Move plotting scripts to same folder

- Also uploaded pre-trained models to Zenodo
- Still untested

.gitignore

@@ -127,3 +127,5 @@ dmypy.json
 
 # Pyre type checker
 .pyre/
+
+*.zip

README.md

@@ -4,7 +4,7 @@ Multi-fidelity Generative Deep Learning Turbulent Flows [[FoDS]()][[ArXiv](https
 [Nicholas Geneva](https://nicholasgeneva.com/), [Nicholas Zabaras](https://cics.nd.edu)
 
 ---
-[![Documentation Status](https://readthedocs.org/projects/deep-turbulence/badge/?version=latest)](https://deep-turbulence.readthedocs.io/en/latest/?badge=latest) [![dataset](https://zenodo.org/badge/DOI/10.5281/zenodo.4298896.svg)](https://doi.org/10.5281/zenodo.4298896) [![liscense](https://img.shields.io/github/license/zabaras/deep-turbulence)](https://github.com/zabaras/deep-turbulence/blob/master/LICENSE)
+[![Documentation Status](https://readthedocs.org/projects/deep-turbulence/badge/?version=latest)](https://deep-turbulence.readthedocs.io/en/latest/?badge=latest) [![dataset](https://zenodo.org/badge/DOI/10.5281/zenodo.4311698.svg)](https://doi.org/10.5281/zenodo.4311698) [![liscense](https://img.shields.io/github/license/zabaras/deep-turbulence)](https://github.com/zabaras/deep-turbulence/blob/master/LICENSE)
 
 A novel multi-fidelity deep generative model is introduced for the surrogate modeling of high-fidelity turbulent flow fields given the solution of a computationally inexpensive but inaccurate low-fidelity solver.
 

docs/source/start.rst

@@ -20,8 +20,9 @@ Start with cloning the repository to your local machine using:
 2. Downloading Data
 -------------------
 To download both the training and testing data for both of the numerical examples, visit the following Zenodo repository.
-`https://doi.org/10.5281/zenodo.4298896 <https://doi.org/10.5281/zenodo.4298896>`_
+`https://doi.org/10.5281/zenodo.4311698 <https://doi.org/10.5281/zenodo.4311698>`_
 
+The data-sets are compressed tarfiles, notes that pre-trained models are also included in this repository.
 Extract the data using ``tar –xvzf data-set-name.tar.gz``, and move the training and testing data the desired directory.
 
 The default directories in the config are:
@@ -64,6 +65,12 @@ For example:
 -----------------------------
 Alternatively, you can skip the training and simply run a pre-trained models.
 These can be found in the ``examples`` folder where there are several scripts that demonstrate how to load and test a model.
+Download the pre-trained models from the Zenodo repository.
+These are the two files that are ``.zip`` folders, simply place them in the examples directory.
+
+.. warning::
+ Do **not** unzip the workspace files, they should be just copied as is. TM-Glow will unzip them for you and read necessary data.
+
 The general work flow of running a model for prediction is:
 
 1. Load model from the workspace zip file.
@@ -73,4 +80,5 @@ The general work flow of running a model for prediction is:
 
 .. note::
  These programs require you download the *testing* data from the Zenodo data repository. The default location this data should be unzipped to is discussed in step :ref:`download-data`.
- However, one can easily customize the data location in each script.
+ However, one can easily customize the data location in each script.
+

examples/backward-step/plotVelMagSamples.py → examples/plotBackStepVelocity.py

@@ -127,10 +127,9 @@ def plotPredictionSeries(input0, yPred0, yTarget0, bidx=0, nsteps=10, stride=1,
 
  log = Log(args, record=False)
 
- model_name = 'nsWorkspace'
- model_id = args.epoch_start
-
  # Load pretrained model parameters
+ model_name = 'bstepWorkspace'
+ model_id = 400
  model_path = '.'
  args, model_state_dict, optimizer_state_dict = loadWorkspace(args, model_path, file_name=model_name, file_id=model_id)
  scheduler = None

examples/cylinder-array/plotVelMagSamples.py → examples/plotCylinderVelocity.py

@@ -125,10 +125,9 @@ def plotPredictionSeries(input0, yPred0, yTarget0, bidx=0, nsteps=20, stride=1,
 
  log = Log(args, record=False)
 
- model_name = 'nsWorkspace'
- model_id = args.epoch_start
-
  # Load pretrained model parameters
+ model_name = 'cylinderWorkspace'
+ model_id = 400
  model_path='.'
  args, model_state_dict, optimizer_state_dict = loadWorkspace(args, model_path, file_name=model_name, file_id=model_id)
  scheduler = None