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.

Sign up for GitHub

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

Jump to bottom

call #268

Open
wants to merge 236 commits into
base: dev
Choose a base branch
from
Open

call #268

wants to merge 236 commits into from

Commits on Dec 17, 2023

  1. add trtllm support 

    Signed-off-by: jiemingz <[email protected]>
    @jiemingz
    jiemingz committed Dec 17, 2023
    Configuration menu
    Copy the full SHA
    5e67140 View commit details
    Browse the repository at this point in the history

  2. dataloader reshard 

    Signed-off-by: jiemingz <[email protected]>
    @jiemingz
    jiemingz committed Dec 17, 2023
    Configuration menu
    Copy the full SHA
    d955bb4 View commit details
    Browse the repository at this point in the history

Commits on Feb 22, 2024

  1. cleanup code and support trtllm 0.7 

    Signed-off-by: jiemingz <=>
    jiemingz committed Feb 22, 2024
    Configuration menu
    Copy the full SHA
    f75ae54 View commit details
    Browse the repository at this point in the history

Commits on Feb 23, 2024

  1. cleanup and sync with nemo exprot 

    Signed-off-by: Jimmy Zhang <[email protected]>
    jiemingz committed Feb 23, 2024
    Configuration menu
    Copy the full SHA
    25d8aae View commit details
    Browse the repository at this point in the history

  2. remove resharding 

    Signed-off-by: Jimmy Zhang <[email protected]>
    jiemingz committed Feb 23, 2024
    Configuration menu
    Copy the full SHA
    5aa9dd4 View commit details
    Browse the repository at this point in the history

  3. fix topp and refactor 

    Signed-off-by: jiemingz <=>
    jiemingz committed Feb 23, 2024
    Configuration menu
    Copy the full SHA
    ac8933c View commit details
    Browse the repository at this point in the history

Commits on Feb 24, 2024

  1. clamp tokens 

    Signed-off-by: jiemingz <=>
    jiemingz committed Feb 24, 2024
    Configuration menu
    Copy the full SHA
    f023c7d View commit details
    Browse the repository at this point in the history

Commits on Feb 29, 2024

  1. resharding PP support 

    Signed-off-by: jiemingz <=>
    jiemingz committed Feb 29, 2024
    Configuration menu
    Copy the full SHA
    46828bf View commit details
    Browse the repository at this point in the history

Commits on Mar 6, 2024

  1. add mbs support 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm authored and jiemingz committed Mar 6, 2024
    Configuration menu
    Copy the full SHA
    8d50946 View commit details
    Browse the repository at this point in the history

Commits on Mar 7, 2024

  1. Added nondivisible batch support on critic 

    Signed-off-by: jiemingz <=>
    jiemingz committed Mar 7, 2024
    Configuration menu
    Copy the full SHA
    0a0893a View commit details
    Browse the repository at this point in the history

Commits on Mar 13, 2024

  1. format only 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 13, 2024
    Configuration menu
    Copy the full SHA
    c25cf6e View commit details
    Browse the repository at this point in the history

  2. log prob batch 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 13, 2024
    Configuration menu
    Copy the full SHA
    36f9664 View commit details
    Browse the repository at this point in the history

  3. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 13, 2024
    Configuration menu
    Copy the full SHA
    e52a35d View commit details
    Browse the repository at this point in the history

Commits on Mar 14, 2024

  1. debug 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 14, 2024
    Configuration menu
    Copy the full SHA
    1ae9987 View commit details
    Browse the repository at this point in the history

  2. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 14, 2024
    Configuration menu
    Copy the full SHA
    60aaecb View commit details
    Browse the repository at this point in the history

Commits on Mar 17, 2024

  1. test 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 17, 2024
    Configuration menu
    Copy the full SHA
    d68ed96 View commit details
    Browse the repository at this point in the history

Commits on Mar 19, 2024

  1. dummy 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 19, 2024
    Configuration menu
    Copy the full SHA
    b2bfa6f View commit details
    Browse the repository at this point in the history

Commits on Mar 20, 2024

  1. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 20, 2024
    Configuration menu
    Copy the full SHA
    297255d View commit details
    Browse the repository at this point in the history

  2. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 20, 2024
    Configuration menu
    Copy the full SHA
    d4910b7 View commit details
    Browse the repository at this point in the history

  3. restore critic 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 20, 2024
    Configuration menu
    Copy the full SHA
    4a52afe View commit details
    Browse the repository at this point in the history

Commits on Mar 21, 2024

  1. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    2140f36 View commit details
    Browse the repository at this point in the history

  2. add logging 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    e7ea083 View commit details
    Browse the repository at this point in the history

  3. fix critic client 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    6cd8923 View commit details
    Browse the repository at this point in the history

  4. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    887874d View commit details
    Browse the repository at this point in the history

  5. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    6e2ca2d View commit details
    Browse the repository at this point in the history

  6. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    bacd786 View commit details
    Browse the repository at this point in the history

  7. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    17b04ca View commit details
    Browse the repository at this point in the history

  8. fix typo 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    9fee58b View commit details
    Browse the repository at this point in the history

  9. better train timing 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    61ba204 View commit details
    Browse the repository at this point in the history

  10. remove prints 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    c9dad2a View commit details
    Browse the repository at this point in the history

  11. with timing 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    e5f68e2 View commit details
    Browse the repository at this point in the history

  12. delete unused func 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    74791ae View commit details
    Browse the repository at this point in the history

  13. add critic logging 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Mar 21, 2024
    Configuration menu
    Copy the full SHA
    e40ebd6 View commit details
    Browse the repository at this point in the history

Commits on Apr 1, 2024

  1. add 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 1, 2024
    Configuration menu
    Copy the full SHA
    72ba6c6 View commit details
    Browse the repository at this point in the history

