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 quantization conv and matmul #10327

Closed
wants to merge 84 commits into from
Closed

add quantization conv and matmul #10327

wants to merge 84 commits into from

Commits on Aug 18, 2023

  1. change for sd torch graph compile

    @strint
    strint committed Aug 18, 2023
    Configuration menu
    Copy the full SHA
    2cfde9d View commit details
    Browse the repository at this point in the history

  2. auto format by CI

    @oneflow-ci-bot
    oneflow-ci-bot committed Aug 18, 2023
    Configuration menu
    Copy the full SHA
    f04cc42 View commit details
    Browse the repository at this point in the history

  3. Update python/oneflow/nn/graph/graph.py 

    Co-authored-by: Houjiang Chen <[email protected]>
    @strint @hjchen2
    strint and hjchen2 authored Aug 18, 2023
    Configuration menu
    Copy the full SHA
    c30278e View commit details
    Browse the repository at this point in the history

  4. auto format by CI

    @oneflow-ci-bot
    oneflow-ci-bot committed Aug 18, 2023
    Configuration menu
    Copy the full SHA
    b0d72a3 View commit details
    Browse the repository at this point in the history

  5. restore md

    @strint
    strint committed Aug 18, 2023
    Configuration menu
    Copy the full SHA
    ed27149 View commit details
    Browse the repository at this point in the history

Commits on Aug 21, 2023

  1. update cutlass and support int8 conv

    @hjchen2
    hjchen2 committed Aug 21, 2023
    Configuration menu
    Copy the full SHA
    df4a31d View commit details
    Browse the repository at this point in the history

Commits on Aug 22, 2023

  1. tuning int8 conv kernels

    @hjchen2
    hjchen2 committed Aug 22, 2023
    Configuration menu
    Copy the full SHA
    cc82f66 View commit details
    Browse the repository at this point in the history

  2. fix

    @hjchen2
    hjchen2 committed Aug 22, 2023
    Configuration menu
    Copy the full SHA
    1e3c555 View commit details
    Browse the repository at this point in the history

  3. rm incompatible int8 conv implementation

    @hjchen2
    hjchen2 committed Aug 22, 2023
    Configuration menu
    Copy the full SHA
    19770b1 View commit details
    Browse the repository at this point in the history

  4. Merge branch 'master' into support_sd_torch_compile

    @hjchen2
    hjchen2 authored Aug 22, 2023
    Configuration menu
    Copy the full SHA
    7bfd55a View commit details
    Browse the repository at this point in the history

  5. half reduce_min and reduce_max

    @hjchen2
    hjchen2 committed Aug 22, 2023
    Configuration menu
    Copy the full SHA
    c8110e8 View commit details
    Browse the repository at this point in the history

  6. Merge remote-tracking branch 'origin/support_sd_torch_compile' into d… 

    …ev_int8_conv
    @hjchen2
    hjchen2 committed Aug 22, 2023
    Configuration menu
    Copy the full SHA
    dc5db78 View commit details
    Browse the repository at this point in the history

Commits on Aug 23, 2023

  1. fake dynamo

    @strint
    strint committed Aug 23, 2023
    Configuration menu
    Copy the full SHA
    30435f9 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'master' of https://github.com/Oneflow-Inc/oneflow into … 

    …fix_dynamo_mock_error
    @strint
    strint committed Aug 23, 2023
    Configuration menu
    Copy the full SHA
    6cfa893 View commit details
    Browse the repository at this point in the history

  3. auto format by CI

    @oneflow-ci-bot
    oneflow-ci-bot committed Aug 23, 2023
    Configuration menu
    Copy the full SHA
    69bbc0e View commit details
    Browse the repository at this point in the history

  4. warmup int8 conv algo

    @hjchen2
    hjchen2 committed Aug 23, 2023
    Configuration menu
    Copy the full SHA
    6158db3 View commit details
    Browse the repository at this point in the history

  5. general argstree

    @strint
    strint committed Aug 23, 2023
    Configuration menu
    Copy the full SHA
    998c501 View commit details
    Browse the repository at this point in the history

