Contains:
- Tutorials for simple ptychography reconstruction applications with tensorflow (contained in tensorflow_tutorials). It is safe to ignore the python package setup procedure.
- A generic, modular tensorflow-based simulation and reconstruction framework in ptychoSampling. The documentation, however, is quite sparse and sometimes unchanged from that for older versions of the code.
- Simulation and reconstruction examples for far-field, near-field, and bragg ptychography cases.
Does not contain (for now):
- Tests and timing benchmarks.
Notes:
-
For ease of application, the forward model simulation code uses numpy. Only the reconstruction code uses Tensorflow.
-
Uses Tensorflow 1.14 for now. In the future, I am planning on switching away from the static computational graphs to a dynamic framework (Tensorflow 2.0, Pytorch, Autograd, Jax, etc) for ease of usage.