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

Add HQQ quantization support #29637

Merged
merged 78 commits into from
May 2, 2024
Merged

Add HQQ quantization support #29637

merged 78 commits into from
May 2, 2024

Commits on Apr 24, 2024

  1. update HQQ transformers integration

    @mobicham
    mobicham committed Apr 24, 2024
    Configuration menu
    Copy the full SHA
    bbc68fe View commit details
    Browse the repository at this point in the history

  2. Merge branch 'huggingface:main' into stable

    @mobicham
    mobicham committed Apr 24, 2024
    Configuration menu
    Copy the full SHA
    2a1f224 View commit details
    Browse the repository at this point in the history

  3. push import_utils.py

    @mobicham
    mobicham committed Apr 24, 2024
    Configuration menu
    Copy the full SHA
    e1e5df6 View commit details
    Browse the repository at this point in the history

  4. add force_hooks check in modeling_utils.py

    @mobicham
    mobicham committed Apr 24, 2024
    Configuration menu
    Copy the full SHA
    0192b03 View commit details
    Browse the repository at this point in the history

  5. fix | with Optional

    @mobicham
    mobicham committed Apr 24, 2024
    Configuration menu
    Copy the full SHA
    823de37 View commit details
    Browse the repository at this point in the history

  6. force bias as param

    @mobicham
    mobicham committed Apr 24, 2024
    Configuration menu
    Copy the full SHA
    08d7b8e View commit details
    Browse the repository at this point in the history

  7. check bias is Tensor

    @mobicham
    mobicham committed Apr 24, 2024
    Configuration menu
    Copy the full SHA
    e1fa6c9 View commit details
    Browse the repository at this point in the history

  8. force forward for multi-gpu

    @mobicham
    mobicham committed Apr 24, 2024
    Configuration menu
    Copy the full SHA
    6e854ca View commit details
    Browse the repository at this point in the history

Commits on Apr 25, 2024

  1. review fixes pass

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    2b9f271 View commit details
    Browse the repository at this point in the history

  2. remove torch grad()

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    5bb9ca2 View commit details
    Browse the repository at this point in the history

  3. if any key in linear_tags fix

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    392e7c5 View commit details
    Browse the repository at this point in the history

  4. add cpu/disk check

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    20f9ad5 View commit details
    Browse the repository at this point in the history

  5. isinstance return

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    3a5679a View commit details
    Browse the repository at this point in the history

  6. add multigpu test + refactor tests

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    7a1bbca View commit details
    Browse the repository at this point in the history

  7. clean hqq_utils imports in hqq.py

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    65b2887 View commit details
    Browse the repository at this point in the history

  8. clean hqq_utils imports in quantizer_hqq.py

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    bba74cd View commit details
    Browse the repository at this point in the history

  9. delete hqq_utils.py

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    de88c2a View commit details
    Browse the repository at this point in the history

  10. Delete src/transformers/utils/hqq_utils.py

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    651a586 View commit details
    Browse the repository at this point in the history

  11. ruff init

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    d07ea85 View commit details
    Browse the repository at this point in the history

  12. remove torch.float16 from __init__ in test

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    dedf69e View commit details
    Browse the repository at this point in the history

  13. refactor test

    @mobicham
    mobicham committed Apr 25, 2024
    Configuration menu
    Copy the full SHA
    0edf8a4 View commit details
    Browse the repository at this point in the history

Commits on Apr 26, 2024

  1. isinstance -> type in quantizer_hqq.py

    @mobicham
    mobicham committed Apr 26, 2024
    Configuration menu
    Copy the full SHA
    c7ec123 View commit details
    Browse the repository at this point in the history

Commits on Apr 29, 2024

  1. cpu/disk device_map check in quantizer_hqq.py

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    5283ac2 View commit details
    Browse the repository at this point in the history

  2. remove type(module) nn.linear check in quantizer_hqq.py

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    15daeb4 View commit details
    Browse the repository at this point in the history

  3. add BaseQuantizeConfig import inside HqqConfig init

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    bc4bc73 View commit details
    Browse the repository at this point in the history

  4. remove hqq import in hqq.py

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    b54e87b View commit details
    Browse the repository at this point in the history

  5. remove accelerate import from test_hqq.py

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    0f9698a View commit details
    Browse the repository at this point in the history

  6. quant config.py doc update

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    d31837f View commit details
    Browse the repository at this point in the history

  7. add hqqconfig to main_classes doc

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    b8f792c View commit details
    Browse the repository at this point in the history

  8. Merge branch 'huggingface:main' into stable

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    8b84cb1 View commit details
    Browse the repository at this point in the history

  9. make style

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    9a061e5 View commit details
    Browse the repository at this point in the history

  10. __init__ fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    8612282 View commit details
    Browse the repository at this point in the history

  11. ruff __init__

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    b786793 View commit details
    Browse the repository at this point in the history

  12. skip_modules list

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    e7ba717 View commit details
    Browse the repository at this point in the history

  13. hqqconfig format fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    3a38f21 View commit details
    Browse the repository at this point in the history