  6. Merge branch 'fix_dynamo_mock_error' of https://github.com/Oneflow-In… 

    …c/oneflow into fix_dynamo_mock_error
    @strint
    strint committed Aug 23, 2023
    Configuration menu
    Copy the full SHA
    eccb5f1 View commit details
    Browse the repository at this point in the history

  7. auto format by CI

    @oneflow-ci-bot
    oneflow-ci-bot committed Aug 23, 2023
    Configuration menu
    Copy the full SHA
    44a3710 View commit details
    Browse the repository at this point in the history

Commits on Aug 24, 2023

  1. refine repr of argstree

    @strint
    strint committed Aug 24, 2023
    Configuration menu
    Copy the full SHA
    3619406 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'fix_dynamo_mock_error' of https://github.com/Oneflow-In… 

    …c/oneflow into fix_dynamo_mock_error
    @strint
    strint committed Aug 24, 2023
    Configuration menu
    Copy the full SHA
    93d0fe3 View commit details
    Browse the repository at this point in the history

  3. auto format by CI

    @oneflow-ci-bot
    oneflow-ci-bot committed Aug 24, 2023
    Configuration menu
    Copy the full SHA
    ccac35e View commit details
    Browse the repository at this point in the history

  4. graph allow any input

    @strint
    strint committed Aug 24, 2023
    Configuration menu
    Copy the full SHA
    025d4c5 View commit details
    Browse the repository at this point in the history

  5. Merge branch 'fix_dynamo_mock_error' of https://github.com/Oneflow-In… 

    …c/oneflow into fix_dynamo_mock_error
    @strint
    strint committed Aug 24, 2023
    Configuration menu
    Copy the full SHA
    679e0d5 View commit details
    Browse the repository at this point in the history

  6. save support any input

    @strint
    strint committed Aug 24, 2023
    Configuration menu
    Copy the full SHA
    354cf21 View commit details
    Browse the repository at this point in the history

Commits on Aug 25, 2023

  1. fix

    @jackalcooper
    jackalcooper committed Aug 25, 2023
    Configuration menu
    Copy the full SHA
    99b3d60 View commit details
    Browse the repository at this point in the history

  2. fix gsize>64

    @jackalcooper
    jackalcooper committed Aug 25, 2023
    Configuration menu
    Copy the full SHA
    8347ed8 View commit details
    Browse the repository at this point in the history

  3. Merge branch 'add-kMaxInputCount-in-GroupedMatmulFunctor' of https://… 

    …github.com/Oneflow-Inc/oneflow into sdxl_dev
    @strint
    strint committed Aug 25, 2023
    Configuration menu
    Copy the full SHA
    6705366 View commit details
    Browse the repository at this point in the history

Commits on Aug 26, 2023

  1. compile with cutlass extension

    @hjchen2
    hjchen2 committed Aug 26, 2023
    Configuration menu
    Copy the full SHA
    dde053c View commit details
    Browse the repository at this point in the history

  2. fix

    @hjchen2
    hjchen2 committed Aug 26, 2023
    Configuration menu
    Copy the full SHA
    cb84faf View commit details
    Browse the repository at this point in the history

Commits on Aug 27, 2023

  1. add conv2d quant op

    @hjchen2
    hjchen2 committed Aug 27, 2023
    Configuration menu
    Copy the full SHA
    5558ed3 View commit details
    Browse the repository at this point in the history

  2. add conv2d quant kernel

    @hjchen2
    hjchen2 committed Aug 27, 2023
    Configuration menu
    Copy the full SHA
    a947f44 View commit details
    Browse the repository at this point in the history

Commits on Aug 28, 2023

  1. refactor quantization kernel and fix tuning warmup pass

    @hjchen2
    hjchen2 committed Aug 28, 2023
    Configuration menu
    Copy the full SHA
    7a38a9f View commit details
    Browse the repository at this point in the history

  2. fix

    @hjchen2
    hjchen2 committed Aug 28, 2023
    Configuration menu
    Copy the full SHA
    e48f738 View commit details
    Browse the repository at this point in the history

