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.

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

feat: add loading backbone pretrained for multiclass detection, new elements for kie predictor #6

Merged
merged 8 commits into from
Dec 5, 2022

Conversation

aminemindee
Copy link
Owner

No description provided.

@aminemindee aminemindee changed the base branch from feat/multiclass to main November 29, 2022 17:20
@aminemindee aminemindee changed the base branch from main to feat/multiclass November 30, 2022 14:46
@aminemindee aminemindee merged commit aa15f43 into feat/multiclass Dec 5, 2022
aminemindee added a commit that referenced this pull request Dec 5, 2022
…lements for kie predictor (#6)

* feat: ✨ add load backbone

* feat: change kie predictor out

* fix new elements for kie, dataset when class is empty and fix and add tests

* fix api kie route

* fix evaluate kie script

* fix black

* remove commented code

* update README
aminemindee added a commit that referenced this pull request Jan 27, 2023
…redictor (mindee#1097)

* feat: add handling of multiclass format in detection dataset loading with all transforms

* fix: fix loss computation and make training work

* feat: make loss computation vectorized and change target building to handle better class ids

* feat: add multiclass to pytorch and fix tests

* feat: add doc about Pages changes and multilabel dataset for training

* feat: fix api dockerfile and make it work with new changes

* fix reference tests

* refactor: refactor invert dict list and list dict function into one simpler

* fix: style and mypy

* docs: make it more clear for new data format

* explain why python version was upped

* add assert on length of tuple

* feat: add class names can be obtained from model config

* fix: prioritize class_names from dataset over model config

* fix: fix show samples in training

* fix: add check when target is dict and all values are numpy arrays

* fix: make detection target always dict and remove unnecessary made code from it

* fix: script detection evaluation tests and dataset tests with target as dict

* fix tests also on pytorch

* feat: Add kie predictor and io elements and visualization that come with it

* fix: revert ocr predictor to old format

* fix tests and add test for kie predictor

* up project version to 0.7.0

* update api to fix it and add kie route

* fix api version

* feat: sort class names to always have the same order.

* sort imports to avoid cyclic imports

* fix class_names default, use of tf_is_available avoid and copyright dates

* feat: update readme and doc with kie predictor

* feat: add loading backbone pretrained for multiclass detection, new elements for kie predictor (#6)

* feat: ✨ add load backbone

* feat: change kie predictor out

* fix new elements for kie, dataset when class is empty and fix and add tests

* fix api kie route

* fix evaluate kie script

* fix black

* remove commented code

* update README

* fix mypy
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant