eynollah/train/requirements.txt

ocrd-fork-sacred >= 0.8.7.post1
seaborn
numpy
tqdm
imutils
scipy
tensorflow-addons # for connected_components, depublished and only compatible with tensorflow < 2.16
tensorflow < 2.16 # for tensorflow-addons, so only needed in training
tf-keras < 2.16 # avoid keras 3 (also needs TF_USE_LEGACY_KERAS=1)
protobuf < 5 # for tensorflow-addons, so only needed in training
torch
transformers <= 4.30.2 ; python_version < '3.10'
transformers >= 5 ; python_version >= '3.10'
eynollah-fork-tf2onnx == 1.17.0.post2
ml_dtypes >= 0.5
training deps: use sacred fork w/o pkg_resources, pin tf/tf_keras, protobuf packages to work with tensorflow_addons 2026-05-07 18:09:27 +02:00			`ocrd-fork-sacred >= 0.8.7.post1`
adding requirements 2024-04-04 11:30:12 +02:00			`seaborn`
training: use proper Keras callbacks and top-level loop 2026-01-22 11:25:00 +01:00			`numpy`
adding requirements 2024-04-04 11:30:12 +02:00			`tqdm`
			`imutils`
code formatting with black; typos 2024-04-10 22:20:23 +02:00			`scipy`
training deps: use sacred fork w/o pkg_resources, pin tf/tf_keras, protobuf packages to work with tensorflow_addons 2026-05-07 18:09:27 +02:00			`tensorflow-addons # for connected_components, depublished and only compatible with tensorflow < 2.16`
			`tensorflow < 2.16 # for tensorflow-addons, so only needed in training`
training: require tf2onnx and pin ml_dtypes >= 0.5 2026-07-09 14:37:51 +02:00			`tf-keras < 2.16 # avoid keras 3 (also needs TF_USE_LEGACY_KERAS=1)`
training deps: use sacred fork w/o pkg_resources, pin tf/tf_keras, protobuf packages to work with tensorflow_addons 2026-05-07 18:09:27 +02:00			`protobuf < 5 # for tensorflow-addons, so only needed in training`
integrating transformer ocr 2026-02-03 19:45:50 +01:00			`torch`
			`transformers <= 4.30.2 ; python_version < '3.10'`
			`transformers >= 5 ; python_version >= '3.10'`
upgrade tf2onnx fork dep, remove spurious tf-data dep 2026-07-09 17:30:48 +02:00			`eynollah-fork-tf2onnx == 1.17.0.post2`
training: require tf2onnx and pin ml_dtypes >= 0.5 2026-07-09 14:37:51 +02:00			`ml_dtypes >= 0.5`