Skip to content

Latest commit

 

History

History
90 lines (62 loc) · 2.89 KB

README.md

File metadata and controls

90 lines (62 loc) · 2.89 KB
Raw

Logo

STEMseg

Help the processing of 4D-STEM data, specifically automatically clustering similar pixels in SEND pencil beam data
View Demo · Report Bug · Request Feature

Table of Contents
  1. About The Project
  2. Running on DLS-Jupyterhub
  3. Contact

About The Project

Intended for processing 4D-STEM datasets using an autoencoder trainer and a DataWrapper for processing the samples. The example notebooks should walkthrough the process of training the encoder and using it to cluster regions with similar diffraction data automatically.

Running on the DLS Jupyterhub

  • Make a copy of the example-notebooks

  • Copy the folders stemseg and stemutils (under modules folder) to the notebook's local directory or /home/{YOUR FED ID}/.local/lib/python3.7/site-packages

  • Boot up a GPU kernel having replaced the container image with: CONTAINER_IMAGE=gcr.io/diamond-pubreg/container-tools/jhub-notebook:cuda10.1

  • Select the EPSIC3.7 environment to run the example notebooks (if not a default)

  • First work through TrainModels.ipynb to create a trained VAE for your datasets

  • Then work through TrainInvestigator.ipynb to use the model to cluster

Contact

Andy Bridger - [email protected]

Project Link: https://github.com/andy-bridger/stemsegment

(back to top)