Commits on Apr 6, 2024

  1. cleanup 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 6, 2024
    Configuration menu
    Copy the full SHA
    bfb61e4 View commit details
    Browse the repository at this point in the history

  2. update 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 6, 2024
    Configuration menu
    Copy the full SHA
    21206c5 View commit details
    Browse the repository at this point in the history

  3. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 6, 2024
    Configuration menu
    Copy the full SHA
    148acf4 View commit details
    Browse the repository at this point in the history

  4. fix bug 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 6, 2024
    Configuration menu
    Copy the full SHA
    d7c9990 View commit details
    Browse the repository at this point in the history

  5. fix bug 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 6, 2024
    Configuration menu
    Copy the full SHA
    537d6e5 View commit details
    Browse the repository at this point in the history

  6. test 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 6, 2024
    Configuration menu
    Copy the full SHA
    47400ba View commit details
    Browse the repository at this point in the history

  7. fix bug 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 6, 2024
    Configuration menu
    Copy the full SHA
    d7b2b23 View commit details
    Browse the repository at this point in the history

Commits on Apr 7, 2024

  1. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 7, 2024
    Configuration menu
    Copy the full SHA
    ce76226 View commit details
    Browse the repository at this point in the history

  2. add 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 7, 2024
    Configuration menu
    Copy the full SHA
    8edf534 View commit details
    Browse the repository at this point in the history

Commits on Apr 8, 2024

  1. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    6379a2e View commit details
    Browse the repository at this point in the history

  2. fix again 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    eadae31 View commit details
    Browse the repository at this point in the history

  3. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    e2b97d9 View commit details
    Browse the repository at this point in the history

  4. fix mean 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    d9bdf7c View commit details
    Browse the repository at this point in the history

  5. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    1c7d215 View commit details
    Browse the repository at this point in the history

  6. add debug 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    3638301 View commit details
    Browse the repository at this point in the history

  7. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    4cca85f View commit details
    Browse the repository at this point in the history

  8. add data iter for VP 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    1b19bdd View commit details
    Browse the repository at this point in the history

  9. move 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    3f045ae View commit details
    Browse the repository at this point in the history

  10. fixing 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    3c9fe3d View commit details
    Browse the repository at this point in the history

  11. add 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    f36f394 View commit details
    Browse the repository at this point in the history

  12. chunking needs to be moved out 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    5211bc2 View commit details
    Browse the repository at this point in the history

  13. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 8, 2024
    Configuration menu
    Copy the full SHA
    0f59edf View commit details
    Browse the repository at this point in the history

Commits on Apr 9, 2024

  1. fix metrics 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 9, 2024
    Configuration menu
    Copy the full SHA
    c3fe2f7 View commit details
    Browse the repository at this point in the history

  2. fix dtype 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 9, 2024
    Configuration menu
    Copy the full SHA
    5d3e07d View commit details
    Browse the repository at this point in the history

  3. merge 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 9, 2024
    Configuration menu
    Copy the full SHA
    15887e5 View commit details
    Browse the repository at this point in the history

  4. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 9, 2024
    Configuration menu
    Copy the full SHA
    2ad76ba View commit details
    Browse the repository at this point in the history

Commits on Apr 10, 2024

  1. make the global id management into a class 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 10, 2024
    Configuration menu
    Copy the full SHA
    9d9a6b6 View commit details
    Browse the repository at this point in the history

Commits on Apr 11, 2024

  1. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 11, 2024
    Configuration menu
    Copy the full SHA
    d6fb55d View commit details
    Browse the repository at this point in the history

Commits on Apr 17, 2024

  1. trtllm0.9 changes (#149) 

    * trtllm0.9 changes

Signed-off-by: jiemingz <=>

* fix typos

Signed-off-by: jiemingz <=>

* address comments

Signed-off-by: jiemingz <=>

* fixes

Signed-off-by: jiemingz <=>

* fix

Signed-off-by: jiemingz <=>

* fix nemo generations with PP

Signed-off-by: jiemingz <=>

* add engine_unload

Signed-off-by: jiemingz <=>

* cleanup trtllm

Signed-off-by: jiemingz <=>

* address comments

Signed-off-by: jiemingz <=>

---------

Signed-off-by: jiemingz <=>
Co-authored-by: jiemingz <=>
    @jiemingz
    jiemingz authored Apr 17, 2024
    Configuration menu
    Copy the full SHA
    0983164 View commit details
    Browse the repository at this point in the history

  2. [pre-commit.ci] auto fixes from pre-commit.com hooks 

    for more information, see https://pre-commit.ci
    @pre-commit-ci
    pre-commit-ci[bot] committed Apr 17, 2024
    Configuration menu
    Copy the full SHA
    fe765cd View commit details
    Browse the repository at this point in the history

  3. trtllm patch file 

    Signed-off-by: Jimmy Zhang <[email protected]>
    @jiemingz
    jiemingz committed Apr 17, 2024
    Configuration menu
    Copy the full SHA
    dfac922 View commit details
    Browse the repository at this point in the history

  4. dockerfile 

    Signed-off-by: Jimmy Zhang <[email protected]>
    @jiemingz
    jiemingz committed Apr 17, 2024
    Configuration menu
    Copy the full SHA
    c159aa3 View commit details
    Browse the repository at this point in the history

Commits on Apr 18, 2024

  1. fix build 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 18, 2024
    Configuration menu
    Copy the full SHA
    d81caef View commit details
    Browse the repository at this point in the history

  2. fix bug 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 18, 2024
    Configuration menu
    Copy the full SHA
    92c19f6 View commit details
    Browse the repository at this point in the history

Commits on Apr 19, 2024

  1. add groupnorm build 

    Signed-off-by: Jimmy Zhang <[email protected]>
    @jiemingz
    jiemingz committed Apr 19, 2024
    Configuration menu
    Copy the full SHA
    7088f54 View commit details
    Browse the repository at this point in the history

