You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Mike Gerber e2a3f7e10c | 5 years ago | |
---|---|---|
data@f817209ba7 | 5 years ago | |
LICENSE | 5 years ago | |
README.md | 5 years ago | |
qurator_data_lib.sh | 5 years ago | |
requirements.txt | 5 years ago | |
train.sh | 5 years ago | |
upload-prepare.sh | 5 years ago |
README.md
Train a GT4HistOCR Calamari model
train.sh
trains a Calamari model based on GT4HistOCR. Or rather 5 using
cross-validation to use for confidence voting. This repository mainly
serves as documentation of the provenance of the model published at
https://qurator-data.de/calamari-models/, not as the definitive guide to
training such a model.
Trained models
For a finished model have a look here: https://qurator-data.de/calamari-models/
Training your own model
If you really want to, you can use this script to train your own. It takes about 1 week on a Nvidia RTX 2080 GPU. Please use requirements.txt in that case to setup a virtualenv.
train.sh
is able to download GT4HistOCR from the web if the data
submodule
is not available, that is if you're not a member of the Qurator team at SBB.