Commits on Aug 29, 2023

  1. Merge branch 'sdxl_dev' of https://github.com/Oneflow-Inc/oneflow int… 

    …o dev_int8_conv
    @hjchen2
    hjchen2 committed Aug 29, 2023
    Configuration menu
    Copy the full SHA
    63e92d3 View commit details
    Browse the repository at this point in the history

Commits on Aug 30, 2023

  1. fix compilation error for cuda12.2

    @hjchen2
    hjchen2 committed Aug 30, 2023
    Configuration menu
    Copy the full SHA
    5500840 View commit details
    Browse the repository at this point in the history

  2. eliminate scalar math op

    @hjchen2
    hjchen2 committed Aug 30, 2023
    2 Configuration menu
    Copy the full SHA
    cc7c49c View commit details
    Browse the repository at this point in the history

  3. refactor

    @hjchen2
    hjchen2 committed Aug 30, 2023
    Configuration menu
    Copy the full SHA
    f1a4397 View commit details
    Browse the repository at this point in the history

Commits on Aug 31, 2023

  1. fuse add to output completely

    @hjchen2
    hjchen2 committed Aug 31, 2023
    Configuration menu
    Copy the full SHA
    a889fc5 View commit details
    Browse the repository at this point in the history

  2. add_quant_matmul

    @clackhan
    clackhan committed Aug 31, 2023
    Configuration menu
    Copy the full SHA
    a80b11d View commit details
    Browse the repository at this point in the history

Commits on Sep 1, 2023

  1. fix unused-but-set-variable error

    @hjchen2
    hjchen2 committed Sep 1, 2023
    Configuration menu
    Copy the full SHA
    cd745cc View commit details
    Browse the repository at this point in the history

  2. conv2d_quant support add_to_output

    @hjchen2
    hjchen2 committed Sep 1, 2023
    Configuration menu
    Copy the full SHA
    4e499e1 View commit details
    Browse the repository at this point in the history

  3. Merge branch 'dev_int8_conv' of https://github.com/Oneflow-Inc/oneflow … 

    …into dev_int8_conv
    @hjchen2
    hjchen2 committed Sep 1, 2023
    Configuration menu
    Copy the full SHA
    c792366 View commit details
    Browse the repository at this point in the history

  4. fix and reformat

    @hjchen2
    hjchen2 committed Sep 1, 2023
    Configuration menu
    Copy the full SHA
    63b449c View commit details
    Browse the repository at this point in the history

  5. update cutlass extension branch

    @hjchen2
    hjchen2 committed Sep 1, 2023
    Configuration menu
    Copy the full SHA
    eef55aa View commit details
    Browse the repository at this point in the history

Commits on Sep 2, 2023

  1. matmul_quant supports add_to_output

    @hjchen2
    hjchen2 committed Sep 2, 2023
    Configuration menu
    Copy the full SHA
    d62f6e4 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'master' into dev_int8_conv

    @hjchen2
    hjchen2 authored Sep 2, 2023
    Configuration menu
    Copy the full SHA
    0fff5e0 View commit details
    Browse the repository at this point in the history

  3. auto format by CI

    @oneflow-ci-bot
    oneflow-ci-bot committed Sep 2, 2023
    Configuration menu
    Copy the full SHA
    418598e View commit details
    Browse the repository at this point in the history

Commits on Sep 3, 2023

  1. support to find the fastest kernel for matmul_quant

    @hjchen2
    hjchen2 committed Sep 3, 2023
    Configuration menu
    Copy the full SHA
    9031d5c View commit details
    Browse the repository at this point in the history

  2. Merge branch 'dev_int8_conv' of https://github.com/Oneflow-Inc/oneflow … 

    …into dev_int8_conv
    @hjchen2
    hjchen2 committed Sep 3, 2023
    Configuration menu
    Copy the full SHA
    f31f2b1 View commit details
    Browse the repository at this point in the history