  2. upgrade to latest te and mcore 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 19, 2024
    Configuration menu
    Copy the full SHA
    472a56c View commit details
    Browse the repository at this point in the history

  3. Merge remote-tracking branch 'origin/dev' into aligner_trt_build 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 19, 2024
    Configuration menu
    Copy the full SHA
    032bf35 View commit details
    Browse the repository at this point in the history

  4. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 19, 2024
    Configuration menu
    Copy the full SHA
    d5f55f5 View commit details
    Browse the repository at this point in the history

Commits on Apr 20, 2024

  1. specify max token 

    Signed-off-by: jiemingz <=>
    jiemingz committed Apr 20, 2024
    Configuration menu
    Copy the full SHA
    c7cdca1 View commit details
    Browse the repository at this point in the history

  2. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 20, 2024
    Configuration menu
    Copy the full SHA
    04d02c8 View commit details
    Browse the repository at this point in the history

  3. Merge remote-tracking branch 'origin/geshen/main_trt' into aligner_tr… 

    …t_build
    @gshennvm
    gshennvm committed Apr 20, 2024
    Configuration menu
    Copy the full SHA
    56ccacf View commit details
    Browse the repository at this point in the history

Commits on Apr 22, 2024

  1. fix critic checkpoint loading 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 22, 2024
    Configuration menu
    Copy the full SHA
    d23865f View commit details
    Browse the repository at this point in the history

  2. add assert 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 22, 2024
    Configuration menu
    Copy the full SHA
    3c21c81 View commit details
    Browse the repository at this point in the history

  3. fix bug 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 22, 2024
    Configuration menu
    Copy the full SHA
    2c99dcb View commit details
    Browse the repository at this point in the history

  4. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 22, 2024
    Configuration menu
    Copy the full SHA
    9e8526d View commit details
    Browse the repository at this point in the history

Commits on Apr 23, 2024

  1. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 23, 2024
    Configuration menu
    Copy the full SHA
    c1daeb9 View commit details
    Browse the repository at this point in the history

  2. update dockerfile 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 23, 2024
    Configuration menu
    Copy the full SHA
    e16c357 View commit details
    Browse the repository at this point in the history

  3. update to 24.03.01 deps 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 23, 2024
    Configuration menu
    Copy the full SHA
    410eaf5 View commit details
    Browse the repository at this point in the history

Commits on Apr 24, 2024

  1. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 24, 2024
    Configuration menu
    Copy the full SHA
    e405432 View commit details
    Browse the repository at this point in the history

  2. update dockerfile 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 24, 2024
    Configuration menu
    Copy the full SHA
    07cfa67 View commit details
    Browse the repository at this point in the history

Commits on Apr 26, 2024

  1. add dockerfileg 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 26, 2024
    Configuration menu
    Copy the full SHA
    b2dfee0 View commit details
    Browse the repository at this point in the history

Commits on Apr 29, 2024

  1. fix trtllm patch 

    Signed-off-by: jiemingz <=>
    jiemingz committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    63cd6b3 View commit details
    Browse the repository at this point in the history

  2. clamp output with warning 

    Signed-off-by: jiemingz <=>
    jiemingz committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    6901348 View commit details
    Browse the repository at this point in the history

  3. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    74a0bb1 View commit details
    Browse the repository at this point in the history

Commits on Apr 30, 2024

  1. remove debug statements 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Apr 30, 2024
    Configuration menu
    Copy the full SHA
    b6a05fd View commit details
    Browse the repository at this point in the history

  2. Merge remote-tracking branch 'origin/main' into aligner_trt_build

    @gshennvm
    gshennvm committed Apr 30, 2024
    Configuration menu
    Copy the full SHA
    db2701b View commit details
    Browse the repository at this point in the history

Commits on May 6, 2024

  1. add debug info 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 6, 2024
    Configuration menu
    Copy the full SHA
    8dd5c59 View commit details
    Browse the repository at this point in the history

  2. bump pytrition version 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 6, 2024
    Configuration menu
    Copy the full SHA
    b5d6f88 View commit details
    Browse the repository at this point in the history

Commits on May 7, 2024

  1. add critic speed 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 7, 2024
    Configuration menu
    Copy the full SHA
    5464827 View commit details
    Browse the repository at this point in the history

  2. critic speedup 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 7, 2024
    Configuration menu
    Copy the full SHA
    00e4298 View commit details
    Browse the repository at this point in the history

  3. Merge remote-tracking branch 'origin/geshen/critic_refactor' into ges… 

    …hen/debug_critic
    @gshennvm
    gshennvm committed May 7, 2024
    Configuration menu
    Copy the full SHA
    fe6864b View commit details
    Browse the repository at this point in the history

Commits on May 13, 2024

  1. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 13, 2024
    Configuration menu
    Copy the full SHA
    80579ec View commit details
    Browse the repository at this point in the history

Commits on May 14, 2024

  1. add pad sequence length 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 14, 2024
    Configuration menu
    Copy the full SHA
    f81f55a View commit details
    Browse the repository at this point in the history

Commits on May 15, 2024

  1. dockerfile 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 15, 2024
    Configuration menu
    Copy the full SHA
    4a034f4 View commit details
    Browse the repository at this point in the history

Commits on May 16, 2024

  1. higher stability 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 16, 2024
    Configuration menu
    Copy the full SHA
    66b5a54 View commit details
    Browse the repository at this point in the history

  2. Merge remote-tracking branch 'origin/main' into geshen/debug_critic

    @gshennvm
    gshennvm committed May 16, 2024
    Configuration menu
    Copy the full SHA
    7841381 View commit details
    Browse the repository at this point in the history

  3. add 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 16, 2024
    Configuration menu
    Copy the full SHA
    1779c51 View commit details
    Browse the repository at this point in the history

Commits on May 24, 2024

  1. add hack for ckpt 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 24, 2024
    Configuration menu
    Copy the full SHA
    e357ef9 View commit details
    Browse the repository at this point in the history

