Your friendly Tensorflow 2.x neighbor.

Documentation: https://jeikeilim.github.io/tfhelper/

• Python 3.7

pip install tfhelper

• HDF5Generator

from tfhelper.dataset import HDF5Generator

train_generator = HDF5Generator("/dataset/path.hdf5", "training_data", "test_data")
train_dataset = train_generator.get_dataset(input_shape=(300, 300), batch_size=16, shuffle=True, n_shuffle=1000)

...
model = Define some model

model.fit(train_dataset)

• allow_gpu_memory_growth

• ConfuseCallback
• ModelSaverCallback
• SparsityCallback
• run_tensorboard
• wait_ctrl_c
• get_tf_callbacks

• keras_model_to_tflite
• parse_config
• predict_tflite_interpreter
• evaluate_tflite_interpreter
• load_pruned_model

• get_transfer_learning_model

• get_cam_image