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

Cudnn support #16

Closed
wants to merge 29 commits into from
Closed

Cudnn support #16

wants to merge 29 commits into from

Commits on Nov 5, 2022

  1. init

    @M1ngXU
    M1ngXU committed Nov 5, 2022
    Configuration menu
    Copy the full SHA
    99d8c8f View commit details
    Browse the repository at this point in the history

  2. added activations

    @M1ngXU
    M1ngXU committed Nov 5, 2022
    Configuration menu
    Copy the full SHA
    e9ae2a6 View commit details
    Browse the repository at this point in the history

Commits on Nov 6, 2022

  1. forward conv

    @M1ngXU
    M1ngXU committed Nov 6, 2022
    Configuration menu
    Copy the full SHA
    88bc500 View commit details
    Browse the repository at this point in the history

Commits on Nov 7, 2022

  1. cudnn runs on correct stream, activation backward

    @M1ngXU
    M1ngXU committed Nov 7, 2022
    Configuration menu
    Copy the full SHA
    9860a89 View commit details
    Browse the repository at this point in the history

Commits on Nov 8, 2022

  1. added convolution backward op

    @M1ngXU
    M1ngXU committed Nov 8, 2022
    Configuration menu
    Copy the full SHA
    f1aa74d View commit details
    Browse the repository at this point in the history

Commits on Nov 9, 2022

  1. added batchnorm (backward might be wrong), better tensor allocation, fmt

    @M1ngXU
    M1ngXU committed Nov 9, 2022
    Configuration menu
    Copy the full SHA
    0fdc45f View commit details
    Browse the repository at this point in the history

Commits on Nov 11, 2022

  1. added tensor ops

    @M1ngXU
    M1ngXU committed Nov 11, 2022
    Configuration menu
    Copy the full SHA
    27cd864 View commit details
    Browse the repository at this point in the history

  2. using cuda cudnn result to propagate cuda errors instead of panicking

    @M1ngXU
    M1ngXU committed Nov 11, 2022
    Configuration menu
    Copy the full SHA
    6040079 View commit details
    Browse the repository at this point in the history

  3. removed tensors from batch norm struct

    @M1ngXU
    M1ngXU committed Nov 11, 2022
    Configuration menu
    Copy the full SHA
    94dd49c View commit details
    Browse the repository at this point in the history

  4. now only using cuda cudnn result, refactored filter

    @M1ngXU
    M1ngXU committed Nov 11, 2022
    Configuration menu
    Copy the full SHA
    fd7b1a9 View commit details
    Browse the repository at this point in the history

  5. removed activation descriptor as activation itself doesn't hold data

    @M1ngXU
    M1ngXU committed Nov 11, 2022
    Configuration menu
    Copy the full SHA
    01bbbcd View commit details
    Browse the repository at this point in the history

  6. split modules into multiple files/dirs

    @M1ngXU
    M1ngXU committed Nov 11, 2022
    Configuration menu
    Copy the full SHA
    f3478fb View commit details
    Browse the repository at this point in the history

  7. more file/dir refactoring

    @M1ngXU
    M1ngXU committed Nov 11, 2022
    Configuration menu
    Copy the full SHA
    c18c16d View commit details
    Browse the repository at this point in the history

  8. refactored conv, now only using desc

    @M1ngXU
    M1ngXU committed Nov 11, 2022
    Configuration menu
    Copy the full SHA
    616f490 View commit details
    Browse the repository at this point in the history

Commits on Nov 12, 2022

  1. splitting batchnorm, adding docs

    @M1ngXU
    M1ngXU committed Nov 12, 2022
    Configuration menu
    Copy the full SHA
    d2ab32c View commit details
    Browse the repository at this point in the history

  2. added docs

    @M1ngXU
    M1ngXU committed Nov 12, 2022
    Configuration menu
    Copy the full SHA
    42c5ad4 View commit details
    Browse the repository at this point in the history

  3. finished doc

    @M1ngXU
    M1ngXU committed Nov 12, 2022
    Configuration menu
    Copy the full SHA
    6eca31e View commit details
    Browse the repository at this point in the history

  4. added softmax

    @M1ngXU
    M1ngXU committed Nov 12, 2022
    Configuration menu
    Copy the full SHA
    a615061 View commit details
    Browse the repository at this point in the history

  5. added pooling

    @M1ngXU
    M1ngXU committed Nov 12, 2022
    Configuration menu
    Copy the full SHA
    916db90 View commit details
    Browse the repository at this point in the history

Commits on Nov 13, 2022

  1. added custom kernel for tensor division

    @M1ngXU
    M1ngXU committed Nov 13, 2022
    Configuration menu
    Copy the full SHA
    f77fccf View commit details
    Browse the repository at this point in the history

  2. refactored cudnn custom kernel

    @M1ngXU
    M1ngXU committed Nov 13, 2022
    Configuration menu
    Copy the full SHA
    9efcadd View commit details
    Browse the repository at this point in the history

  3. implementing into kernel param for tensor data

    @M1ngXU
    M1ngXU committed Nov 13, 2022
    Configuration menu
    Copy the full SHA
    e755961 View commit details
    Browse the repository at this point in the history

  4. renaming as_data to get_data_ref

    @M1ngXU
    M1ngXU committed Nov 13, 2022
    Configuration menu
    Copy the full SHA
    fb0a32e View commit details
    Browse the repository at this point in the history

  5. added scaling to division

    @M1ngXU
    M1ngXU committed Nov 13, 2022
    Configuration menu
    Copy the full SHA
    3892ba4 View commit details
    Browse the repository at this point in the history

  6. added clone_into_new to create new tensors with another data allocation

    @M1ngXU
    M1ngXU committed Nov 13, 2022
    Configuration menu
    Copy the full SHA
    73662c0 View commit details
    Browse the repository at this point in the history

  7. refactored custom kernels, added sin/cos

    @M1ngXU
    M1ngXU committed Nov 13, 2022
    Configuration menu
    Copy the full SHA
    95e710e View commit details
    Browse the repository at this point in the history

Commits on Nov 15, 2022

  1. added reduce and broadcasting, working on exmaple

    @M1ngXU
    M1ngXU committed Nov 15, 2022
    Configuration menu
    Copy the full SHA
    8ef0600 View commit details
    Browse the repository at this point in the history

Commits on Dec 12, 2022

  1. fixed custom kernels

    @M1ngXU
    M1ngXU committed Dec 12, 2022
    Configuration menu
    Copy the full SHA
    b52dbed View commit details
    Browse the repository at this point in the history

Commits on Jan 9, 2023

  1. stash

    @M1ngXU
    M1ngXU committed Jan 9, 2023
    Configuration menu
    Copy the full SHA
    6a23616 View commit details
    Browse the repository at this point in the history