  2. fix conf 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 24, 2024
    Configuration menu
    Copy the full SHA
    6b606e8 View commit details
    Browse the repository at this point in the history

  3. no import 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 24, 2024
    Configuration menu
    Copy the full SHA
    a669837 View commit details
    Browse the repository at this point in the history

  4. add 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 24, 2024
    Configuration menu
    Copy the full SHA
    ada5f45 View commit details
    Browse the repository at this point in the history

Commits on May 25, 2024

  1. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 25, 2024
    Configuration menu
    Copy the full SHA
    393acc6 View commit details
    Browse the repository at this point in the history

  2. run through 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 25, 2024
    Configuration menu
    Copy the full SHA
    b6a4d59 View commit details
    Browse the repository at this point in the history

  3. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 25, 2024
    Configuration menu
    Copy the full SHA
    977e6e7 View commit details
    Browse the repository at this point in the history

  4. adaptive 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 25, 2024
    Configuration menu
    Copy the full SHA
    621718d View commit details
    Browse the repository at this point in the history

  5. output tensor 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 25, 2024
    Configuration menu
    Copy the full SHA
    6109b8b View commit details
    Browse the repository at this point in the history

Commits on May 26, 2024

  1. add logging 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 26, 2024
    Configuration menu
    Copy the full SHA
    866c22b View commit details
    Browse the repository at this point in the history

Commits on May 28, 2024

  1. fix for llama3 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 28, 2024
    Configuration menu
    Copy the full SHA
    02aa2b8 View commit details
    Browse the repository at this point in the history

Commits on May 31, 2024

  1. disable last checkpoint 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed May 31, 2024
    Configuration menu
    Copy the full SHA
    e6f27c5 View commit details
    Browse the repository at this point in the history

Commits on Jun 1, 2024

  1. fix padding bug 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 1, 2024
    Configuration menu
    Copy the full SHA
    c689d2a View commit details
    Browse the repository at this point in the history

  2. add critic warmup 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 1, 2024
    Configuration menu
    Copy the full SHA
    cd4aaa5 View commit details
    Browse the repository at this point in the history

Commits on Jun 8, 2024

  1. Revert "add" 

    This reverts commit 1779c51.
    @gshennvm
    gshennvm committed Jun 8, 2024
    Configuration menu
    Copy the full SHA
    993e358 View commit details
    Browse the repository at this point in the history

  2. Merge remote-tracking branch 'origin/main' into geshen/trt_llm_to_main

    @gshennvm
    gshennvm committed Jun 8, 2024
    Configuration menu
    Copy the full SHA
    28fcaf3 View commit details
    Browse the repository at this point in the history

Commits on Jun 11, 2024

  1. fix module missing bug 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 11, 2024
    Configuration menu
    Copy the full SHA
    ef347e5 View commit details
    Browse the repository at this point in the history

  2. Ensure critic server does not squeeze out a singleton batch dim (#199) 

    * Make sure critic does not squeeze out a batch_dim=1

Signed-off-by: Terry Kong <[email protected]>

* remove breakpoints

Signed-off-by: Terry Kong <[email protected]>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Remove references to non-nemo samplers

Signed-off-by: Terry Kong <[email protected]>

---------

Signed-off-by: Terry Kong <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
    @terrykong @pre-commit-ci
    terrykong and pre-commit-ci[bot] authored Jun 11, 2024
    Configuration menu
    Copy the full SHA
    752d0bd View commit details
    Browse the repository at this point in the history

Commits on Jun 12, 2024

  1. Merge branch 'geshen/llama3_rlhf' into geshen/trt_llm_to_main

    @gshennvm
    gshennvm committed Jun 12, 2024
    Configuration menu
    Copy the full SHA
    78e6536 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'geshen/trt_llm_to_main' of github.com:NVIDIA/NeMo-Align… 

    …er into geshen/trt_llm_to_main
    @gshennvm
    gshennvm committed Jun 12, 2024
    Configuration menu
    Copy the full SHA
    4de3eeb View commit details
    Browse the repository at this point in the history

  3. TRTLLM PP wrong format WAR 

    Signed-off-by: Jimmy Zhang <[email protected]>
    @jiemingz @gshennvm
    jiemingz authored and gshennvm committed Jun 12, 2024
    Configuration menu
    Copy the full SHA
    8a39881 View commit details
    Browse the repository at this point in the history

  4. docker file branch 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 12, 2024
    Configuration menu
    Copy the full SHA
    666e969 View commit details
    Browse the repository at this point in the history

  5. fix config 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 12, 2024
    Configuration menu
    Copy the full SHA
    3bec1bc View commit details
    Browse the repository at this point in the history

  6. remove prints 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 12, 2024
    Configuration menu
    Copy the full SHA
    3e7ca5f View commit details
    Browse the repository at this point in the history

  7. remove print 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 12, 2024
    Configuration menu
    Copy the full SHA
    12a0aae View commit details
    Browse the repository at this point in the history

  8. remove unneeded statement 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 12, 2024
    Configuration menu
    Copy the full SHA
    3956b6d View commit details
    Browse the repository at this point in the history

  9. no save topk 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 12, 2024
    Configuration menu
    Copy the full SHA
    e090663 View commit details
    Browse the repository at this point in the history

Commits on Jun 24, 2024

  1. Merge remote-tracking branch 'origin/main' into geshen/trt_llm_to_main

    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    af83947 View commit details
    Browse the repository at this point in the history

  2. [pre-commit.ci] auto fixes from pre-commit.com hooks 

    for more information, see https://pre-commit.ci
    @pre-commit-ci
    pre-commit-ci[bot] committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    cc03b76 View commit details
    Browse the repository at this point in the history

  3. critic speedup 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    605bda1 View commit details
    Browse the repository at this point in the history