Commits on Sep 4, 2023

  1. fuse gelu quant

    @hjchen2
    hjchen2 committed Sep 4, 2023
    Configuration menu
    Copy the full SHA
    a857441 View commit details
    Browse the repository at this point in the history

  2. add_prune_redundant_quantization_op_pass

    @clackhan
    clackhan committed Sep 4, 2023
    Configuration menu
    Copy the full SHA
    9ac9a1c View commit details
    Browse the repository at this point in the history

  3. auto format by CI

    @oneflow-ci-bot
    oneflow-ci-bot committed Sep 4, 2023
    Configuration menu
    Copy the full SHA
    2d45cc6 View commit details
    Browse the repository at this point in the history

  4. optimize activation dynamic quantization

    @hjchen2
    hjchen2 committed Sep 4, 2023
    Configuration menu
    Copy the full SHA
    9af9b8d View commit details
    Browse the repository at this point in the history

  5. Merge branch 'dev_int8_conv' of https://github.com/Oneflow-Inc/oneflow … 

    …into dev_int8_conv
    @hjchen2
    hjchen2 committed Sep 4, 2023
    Configuration menu
    Copy the full SHA
    60b8611 View commit details
    Browse the repository at this point in the history

Commits on Sep 5, 2023

  1. optimize

    @hjchen2
    hjchen2 committed Sep 5, 2023
    Configuration menu
    Copy the full SHA
    4a5a5c6 View commit details
    Browse the repository at this point in the history

  2. impl mlir pass

    @clackhan
    clackhan committed Sep 5, 2023
    Configuration menu
    Copy the full SHA
    42c1ee9 View commit details
    Browse the repository at this point in the history

  3. Merge branch 'dev_int8_conv' of https://github.com/Oneflow-Inc/oneflow … 

    …into dev_int8_conv
    @clackhan
    clackhan committed Sep 5, 2023
    Configuration menu
    Copy the full SHA
    af2f7d4 View commit details
    Browse the repository at this point in the history

  4. auto format by CI

    @oneflow-ci-bot
    oneflow-ci-bot committed Sep 5, 2023
    Configuration menu
    Copy the full SHA
    7a5ba06 View commit details
    Browse the repository at this point in the history

  5. refine

    @clackhan
    clackhan committed Sep 5, 2023
    Configuration menu
    Copy the full SHA
    d294ef2 View commit details
    Browse the repository at this point in the history

  6. Merge branch 'dev_int8_conv' of https://github.com/Oneflow-Inc/oneflow … 

    …into dev_int8_conv
    @clackhan
    clackhan committed Sep 5, 2023
    Configuration menu
    Copy the full SHA
    c384f31 View commit details
    Browse the repository at this point in the history

  7. update

    @hjchen2
    hjchen2 committed Sep 5, 2023
    Configuration menu
    Copy the full SHA
    d414feb View commit details
    Browse the repository at this point in the history

  8. Merge branch 'dev_int8_conv' of https://github.com/Oneflow-Inc/oneflow … 

    …into dev_int8_conv
    @hjchen2
    hjchen2 committed Sep 5, 2023
    Configuration menu
    Copy the full SHA
    1388490 View commit details
    Browse the repository at this point in the history

Commits on Sep 6, 2023

  1. fmt

    @jackalcooper
    jackalcooper committed Sep 6, 2023
    Configuration menu
    Copy the full SHA
    13c0669 View commit details
    Browse the repository at this point in the history

  2. prune_reduntant_quant_from_input_op

    @clackhan
    clackhan committed Sep 6, 2023
    Configuration menu
    Copy the full SHA
    14c4bba View commit details
    Browse the repository at this point in the history

  3. Merge branch 'dev_int8_conv' of https://github.com/Oneflow-Inc/oneflow … 

    …into dev_int8_conv
    @clackhan
    clackhan committed Sep 6, 2023
    Configuration menu
    Copy the full SHA
    28ec052 View commit details
    Browse the repository at this point in the history

  4. refine

    @clackhan
    clackhan committed Sep 6, 2023
    Configuration menu
    Copy the full SHA
    d30b33a View commit details
    Browse the repository at this point in the history

