  Show description for e274ca0
  dependabot[bot]

  authored
  e274ca0

• Bump pytest from 7.4.3 to 7.4.4 (#2807)

  Show description for 9110c57
  dependabot[bot]

  authored
  9110c57

• Avoid futures on close for MosaicML logger (#2804)

  Show description for ed4e07c
  mvpatel2000

  authored
  ed4e07c

• check (#2812)

  mvpatel2000

  authored
  ee7cb69

• Better communication computation overlap (#2811)

  Show description for 52ac18c
  snarayan21

  authored
  52ac18c

Commits on Jan 4, 2024

• Improve error message for speed monitor (#2801)

  Show description for 80b35a7
  mvpatel2000

  authored
  80b35a7

• bump torch version (#2814)

  mvpatel2000

  authored
  f6ca956

• bump vision (#2815)

  mvpatel2000

  authored
  4af5076

• fix rng load (#2816)

  mvpatel2000

  authored
  206a9ea

• Correct multi-unshard stream patching for torch 2.2.0dev, and stream waiting correctness. (#2817)

  Show description for e5240d2
  snarayan21

  authored
  e5240d2

• fix profiler (#2818)

  mvpatel2000

  authored
  2deccf2

Commits on Jan 8, 2024

• Bump traitlets from 5.13.0 to 5.14.1 (#2822)

  Show description for 5592e41
  dependabot[bot]

  authored
  5592e41

• All unshard streams wait on computation every step (#2823)

  Show description for c22c61a
  snarayan21

  authored
  c22c61a

• Add encoding=utf-8 (#2824)

  dakinggg

  authored
  23bc6fb

• Fix import for daily test (#2826)

  Show description for a36fb74
  snarayan21

  authored
  a36fb74

• [MLFlowObjectStore] [1/2] Base implementation for MLFlowObjectStore (#2802)

  Show description for f50dcaf
  

  jerrychen109

  and

  dakinggg

  authored
  f50dcaf

Commits on Jan 9, 2024

• Remove fused layernorm (already deprecated for 2 versions) (#2827)

  Show description for 0aa95e0
  

  mvpatel2000

  and

  Your Name

  authored
  0aa95e0

• checkpoint saver tracks all checkpoints/intervals in state (#2819)

  Show description for 737b462
  

  aspfohl

  and

  mvpatel2000

  authored
  737b462

• code-quality timeout update (#2830)

  Show description for c82dcc4
  aspfohl

  authored
  c82dcc4

• [S] Fix how single value tensors are logged (#2831)

  Show description for 7b70dde
  

  aspfohl

  and

  dakinggg

  authored
  7b70dde

• Adds DTensor Support (#2821)

  Show description for 94e0386
  

  6 peopleauthored
  94e0386

Commits on Jan 10, 2024

• Remove duplicate checkpoint verifications (#2828)

  eracah

  authored
  eb4fbd0

• Fix seed for FSDP wrap (#2833)

  Show description for c48e6fe
  

  3 peopleauthored
  c48e6fe

Commits on Jan 11, 2024

• Remove fsdp patch for comm overlap (#2836)

  mvpatel2000

  authored
  6c63f2e

• allow hsdp (#2838)

  mvpatel2000

  authored
  83fb295

Commits on Jan 12, 2024

• Bump torch 2.1.2 (#2840)

  Show description for 55341aa
  mvpatel2000

  authored
  55341aa

• Upgrade pyright to 1.1.310 (#2841)

  b-chu

  authored
  2ff7c27

• [MLFlowObjectStore] [2/2] Support checkpointing with MLFlow (#2810)

  Show description for 56fa4bd
  jerrychen109

  authored
  56fa4bd

Commits on Jan 13, 2024

• update nightly to torch 2.3 (#2842)

  Show description for c9f0c21
  

  j316chuck

  and

  mvpatel2000

  authored
  c9f0c21

• Pin sphinxcontrib applehelp (#2854)

  Show description for c19fd36
  mvpatel2000

  authored
  c19fd36

• Update setup.py (#2855)

  j316chuck

  authored
  027c3d0

Commits on Jan 14, 2024

• Torch 2.3 patch (#2849)

  Show description for a2ae299
  

  3 peopleauthored
  a2ae299

Commits on Jan 15, 2024

• Update mosaicml-cli requirement from <0.6,>=0.5.25 to >=0.5.25,<0.7 (#2866)

  Show description for abff6de
  dependabot[bot]

  authored
  abff6de

• Rewrite to use individual state functions (#2860)

  Show description for d497d8f
  mvpatel2000

  authored
  d497d8f

• Add custom stopping criteria to ICL generate tasks (#2800)

  Show description for 1bc8d0a
  

  bmosaicml

  and

  dakinggg

  authored
  1bc8d0a

Commits on Jan 16, 2024

• Add save_ignore_keys (#2868)

  Show description for 31ea664
  

  mvpatel2000

  and

  dakinggg

  authored
  31ea664

• fix conflicts

  cli99

  committed
  f5978a8