  4. [pre-commit.ci] auto fixes from pre-commit.com hooks 

    for more information, see https://pre-commit.ci
    @pre-commit-ci
    pre-commit-ci[bot] committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    b3dedfd View commit details
    Browse the repository at this point in the history

  5. fix

    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    9fb90ff View commit details
    Browse the repository at this point in the history

  6. Merge branch 'geshen/critic_speedup' of github.com:NVIDIA/NeMo-Aligne… 

    …r into geshen/critic_speedup
    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    bf62bcc View commit details
    Browse the repository at this point in the history

  7. [pre-commit.ci] auto fixes from pre-commit.com hooks 

    for more information, see https://pre-commit.ci
    @pre-commit-ci
    pre-commit-ci[bot] committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    aea50ad View commit details
    Browse the repository at this point in the history

  8. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    0a9416e View commit details
    Browse the repository at this point in the history

  9. Merge branch 'geshen/critic_speedup' of github.com:NVIDIA/NeMo-Aligne… 

    …r into geshen/critic_speedup
    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    41ffeb5 View commit details
    Browse the repository at this point in the history

  10. add parallel_state

    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    e3d85bf View commit details
    Browse the repository at this point in the history

  11. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    1225041 View commit details
    Browse the repository at this point in the history

  12. rename 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    5315fc5 View commit details
    Browse the repository at this point in the history

  13. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    9c6141b View commit details
    Browse the repository at this point in the history

  14. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 24, 2024
    Configuration menu
    Copy the full SHA
    2cef93e View commit details
    Browse the repository at this point in the history

Commits on Jun 25, 2024

  1. pull changes from degert/spin-trt-beta (#220) 

    * pull changes from degert/spin-trt-beta

* add back import

Signed-off-by: Gerald Shen <[email protected]>

---------

Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm authored Jun 25, 2024
    Configuration menu
    Copy the full SHA
    f5bd8c5 View commit details
    Browse the repository at this point in the history

  2. add text input support 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 25, 2024
    Configuration menu
    Copy the full SHA
    cbd7095 View commit details
    Browse the repository at this point in the history

  3. clean up location of tokenization 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 25, 2024
    Configuration menu
    Copy the full SHA
    c72e4c1 View commit details
    Browse the repository at this point in the history

  4. remove useless imports 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 25, 2024
    Configuration menu
    Copy the full SHA
    220cf17 View commit details
    Browse the repository at this point in the history

Commits on Jun 26, 2024

  1. refactor rm server 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 26, 2024
    Configuration menu
    Copy the full SHA
    1ed3c46 View commit details
    Browse the repository at this point in the history

  2. remove run rm file 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 26, 2024
    Configuration menu
    Copy the full SHA
    aa4249d View commit details
    Browse the repository at this point in the history

  3. remove preferred batch size logic 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 26, 2024
    Configuration menu
    Copy the full SHA
    22615b7 View commit details
    Browse the repository at this point in the history

  4. add comment 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 26, 2024
    Configuration menu
    Copy the full SHA
    95a9507 View commit details
    Browse the repository at this point in the history

  5. allow users to specify their own preferred batch size 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 26, 2024
    Configuration menu
    Copy the full SHA
    0fdb124 View commit details
    Browse the repository at this point in the history

  6. add comments and add changelog 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 26, 2024
    Configuration menu
    Copy the full SHA
    031287c View commit details
    Browse the repository at this point in the history

  7. update changelog 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 26, 2024
    Configuration menu
    Copy the full SHA
    fa37930 View commit details
    Browse the repository at this point in the history

  8. remove old reward model callable 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 26, 2024
    Configuration menu
    Copy the full SHA
    c83a98d View commit details
    Browse the repository at this point in the history

  9. inference should be done with collect loss data =True

    @gshennvm
    gshennvm committed Jun 26, 2024
    Configuration menu
    Copy the full SHA
    a9f722a View commit details
    Browse the repository at this point in the history

Commits on Jun 27, 2024

  1. Merge branch 'geshen/critic_speedup' into geshen/trt_llm_to_main 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 27, 2024
    Configuration menu
    Copy the full SHA
    15460d2 View commit details
    Browse the repository at this point in the history

  2. fix issues with merge 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 27, 2024
    Configuration menu
    Copy the full SHA
    f6c09ea View commit details
    Browse the repository at this point in the history

  3. cleanup configs 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 27, 2024
    Configuration menu
    Copy the full SHA
    1d85152 View commit details
    Browse the repository at this point in the history

  4. add strip to sequence length 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 27, 2024
    Configuration menu
    Copy the full SHA
    b3078c5 View commit details
    Browse the repository at this point in the history

  5. Merge remote-tracking branch 'origin/geshen/critic_speedup' into gesh… 

    …en/trt_llm_to_main
    @gshennvm
    gshennvm committed Jun 27, 2024
    Configuration menu
    Copy the full SHA
    1eb0823 View commit details
    Browse the repository at this point in the history

  6. change

    @gshennvm
    gshennvm committed Jun 27, 2024
    Configuration menu
    Copy the full SHA
    6203e90 View commit details
    Browse the repository at this point in the history

  7. fix

    @gshennvm
    gshennvm committed Jun 27, 2024
    Configuration menu
    Copy the full SHA
    e51c45f View commit details
    Browse the repository at this point in the history

Commits on Jun 28, 2024

  1. clean actor 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 28, 2024
    Configuration menu
    Copy the full SHA
    1b220f1 View commit details
    Browse the repository at this point in the history

  2. backwards compatibility in actor

    @gshennvm
    gshennvm committed Jun 28, 2024
    Configuration menu
    Copy the full SHA
    0815b56 View commit details
    Browse the repository at this point in the history

  3. Apply suggestions from code review 

    Co-authored-by: Olivier Delalleau <[email protected]>
    @gshennvm @odelalleau
    gshennvm and odelalleau authored Jun 28, 2024
    Configuration menu
    Copy the full SHA
    8d75cbf View commit details
    Browse the repository at this point in the history

