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

[Inference] Update fakequant #9140

Merged
merged 25 commits into from
Sep 14, 2024
Merged

[Inference] Update fakequant #9140

merged 25 commits into from
Sep 14, 2024

Commits on Aug 28, 2024

  1. 1. add a8w8(fp8) a8w8c8(int8) quant_type support 

    2. add llama3.1 and qwen2 ptq config
3. update quantization.md
    @lixcli
    lixcli committed Aug 28, 2024
    Configuration menu
    Copy the full SHA
    005f2ad View commit details
    Browse the repository at this point in the history

  2. fix load_quant_model bug

    @lixcli
    lixcli committed Aug 28, 2024
    Configuration menu
    Copy the full SHA
    e56d9c4 View commit details
    Browse the repository at this point in the history

  3. fix load quant bug

    @lixcli
    lixcli committed Aug 28, 2024
    Configuration menu
    Copy the full SHA
    e2b9a49 View commit details
    Browse the repository at this point in the history

  4. update ll/README.md

    @lixcli
    lixcli committed Aug 28, 2024
    Configuration menu
    Copy the full SHA
    d21ace7 View commit details
    Browse the repository at this point in the history

  5. remove useless code

    @lixcli
    lixcli committed Aug 28, 2024
    Configuration menu
    Copy the full SHA
    e89372c View commit details
    Browse the repository at this point in the history

  6. update quant observer config

    @lixcli
    lixcli committed Aug 28, 2024
    Configuration menu
    Copy the full SHA
    e7160d3 View commit details
    Browse the repository at this point in the history

Commits on Aug 29, 2024

  1. Merge branch 'develop' of https://github.com/PaddlePaddle/PaddleNLP i… 

    …nto add_new_fakequant_type
    @lixcli
    lixcli committed Aug 29, 2024
    Configuration menu
    Copy the full SHA
    630b3d6 View commit details
    Browse the repository at this point in the history

  2. resolve wrong modify

    @lixcli
    lixcli committed Aug 29, 2024
    Configuration menu
    Copy the full SHA
    7032bf2 View commit details
    Browse the repository at this point in the history

  3. fix prepare_qconfig

    @lixcli
    lixcli committed Aug 29, 2024
    Configuration menu
    Copy the full SHA
    323c465 View commit details
    Browse the repository at this point in the history

  4. remove unuse files

    @lixcli
    lixcli committed Aug 29, 2024
    Configuration menu
    Copy the full SHA
    df416ac View commit details
    Browse the repository at this point in the history

  5. update quantization.md

    @lixcli
    lixcli committed Aug 29, 2024
    Configuration menu
    Copy the full SHA
    db61a99 View commit details
    Browse the repository at this point in the history

Commits on Aug 30, 2024

  1. reformat quantization.md and argument.py

    @lixcli
    lixcli committed Aug 30, 2024
    Configuration menu
    Copy the full SHA
    f114947 View commit details
    Browse the repository at this point in the history

  2. update prepare data method for ceval ptq

    @lixcli
    lixcli committed Aug 30, 2024
    Configuration menu
    Copy the full SHA
    8d26cb1 View commit details
    Browse the repository at this point in the history

Commits on Sep 3, 2024

  1. fix wint4 bug

    @lixcli
    lixcli committed Sep 3, 2024
    Configuration menu
    Copy the full SHA
    cebf8f0 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'develop' of https://github.com/lixcli/PaddleNLP

    @lixcli
    lixcli committed Sep 3, 2024
    Configuration menu
    Copy the full SHA
    bc67b75 View commit details
    Browse the repository at this point in the history

  3. fix wint4 config bug

    @lixcli
    lixcli committed Sep 3, 2024
    Configuration menu
    Copy the full SHA
    5f49b75 View commit details
    Browse the repository at this point in the history

Commits on Sep 6, 2024

  1. use independent avg/abs_max observer

    @lixcli
    lixcli committed Sep 6, 2024
    Configuration menu
    Copy the full SHA
    82138f2 View commit details
    Browse the repository at this point in the history

  2. rename fp8 quant_type

    @lixcli
    lixcli committed Sep 6, 2024
    Configuration menu
    Copy the full SHA
    9d4ffa5 View commit details
    Browse the repository at this point in the history

  3. Merge branch 'develop' of https://github.com/lixcli/PaddleNLP into up… 

    …date_fakequant_0906
    @lixcli
    lixcli committed Sep 6, 2024
    Configuration menu
    Copy the full SHA
    6199e87 View commit details
    Browse the repository at this point in the history

  4. update quantization.md

    @lixcli
    lixcli committed Sep 6, 2024
    Configuration menu
    Copy the full SHA
    5023827 View commit details
    Browse the repository at this point in the history

  5. remove ceval in run_finetune.py

    @lixcli
    lixcli committed Sep 6, 2024
    Configuration menu
    Copy the full SHA
    fd8df17 View commit details
    Browse the repository at this point in the history

Commits on Sep 10, 2024

  1. Merge branch 'develop' into update_fakequant_0906

    @lixcli
    lixcli authored Sep 10, 2024
    Configuration menu
    Copy the full SHA
    fd5fee6 View commit details
    Browse the repository at this point in the history

  2. update quantization.md 

    remove unuse code
    @lixcli
    lixcli committed Sep 10, 2024
    Configuration menu
    Copy the full SHA
    ebadeb1 View commit details
    Browse the repository at this point in the history

  3. Merge branch update_fakequant_0906 of https://github.com/lixcli/Paddl… 

    …eNLP into update_0910
    @lixcli
    lixcli committed Sep 10, 2024
    Configuration menu
    Copy the full SHA
    8fe32c5 View commit details
    Browse the repository at this point in the history

  4. update quantization.md

    @lixcli
    lixcli committed Sep 10, 2024
    Configuration menu
    Copy the full SHA
    8f58690 View commit details
    Browse the repository at this point in the history