  14. hqqconfig doc fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    9eee213 View commit details
    Browse the repository at this point in the history

  15. hqqconfig doc fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    03cc8e6 View commit details
    Browse the repository at this point in the history

  16. hqqconfig doc fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    96bd141 View commit details
    Browse the repository at this point in the history

  17. hqqconfig doc fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    713d226 View commit details
    Browse the repository at this point in the history

  18. hqqconfig doc fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    dad9a60 View commit details
    Browse the repository at this point in the history

  19. hqqconfig doc fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    67c0985 View commit details
    Browse the repository at this point in the history

  20. hqqconfig doc fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    94c393a View commit details
    Browse the repository at this point in the history

  21. hqqconfig doc fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    35fc9f5 View commit details
    Browse the repository at this point in the history

  22. hqqconfig doc fix

    @mobicham
    mobicham committed Apr 29, 2024
    Configuration menu
    Copy the full SHA
    06f6497 View commit details
    Browse the repository at this point in the history

Commits on Apr 30, 2024

  1. test_hqq.py remove mistral comment

    @mobicham
    mobicham committed Apr 30, 2024
    Configuration menu
    Copy the full SHA
    25fde9c View commit details
    Browse the repository at this point in the history

  2. remove self.using_multi_gpu is False

    @mobicham
    mobicham committed Apr 30, 2024
    Configuration menu
    Copy the full SHA
    ee50516 View commit details
    Browse the repository at this point in the history

  3. torch_dtype default val set and logger.info

    @mobicham
    mobicham committed Apr 30, 2024
    Configuration menu
    Copy the full SHA
    01d798a View commit details
    Browse the repository at this point in the history

Commits on May 2, 2024

  1. hqq.py isinstance fix

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    a909ca8 View commit details
    Browse the repository at this point in the history

  2. remove torch=None

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    c466c89 View commit details
    Browse the repository at this point in the history

  3. torch_device test_hqq

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    d522fed View commit details
    Browse the repository at this point in the history

  4. rename test_hqq

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    a09e90f View commit details
    Browse the repository at this point in the history

  5. MODEL_ID in test_hqq

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    5bdf40f View commit details
    Browse the repository at this point in the history

  6. quantizer_hqq setattr fix

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    e693d47 View commit details
    Browse the repository at this point in the history

  7. quantizer_hqq typo fix

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    f5cabe5 View commit details
    Browse the repository at this point in the history

  8. imports quantizer_hqq.py

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    5ede086 View commit details
    Browse the repository at this point in the history

  9. isinstance quantizer_hqq

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    c86000b View commit details
    Browse the repository at this point in the history

  10. hqq_layer.bias reformat quantizer_hqq

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    7d3e083 View commit details
    Browse the repository at this point in the history

  11. Step 2 as comment in quantizer_hqq

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    082dfea View commit details
    Browse the repository at this point in the history

  12. prepare_for_hqq_linear() comment

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    667f1ad View commit details
    Browse the repository at this point in the history

  13. keep_in_fp32_modules fix

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    e0cd784 View commit details
    Browse the repository at this point in the history

  14. HqqHfQuantizer reformat

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    5d3b504 View commit details
    Browse the repository at this point in the history

  15. quantization.md hqqconfig

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    cc1961c View commit details
    Browse the repository at this point in the history

  16. quantization.md model example reformat

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    9aa9e15 View commit details
    Browse the repository at this point in the history

  17. quantization.md # space

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    9273e21 View commit details
    Browse the repository at this point in the history

  18. quantization.md space })

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    f29e7a4 View commit details
    Browse the repository at this point in the history

  19. quantization.md space })

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    5168852 View commit details
    Browse the repository at this point in the history

  20. quantization_config fix doc 

    Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
    @mobicham @amyeroberts
    mobicham and amyeroberts committed May 2, 2024
    Configuration menu
    Copy the full SHA
    0dfe080 View commit details
    Browse the repository at this point in the history

  21. axis value check in quantization_config

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    2934052 View commit details
    Browse the repository at this point in the history

  22. format

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    bc7cf4e View commit details
    Browse the repository at this point in the history

  23. dynamic config explanation

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    d33f944 View commit details
    Browse the repository at this point in the history

  24. quant config method in quantization.md

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    3522f0a View commit details
    Browse the repository at this point in the history

  25. remove shard-level progress

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    cc14c21 View commit details
    Browse the repository at this point in the history

  26. .cuda fix modeling_utils

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    1e81036 View commit details
    Browse the repository at this point in the history

  27. test_hqq fixes

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    ca07f5a View commit details
    Browse the repository at this point in the history

  28. Merge branch 'huggingface:main' into stable

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    4cc776e View commit details
    Browse the repository at this point in the history

  29. make fix-copies

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    3d777ed View commit details
    Browse the repository at this point in the history

  30. Merge branch 'huggingface:main' into stable

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    b808858 View commit details
    Browse the repository at this point in the history

  31. Merge branch 'huggingface:main' into stable

    @mobicham
    mobicham committed May 2, 2024
    Configuration menu
    Copy the full SHA
    5e71139 View commit details
    Browse the repository at this point in the history