  4. modify changelog

    @gshennvm
    gshennvm committed Jun 28, 2024
    Configuration menu
    Copy the full SHA
    ab4c549 View commit details
    Browse the repository at this point in the history

  5. fixup! modify changelog 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jun 28, 2024
    Configuration menu
    Copy the full SHA
    82fb3a1 View commit details
    Browse the repository at this point in the history

  6. add comments to ppo_critic config

    @gshennvm
    gshennvm committed Jun 28, 2024
    Configuration menu
    Copy the full SHA
    12f85d2 View commit details
    Browse the repository at this point in the history

  7. add note on breaking change in inference rm

    @gshennvm
    gshennvm committed Jun 28, 2024
    Configuration menu
    Copy the full SHA
    f90f4d6 View commit details
    Browse the repository at this point in the history

  8. change inference mbs to 4

    @gshennvm
    gshennvm committed Jun 28, 2024
    Configuration menu
    Copy the full SHA
    527557d View commit details
    Browse the repository at this point in the history

  9. add comments for inference rm config

    @gshennvm
    gshennvm committed Jun 28, 2024
    Configuration menu
    Copy the full SHA
    fe9d288 View commit details
    Browse the repository at this point in the history

  10. revert gbs flag back to previous in ppo critic

    @gshennvm
    gshennvm committed Jun 28, 2024
    Configuration menu
    Copy the full SHA
    f3124d3 View commit details
    Browse the repository at this point in the history

  11. delete unused variable

    @gshennvm
    gshennvm committed Jun 28, 2024
    Configuration menu
    Copy the full SHA
    efadcae View commit details
    Browse the repository at this point in the history

  12. Update nemo_aligner/algorithms/critic_server_trainer.py 

    Co-authored-by: Olivier Delalleau <[email protected]>
    @gshennvm @odelalleau
    gshennvm and odelalleau authored Jun 28, 2024
    Configuration menu
    Copy the full SHA
    35a2895 View commit details
    Browse the repository at this point in the history

Commits on Jun 29, 2024

  1. remove add_eos arg, and update attribute annotate script

    @gshennvm
    gshennvm committed Jun 29, 2024
    Configuration menu
    Copy the full SHA
    4487932 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'geshen/critic_speedup' of github.com:NVIDIA/NeMo-Aligne… 

    …r into geshen/critic_speedup
    @gshennvm
    gshennvm committed Jun 29, 2024
    Configuration menu
    Copy the full SHA
    7e7f27b View commit details
    Browse the repository at this point in the history

  3. no mutation on inputs when processing them for inference

    @gshennvm
    gshennvm committed Jun 29, 2024
    Configuration menu
    Copy the full SHA
    e9c7b39 View commit details
    Browse the repository at this point in the history

  4. fix bug when padding

    @gshennvm
    gshennvm committed Jun 29, 2024
    Configuration menu
    Copy the full SHA
    c6f6da4 View commit details
    Browse the repository at this point in the history

  5. add comment for forward_micro_batch_size in training_rm.yaml

    @gshennvm
    gshennvm committed Jun 29, 2024
    Configuration menu
    Copy the full SHA
    ebb69f4 View commit details
    Browse the repository at this point in the history

  6. change non_blocking to use sync

    @gshennvm
    gshennvm committed Jun 29, 2024
    Configuration menu
    Copy the full SHA
    2775e81 View commit details
    Browse the repository at this point in the history

Commits on Jun 30, 2024

  1. Merge branch 'geshen/critic_speedup' into geshen/trt_llm_to_main

    @gshennvm
    gshennvm committed Jun 30, 2024
    Configuration menu
    Copy the full SHA
    fe0399f View commit details
    Browse the repository at this point in the history

Commits on Jul 1, 2024

  1. nemo export api changes 

    Signed-off-by: Jimmy Zhang <[email protected]>
    @jiemingz @gshennvm
    jiemingz authored and gshennvm committed Jul 1, 2024
    Configuration menu
    Copy the full SHA
    ffa253f View commit details
    Browse the repository at this point in the history

  2. upgrade to newer nemo export 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 1, 2024
    Configuration menu
    Copy the full SHA
    7ca9e34 View commit details
    Browse the repository at this point in the history

  3. fix imports 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 1, 2024
    Configuration menu
    Copy the full SHA
    8181168 View commit details
    Browse the repository at this point in the history

  4. Communicator hang fix in the actor loop (#200) 

    * Communicator hang fix in the actor loop

Signed-off-by: Terry Kong <[email protected]>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Signed-off-by: Terry Kong <[email protected]>

* Update examples/nlp/gpt/train_gpt_ppo_actor.py

Co-authored-by: Gerald Shen <[email protected]>
Signed-off-by: Terry Kong <[email protected]>

---------

Signed-off-by: Terry Kong <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Gerald Shen <[email protected]>
    @terrykong @pre-commit-ci @gshennvm
    3 people authored Jul 1, 2024
    Configuration menu
    Copy the full SHA
    4d0853d View commit details
    Browse the repository at this point in the history

Commits on Jul 3, 2024

  1. add nemo guard for when things don't stop properly 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 3, 2024
    Configuration menu
    Copy the full SHA
    ec548b8 View commit details
    Browse the repository at this point in the history

  2. cleanup communicator clean 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 3, 2024
    Configuration menu
    Copy the full SHA
    ce7a07f View commit details
    Browse the repository at this point in the history

  3. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 3, 2024
    Configuration menu
    Copy the full SHA
    bb2fc48 View commit details
    Browse the repository at this point in the history

Commits on Jul 11, 2024

  1. critic speedup 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 11, 2024
    Configuration menu
    Copy the full SHA
    606f690 View commit details
    Browse the repository at this point in the history

  2. middle of PP should be broadcasted as well 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 11, 2024
    Configuration menu
    Copy the full SHA
    f48dc29 View commit details
    Browse the repository at this point in the history

  3. update with critic changes 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 11, 2024
    Configuration menu
    Copy the full SHA
    708bc24 View commit details
    Browse the repository at this point in the history

  4. Merge remote-tracking branch 'origin/main' into geshen/trt_llm_to_main

    @gshennvm
    gshennvm committed Jul 11, 2024
    Configuration menu
    Copy the full SHA
    48ad685 View commit details
    Browse the repository at this point in the history

  5. general cleanup 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 11, 2024
    Configuration menu
    Copy the full SHA
    d053475 View commit details
    Browse the repository at this point in the history

  6. add checker for trt 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 11, 2024
    Configuration menu
    Copy the full SHA
    0b4a92d View commit details
    Browse the repository at this point in the history

  7. remove comments 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 11, 2024
    Configuration menu
    Copy the full SHA
    b72a5ec View commit details
    Browse the repository at this point in the history

Commits on Jul 12, 2024

  1. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    984acaa View commit details
    Browse the repository at this point in the history

  2. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    c11e1d7 View commit details
    Browse the repository at this point in the history

  3. another fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    14a9926 View commit details
    Browse the repository at this point in the history

  4. add typing 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    7c2fc3e View commit details
    Browse the repository at this point in the history

  5. cleanup 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    fe02867 View commit details
    Browse the repository at this point in the history

  6. ppo trainer should use stop and get time 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    02ad2fa View commit details
    Browse the repository at this point in the history

  7. add some comments 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    24c53be View commit details
    Browse the repository at this point in the history

  8. critic warmup should have good default 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    8a25e5e View commit details
    Browse the repository at this point in the history

  9. added ppo in changelog 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    24f138a View commit details
    Browse the repository at this point in the history

  10. add comments 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    9c72c53 View commit details
    Browse the repository at this point in the history

  11. Avoids crash in PPOTrainer if using adafactor w/o learning rate (#234) 

    Signed-off-by: Terry Kong <[email protected]>
    @terrykong
    terrykong authored Jul 12, 2024
    Configuration menu
    Copy the full SHA
    5ed9cd8 View commit details
    Browse the repository at this point in the history

  12. rename 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    8b6627a View commit details
    Browse the repository at this point in the history

  13. Merge branch 'geshen/trt_llm_to_main' of github.com:NVIDIA/NeMo-Align… 

    …er into geshen/trt_llm_to_main
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    56032c8 View commit details
    Browse the repository at this point in the history

  14. Raise exceptions if using trtllm and use_Greedy in sampling params is… 

    … set but temperature/topk are not correct for greedy (#232)

* Raise exceptions if using trtllm and use_Greedy in sampling params is set but temperature/topk are not correct for greedy

Signed-off-by: Terry Kong <[email protected]>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Signed-off-by: Terry Kong <[email protected]>

* instead of hard erroring, send warnings and then override

Signed-off-by: Terry Kong <[email protected]>

---------

Signed-off-by: Terry Kong <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
    @terrykong @pre-commit-ci
    terrykong and pre-commit-ci[bot] authored Jul 12, 2024
    Configuration menu
    Copy the full SHA
    1e17f8b View commit details
    Browse the repository at this point in the history

  15. fix bugs 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 12, 2024
    Configuration menu
    Copy the full SHA
    e0a94d0 View commit details
    Browse the repository at this point in the history

Commits on Jul 13, 2024

  1. cleanup pad id handling when PP > 1 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 13, 2024
    Configuration menu
    Copy the full SHA
    835b3b3 View commit details
    Browse the repository at this point in the history

Commits on Jul 14, 2024

  1. fix issue with PP > 1 check

    @gshennvm
    gshennvm committed Jul 14, 2024
    Configuration menu
    Copy the full SHA
    2b95331 View commit details
    Browse the repository at this point in the history

  2. add is_end logic 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 14, 2024
    Configuration menu
    Copy the full SHA
    261269a View commit details
    Browse the repository at this point in the history

  3. add is end logic 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 14, 2024
    Configuration menu
    Copy the full SHA
    83ba660 View commit details
    Browse the repository at this point in the history

  4. add is end logic 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 14, 2024
    Configuration menu
    Copy the full SHA
    f3912e7 View commit details
    Browse the repository at this point in the history

  5. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 14, 2024
    Configuration menu
    Copy the full SHA
    09d2783 View commit details
    Browse the repository at this point in the history

  6. fix 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 14, 2024
    Configuration menu
    Copy the full SHA
    5105ed9 View commit details
    Browse the repository at this point in the history

Commits on Jul 15, 2024

  1. fix another bug 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 15, 2024
    Configuration menu
    Copy the full SHA
    a2bf8a0 View commit details
    Browse the repository at this point in the history

  2. Merge remote-tracking branch 'origin/main' into geshen/trt_llm_to_main

    @gshennvm
    gshennvm committed Jul 15, 2024
    Configuration menu
    Copy the full SHA
    d9d45d6 View commit details
    Browse the repository at this point in the history

  3. update changelog 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 15, 2024
    Configuration menu
    Copy the full SHA
    f00d09e View commit details
    Browse the repository at this point in the history

  4. change rc version 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 15, 2024
    Configuration menu
    Copy the full SHA
    d52eab2 View commit details
    Browse the repository at this point in the history

Commits on Jul 16, 2024

  1. Update the hash of the conversion script to include TE fix for mcore … 

    …conversion (#238)

Signed-off-by: Terry Kong <[email protected]>
    @terrykong
    terrykong authored Jul 16, 2024
    Configuration menu
    Copy the full SHA
    12fa459 View commit details
    Browse the repository at this point in the history

  2. update docs 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 16, 2024
    Configuration menu
    Copy the full SHA
    3d0a650 View commit details
    Browse the repository at this point in the history

