Skip to content

Prediction of material microstructure evolution via convolutional LSTM neural networks. Implementation in pytorch.

License

Notifications You must be signed in to change notification settings

Mooon/learning_microstructure_evolution

Repository files navigation

Video Forecasting model (Conv-LSTM) to accelerate the prediction of material microstructure Evolution

Intro

The focus of this project is the development of cost-effective and adaptable software strategies to promote Virtual Testing for the Process-Structure-Property (PSP) modeling.

We investigate the use of deep neural networks for the prediction of microstructure evolution in engineering materials. This investigation enables incremental materials technology improvements, by accelerating the traditional CFD workloads and establishing process-structure linkage.

The project is conducted in collaboration with the Materials and Processes Section TEC-MSP at the ESA's Mechanical Engineering Department at ESTEC, Noordwijk, The Netherlands, within the research group of Integrated Computational Materials Engineering for space relevant Additive Manufacturing processes.

Requirements

  • Python > 3.0
  • Pytorch 1.0

Install

pip3 install -r requirements.txt

Dataset

Our dataset consists of 2D time series datasets (image time series). Labels are automatically generated according to statistical sampling (Dakota). Data quality defined by resolution of numerical datasets (correlated to computation time). The time series images are organized in folders, based on different simulation settings and every folder is listed in a .txt config file. The data are produced and owned by ESA TEC-MSP.

Dataloader

A data sample (a video) is created based on the temporal sequence of images by a custom dataloader. To enumerate all video samples in the dataset, a new .txt config file is manually created that contains a row for each video clip sample in the dataset. Note that the training, validation, and testing datasets have separate config files. Each row is in the format VIDEO_PATH START_FRAME END_FRAME.

A complete tutorial on how to create a video dataloader with pytorch can be found here: [video dataset loading]

Model

The model is a conv-lstm network, which is implemented in PyTorch by NVidia Labs, based on the paper, 'Convolutional Tensor-Train LSTM for Spatio-Temporal Learning', NeurIPS 2020. [project page] [original repo]

Loading the dataset and training the model

  1. create a custom config file for training and testing with config_videocreator.py
  2. Load the videos with ms_dataloader.py
  3. Train the model with ms_model_train.py

Training the model with HPC cluster

In order to submit an already created job script, the command sbatch has to be used.

sbatch train_ms.sh

Testing the model

  • In order to run the test code, you must have trained the model and saved at least one checkpoint as a '.pt' file

License

The conv-tt-lstm model is under Copyright (c) 2020 NVIDIA Corporation. All rights reserved. This work is licensed under a NVIDIA Open Source Non-commercial license.

Contacts

The project for the microstructure prediction is written by Monica Rotulo (monica.rotulo@surf.nl)

About

Prediction of material microstructure evolution via convolutional LSTM neural networks. Implementation in pytorch.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published