  5. fuse dynamic quant conv

    @hjchen2
    hjchen2 committed Sep 6, 2023
    Configuration menu
    Copy the full SHA
    78b51cd View commit details
    Browse the repository at this point in the history

  6. Merge branch 'dev_int8_conv' of https://github.com/Oneflow-Inc/oneflow … 

    …into dev_int8_conv
    @hjchen2
    hjchen2 committed Sep 6, 2023
    Configuration menu
    Copy the full SHA
    0059708 View commit details
    Browse the repository at this point in the history

  7. matmul quant with fiter scale

    @hjchen2
    hjchen2 committed Sep 6, 2023
    Configuration menu
    Copy the full SHA
    a102cc0 View commit details
    Browse the repository at this point in the history

Commits on Sep 7, 2023

  1. refine

    @hjchen2
    hjchen2 committed Sep 7, 2023
    Configuration menu
    Copy the full SHA
    ef397d6 View commit details
    Browse the repository at this point in the history

  2. fuse layer norm and dynamic quant

    @hjchen2
    hjchen2 committed Sep 7, 2023
    Configuration menu
    Copy the full SHA
    f262013 View commit details
    Browse the repository at this point in the history

Commits on Sep 8, 2023

  1. add_grouped_matmul_quant

    @clackhan
    clackhan committed Sep 8, 2023
    Configuration menu
    Copy the full SHA
    54b3dd5 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'dev_int8_conv' of https://github.com/Oneflow-Inc/oneflow … 

    …into dev_int8_conv
    @clackhan
    clackhan committed Sep 8, 2023
    Configuration menu
    Copy the full SHA
    5c64169 View commit details
    Browse the repository at this point in the history

  3. fix

    @hjchen2
    hjchen2 committed Sep 8, 2023
    Configuration menu
    Copy the full SHA
    66f50e5 View commit details
    Browse the repository at this point in the history

  4. optimize

    @hjchen2
    hjchen2 committed Sep 8, 2023
    Configuration menu
    Copy the full SHA
    bfd5319 View commit details
    Browse the repository at this point in the history

Commits on Sep 10, 2023

  1. Merge branch 'dev_int8_conv' of https://github.com/Oneflow-Inc/oneflow … 

    …into dev_int8_conv
    @hjchen2
    hjchen2 committed Sep 10, 2023
    Configuration menu
    Copy the full SHA
    89acc4c View commit details
    Browse the repository at this point in the history

Commits on Sep 13, 2023

  1. update xformers fmha

    @hjchen2
    hjchen2 committed Sep 13, 2023
    Configuration menu
    Copy the full SHA
    a153075 View commit details
    Browse the repository at this point in the history

  2. Revert "update xformers fmha" 

    This reverts commit a153075.
    @hjchen2
    hjchen2 committed Sep 13, 2023
    Configuration menu
    Copy the full SHA
    af2da94 View commit details
    Browse the repository at this point in the history

  3. optimize quant

    @hjchen2
    hjchen2 committed Sep 13, 2023
    Configuration menu
    Copy the full SHA
    26dcc1d View commit details
    Browse the repository at this point in the history

Commits on Sep 15, 2023

  1. add_cutlass_gemm_array_tuner

    @clackhan
    clackhan committed Sep 15, 2023
    Configuration menu
    Copy the full SHA
    397304f View commit details
    Browse the repository at this point in the history

  2. fix

    @hjchen2
    hjchen2 committed Sep 15, 2023
    Configuration menu
    Copy the full SHA
    c1c859b View commit details
    Browse the repository at this point in the history

Commits on Sep 22, 2023

  1. fuse_min_max_observer_and_matmul_quant

    @clackhan
    clackhan committed Sep 22, 2023
    Configuration menu
    Copy the full SHA
    4ba3cc5 View commit details
    Browse the repository at this point in the history

  2. refine

    @clackhan
    clackhan committed Sep 22, 2023
    Configuration menu
    Copy the full SHA
    7e154e1 View commit details
    Browse the repository at this point in the history