Skip to content
This repository has been archived by the owner on Dec 16, 2022. It is now read-only.

Initializing a VilBERT model from a pre-trained transformer #4495

Merged
merged 12 commits into from Aug 4, 2020
Merged

Initializing a VilBERT model from a pre-trained transformer #4495

merged 12 commits into from Aug 4, 2020

Commits on Jul 20, 2020

  1. saving state

    @matt-gardner
    matt-gardner committed Jul 20, 2020
    Copy the full SHA
    c05cdca View commit details
    Browse the repository at this point in the history

  2. Code is running, though it is returning zero gradients (but not None)

    @matt-gardner
    matt-gardner committed Jul 20, 2020
    Copy the full SHA
    13c47be View commit details
    Browse the repository at this point in the history

  3. initial test passing, still working on albert

    @matt-gardner
    matt-gardner committed Jul 20, 2020
    Copy the full SHA
    96a7771 View commit details
    Browse the repository at this point in the history

Commits on Jul 21, 2020

  1. albert works, but bert-base-uncased still gives zero gradients

    @matt-gardner
    matt-gardner committed Jul 21, 2020
    Copy the full SHA
    321c7a9 View commit details
    Browse the repository at this point in the history

Commits on Jul 23, 2020

  1. Loading of weights should now work

    @matt-gardner
    matt-gardner committed Jul 23, 2020
    Copy the full SHA
    5bd4701 View commit details
    Browse the repository at this point in the history

  2. black, flake, mypy

    @matt-gardner
    matt-gardner committed Jul 23, 2020
    Copy the full SHA
    e1da51c View commit details
    Browse the repository at this point in the history

Commits on Jul 24, 2020

  1. Merge remote-tracking branch 'origin/vision' into load_pretrained

    @dirkgr
    dirkgr committed Jul 24, 2020
    Copy the full SHA
    483fc5c View commit details
    Browse the repository at this point in the history

Commits on Jul 27, 2020

  1. remove drop and mask functionality from reader

    @matt-gardner
    matt-gardner committed Jul 27, 2020
    Copy the full SHA
    f3131d9 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'load_pretrained' of github.com:matt-gardner/allennlp in… 

    …to load_pretrained
    @matt-gardner
    matt-gardner committed Jul 27, 2020
    Copy the full SHA
    f4473c4 View commit details
    Browse the repository at this point in the history

  3. make comment better

    @matt-gardner
    matt-gardner committed Jul 27, 2020
    Copy the full SHA
    3ff677c View commit details
    Browse the repository at this point in the history

Commits on Jul 30, 2020

  1. fix tests

    @matt-gardner
    matt-gardner committed Jul 30, 2020
    Copy the full SHA
    6d0fb4d View commit details
    Browse the repository at this point in the history

  2. flake

    @matt-gardner
    matt-gardner committed Jul 30, 2020
    Copy the full SHA
    d40d120 View commit details
    Browse the repository at this point in the history