Commits on Jul 22, 2024

  1. Addresses documentation bugs (#244) 

    Signed-off-by: Terry Kong <[email protected]>
    @terrykong
    terrykong authored Jul 22, 2024
    Configuration menu
    Copy the full SHA
    e103759 View commit details
    Browse the repository at this point in the history

Commits on Jul 23, 2024

  1. Updates microbatch APIs to use megatron's instead of apex's (#241) 

    * Updates microbatch APIs to use megatron's instead of apex's

Signed-off-by: Terry Kong <[email protected]>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* take off protected prefix

Signed-off-by: Terry Kong <[email protected]>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Signed-off-by: Terry Kong <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
    @terrykong @pre-commit-ci
    terrykong and pre-commit-ci[bot] authored Jul 23, 2024
    Configuration menu
    Copy the full SHA
    db3a9cc View commit details
    Browse the repository at this point in the history

Commits on Jul 31, 2024

  1. add pref 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Jul 31, 2024
    Configuration menu
    Copy the full SHA
    7dddeef View commit details
    Browse the repository at this point in the history

Commits on Aug 2, 2024

  1. fix HF links in readme (#248) 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm @terrykong
    gshennvm authored and terrykong committed Aug 2, 2024
    Configuration menu
    Copy the full SHA
    cc26456 View commit details
    Browse the repository at this point in the history

  2. Point the TRTLLM documentation to Nemo docs (#257) 

    * Points to the documentation to the 0.4 branch

* update documentation to point to nemo docs

Signed-off-by: Terry Kong <[email protected]>

---------

Signed-off-by: Terry Kong <[email protected]>
    @terrykong
    terrykong authored Aug 2, 2024
    Configuration menu
    Copy the full SHA
    44701e5 View commit details
    Browse the repository at this point in the history

  3. add scaling efficiency instead 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Aug 2, 2024
    Configuration menu
    Copy the full SHA
    a1f99fd View commit details
    Browse the repository at this point in the history

  4. fix ending 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Aug 2, 2024
    Configuration menu
    Copy the full SHA
    e741e76 View commit details
    Browse the repository at this point in the history

  5. fix logic 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Aug 2, 2024
    Configuration menu
    Copy the full SHA
    e4ff1ec View commit details
    Browse the repository at this point in the history

Commits on Aug 5, 2024

  1. 0.4 doc tech edit (#260) 

    * Revert "add pref"

This reverts commit 7dddeef.

Signed-off-by: Terry Kong <[email protected]>

* Revert "update docs"

This reverts commit 3d0a650.

Signed-off-by: Terry Kong <[email protected]>

* revert steerlm2 docs

Signed-off-by: Terry Kong <[email protected]>

* revert rpo/ipo docs

Signed-off-by: Terry Kong <[email protected]>

* revert trtllm doc change on frontpage

Signed-off-by: Terry Kong <[email protected]>

* Update CHANGELOG.md to 0.4.0

Signed-off-by: Terry Kong <[email protected]>

* remove aview(-1)

Signed-off-by: Terry Kong <[email protected]>

* Addressing #260 (comment)

Signed-off-by: Terry Kong <[email protected]>

* apply next batch of suggestions

Signed-off-by: Terry Kong <[email protected]>

* applied suggestions in #260 (comment)

Signed-off-by: Terry Kong <[email protected]>

* appliying suggestions in #260 (comment)

Signed-off-by: Terry Kong <[email protected]>

* Update dpo.rst

Provided copyedits on file.

Signed-off-by: Terry Kong <[email protected]>

* Update index.rst

Added copyedits to the file.

Signed-off-by: Terry Kong <[email protected]>

* Update rlhf.rst

Copyedited the file.

Signed-off-by: Terry Kong <[email protected]>

* one space

Signed-off-by: Terry Kong <[email protected]>

* revert steerlm2 deletion

Signed-off-by: Terry Kong <[email protected]>

* restore rpo changes

Signed-off-by: Terry Kong <[email protected]>

* restore missing regression comment

Signed-off-by: Terry Kong <[email protected]>

* restore trtllm changeswq

Signed-off-by: Terry Kong <[email protected]>

* restore sterrlm2 in index.rst

Signed-off-by: Terry Kong <[email protected]>

* restore lost spacing

Signed-off-by: Terry Kong <[email protected]>

---------

Signed-off-by: Terry Kong <[email protected]>
Co-authored-by: jgerh <[email protected]>
    @terrykong @jgerh
    terrykong and jgerh authored Aug 5, 2024
    Configuration menu
    Copy the full SHA
    46afc6b View commit details
    Browse the repository at this point in the history

  2. Restore missing newline in performance table 

    Signed-off-by: Terry Kong <[email protected]>
    @terrykong
    terrykong committed Aug 5, 2024
    Configuration menu
    Copy the full SHA
    9d5c71f View commit details
    Browse the repository at this point in the history

Commits on Aug 6, 2024

  1. fix rlhf perf table column headings 

    Signed-off-by: Terry Kong <[email protected]>
    @terrykong
    terrykong committed Aug 6, 2024
    Configuration menu
    Copy the full SHA
    928d4a9 View commit details
    Browse the repository at this point in the history

  2. Updates MBS calculator APIs to reconfigure_num_microbatches_calculator 

    Which was introduced in this mcore commit: NVIDIA/Megatron-LM@32002bb

Signed-off-by: Terry Kong <[email protected]>
    @terrykong
    terrykong committed Aug 6, 2024
    Configuration menu
    Copy the full SHA
    384394f View commit details
    Browse the repository at this point in the history

Commits on Aug 20, 2024

  1. call 

    Signed-off-by: Gerald Shen <[email protected]>
    @gshennvm
    gshennvm committed Aug 20, 2024
    Configuration menu
    Copy the full SHA
    7eb9a97 View commit details
    Browse the repository at this point in the history