Robert Sachunsky
|
342e94e287
|
reinstate ensemble combination and in-memory prediction
|
2 years ago |
|
Robert Sachunsky
|
cade5dda73
|
reintroduce TF stderr / log silencer
|
2 years ago |
Alexander Pacha
|
9ef8259677
|
Merge branch 'master-upstream'
# Conflicts:
# README.md
# sbb_binarize/cli.py
# sbb_binarize/sbb_binarize.py
|
2 years ago |
Konstantin Baierer
|
aeb6804b7d
|
📦 v0.0.11
|
2 years ago |
|
Konstantin Baierer
|
82816b048e
|
📝 changelog
|
2 years ago |
Konstantin Baierer
|
6fb6946531
|
Merge pull request #53 from bertsky/patch-3
ocrd-tool.json: register SavedModel resources
|
2 years ago |
Clemens Neudecker
|
0d8a9f33f5
|
Update README.md
remove patches parameter from readme
|
2 years ago |
vahid
|
7c3f2176f7
|
issue #45 the patches option is omitted and it means that documents will be processed in patches while no patches is not desired by the tool
|
2 years ago |
Robert Sachunsky
|
743ca0dbed
|
Update requirements.txt
|
2 years ago |
|
Robert Sachunsky
|
f8ade5d3da
|
ocrd-tool.json: register SavedModel resources
|
2 years ago |
|
Clemens Neudecker
|
eff7a47852
|
Merge pull request #52 from bertsky/img-allow-alpha
allow alpha channels
|
2 years ago |
|
Clemens Neudecker
|
56ccb39539
|
Update README
* recommend cropping (fix #49)
* document huggingface saved_model
|
2 years ago |
|
Robert Sachunsky
|
a69470135a
|
load_model: also allow SavedModel directories
|
2 years ago |
|
Robert Sachunsky
|
88d64de67b
|
pil2cv: reduce transparency to white
|
2 years ago |
|
Alexander Pacha
|
4a97eacc56
|
Removed --patches flag, as the new version automatically uses patches in an efficient way.
Updated cli.py to correctly load and initialize the changed SbbBinarizer class
|
2 years ago |
|
Alexander Pacha
|
7edf34a815
|
Adding a bit more documentation
|
2 years ago |
|
Alexander Pacha
|
9db2bcc8b0
|
Remove obsolete comment
|
2 years ago |
|
Alexander Pacha
|
1c7e8c5e21
|
Improved formatting
|
2 years ago |
|
Alexander Pacha
|
1baa70e382
|
Improving comments in the code
|
2 years ago |
|
Alexander Pacha
|
4112c6fe71
|
Rewrote binarization script to always use patches, but in a much more efficient way and adding support for batch-conversion with multiple GPUs.
|
2 years ago |
|
Alexander Pacha
|
b0a8b613e8
|
Improved loading of models to allow providing a directory, added a few type-hints and improved the code-style a little bit by running an auto-formatter on the entire file.
|
2 years ago |
|
vahid
|
f11d0b0bf7
|
#42 is resolved
|
2 years ago |
vahidrezanezhad
|
2355d0ec8f
|
Merge pull request #43 from apacha/fix_no_models_available
Raise error when no models were loaded
|
2 years ago |
|
Alexander Pacha
|
55a9fba677
|
Small fix to raise an exception when no model was found instead of failing later with TypeError: 'int' object is not subscriptable.
|
2 years ago |
|
Konstantin Baierer
|
c99be55af2
|
📦 v0.0.10
|
2 years ago |
|
Konstantin Baierer
|
2abd8ae848
|
📝 changelog
|
2 years ago |
|
Konstantin Baierer
|
d1e9389b54
|
Merge pull request #40 from maxnth/fix/cli
use types.Path instead of click.Path to avoid NameError
|
2 years ago |
maxnth
|
dc5a5c791a
|
use types.Path instead of click.Path to avoid NameError
|
2 years ago |
|
Clemens Neudecker
|
e4c1eb2913
|
Merge pull request #37 from bertsky/patch-2
depend on tensorflow not tensorflow-gpu, drop holding h5py
|
3 years ago |
|
Robert Sachunsky
|
fd93a45505
|
depend on tensorflow not tensorflow-gpu, drop holding h5py
|
3 years ago |
|
Clemens Neudecker
|
0877199c68
|
importing types from click is enough
|
3 years ago |
|
vahidrezanezhad
|
792619e830
|
Update cli.py
|
3 years ago |
|
vahidrezanezhad
|
82e0783e01
|
fixing #30. Directory of model(s) is needed not model file name.
|
3 years ago |
|
Clemens Neudecker
|
2119c907c4
|
Merge pull request #36 from qurator-spk/0.0.9
v0.0.9
|
3 years ago |
|
Clemens Neudecker
|
aab51dbe82
|
Update ocrd-tool.json
|
3 years ago |
|
Clemens Neudecker
|
b9bbbe9004
|
Update CHANGELOG.md
|
3 years ago |
|
Clemens Neudecker
|
ff78d4e85c
|
improve usage instructions
|
3 years ago |
|
vahidrezanezhad
|
0fe5232488
|
Merge pull request #35 from cneud/upgrade_to_tf2_with_tf1.compat_session
fix tf2 with tf1.compat session
|
3 years ago |
|
Clemens Neudecker
|
2d4ce8f706
|
don't require Keras explicitely
|
3 years ago |
|
Clemens Neudecker
|
83c60d7b26
|
cleanup
|
3 years ago |
cneud
|
b7aeb1ae8b
|
use tensorflow Keras backend
|
3 years ago |
|
Clemens Neudecker
|
1a4207c614
|
relax numpy version requirement
|
3 years ago |
|
Clemens Neudecker
|
df8e7bfb33
|
Update requirements.txt
|
3 years ago |
|
vahid
|
6296c6a82c
|
Revert "Merge pull request #34 from qurator-spk/upgrade_to_tf2_with_compat_mode"
This reverts commit 593458d321, reversing
changes made to e1575268d7.
|
3 years ago |
|
vahidrezanezhad
|
593458d321
|
Merge pull request #34 from qurator-spk/upgrade_to_tf2_with_compat_mode
actually use TF2 (with TF1.compat mode)
|
3 years ago |
|
Clemens Neudecker
|
33deaaa90d
|
actually use TF2 (with TF1.compat mode)
|
3 years ago |
|
vahidrezanezhad
|
e1575268d7
|
Merge pull request #31 from bertsky/factor-setup
Factor setup
|
3 years ago |
|
Robert Sachunsky
|
73b7b48dfb
|
ocrd-tool.json: add directory content-type for model
|
3 years ago |
|
Robert Sachunsky
|
416b5dd7e2
|
adapt to TF1 deprecation
|
4 years ago |
|
Robert Sachunsky
|
ad50ed287d
|
binarizer: reset TF session prior to predict()
|
4 years ago |