-
Notifications
You must be signed in to change notification settings - Fork 32
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Initial docs generation #32
Conversation
Hi guys, Thank you for the hard work you've been doing with MODNet, which I haven't been able to follow up very much. Just one general comment on documentation, it would be good to get off on the right foot from the beginning (of course :D). I've seen this blog and there is a nice talk from the guy at a pycon conference: https://documentation.divio.com/ Everything does not need to be there from the beginning but if the structure is laid out nicely from the beginning it will for sure help for the future. The "part" of the documentation that is covered by this pull request is mainly about the API reference from what I understand. As you can see in the blog above (and the pycon conference), there are 3 other parts that have other goals. David |
Just to put my suggestions to @ppdebreuck in public:
|
Update:
|
README.md
Outdated
@@ -19,9 +20,13 @@ | |||
|
|||
<a name="introduction"></a> | |||
## Introduction | |||
This repository contains the python package implementing the Material Optimal Descriptor Network (MODNet). It is a supervised machine learning framework for **learning material properties** from the **crystal structure**. The framework is well suited for **limited datasets** and can be used for learning *multiple* properties together by using **joint transfer learning**. | |||
This repository contains the python package implementing the Material Optimal Descriptor Network (MODNet). |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This repository contains the python package implementing the Material Optimal Descriptor Network (MODNet). | |
This repository contains the Python (3.8+) package implementing the Material Optimal Descriptor Network (MODNet). |
README.md
Outdated
This repository contains the python package implementing the Material Optimal Descriptor Network (MODNet). | ||
It is a supervised machine learning framework for **learning material properties** from | ||
either the **composition** or **crystal structure**. The framework is well suited for **limited datasets** | ||
and can be used for learning *multiple* properties together by using **joint transfer learning**. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I would suggest that joint and transfer learning are two separate techniques?
and can be used for learning *multiple* properties together by using **joint transfer learning**. | |
and can be used for learning *multiple* properties together by using **joint learning**. |
This repository also contains two pretrained models that can be used for predicting | ||
the refractive index and vibrational thermodynamics from any crystal structure. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
(Just removing the linebreak)
This repository also contains two pretrained models that can be used for predicting | |
the refractive index and vibrational thermodynamics from any crystal structure. | |
This repository also contains two pretrained models that can be used for predicting the refractive index and vibrational thermodynamics from any crystal structure. |
README.md
Outdated
@@ -34,7 +39,7 @@ See the MODNet papers and repositories below for more details: | |||
|
|||
|
|||
<p align='center'> | |||
<img src="img/MODNet_schematic.PNG" alt="MODNet schematic" /> | |||
<img src="img/MODNet_schematic.png" alt="MODNet schematic" /> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This image does not render for me when I "view file"
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
(just need to either revert this change or fix the image file extension)
The MP MODData on [figshare](https://figshare.com/articles/dataset/Materials_Project_MP_2018_6_MODData/12834275) | ||
(MP_2018.6) is very usefull for predicting a learned property on all structures from the Materials Project: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do we release the MP model too? The example line then shows model predictions...
The MP MODData on [figshare](https://figshare.com/articles/dataset/Materials_Project_MP_2018_6_MODData/12834275) | |
(MP_2018.6) is very usefull for predicting a learned property on all structures from the Materials Project: | |
The MP MODData on [figshare](https://figshare.com/articles/dataset/Materials_Project_MP_2018_6_MODData/12834275) (MP_2018.6) is very useful for predicting a learned property on all structures from the Materials Project: |
README.md
Outdated
|
||
Especially, carefully read the two main classes, `MODData` and `MODNetModel` found in preprocessing and models modules. | ||
A `MODData` instance is used for representing a particular dataset. It contains a list of structures and corresponding properties: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
A `MODData` instance is used for representing a particular dataset. It contains a list of structures and corresponding properties: | |
A `MODData` instance is used for representing a particular dataset. It contains a list of structures and corresponding properties. |
README.md
Outdated
@@ -173,7 +197,7 @@ from modnet.preprocessing import MODData | |||
data = MODData.load('path/dataname') | |||
``` | |||
|
|||
Both the save and load methods use pandas `.read_pickle(...)` and `.load_pickle(...)` which will try to compress/decompress files according to their file extensions (e.g. `".zip"`, `".tgz"` and `".bz2"`. | |||
Both the save and load methods use pandas `.read_pickle(...)` and `.load_pickle(...)` which will try to compress/decompress files according to their file extensions (e.g. `".zip"`, `".tgz"` and `".bz2"). |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Both the save and load methods use pandas `.read_pickle(...)` and `.load_pickle(...)` which will try to compress/decompress files according to their file extensions (e.g. `".zip"`, `".tgz"` and `".bz2"). | |
Both the save and load methods use pandas `.read_pickle(...)` and `.load_pickle(...)` which will try to compress/decompress files according to their file extensions (e.g. `".zip"`, `".tgz"` and `".bz2"). | |
Please note that this method is **unsafe** can load arbitrary Python objects. Care has been taken to check the hashes of data that is automatically loaded from Figshare; if you rely on this feature for your own data then we recommend you do the same. |
README.md
Outdated
model = MODNetModel(targets, | ||
weights, | ||
num_classes = None # only needed for classification, e.g. num_classes = {'is_metal':2} | ||
num_neurons=[[64],[32],[16],[16]], | ||
n_feat=300, | ||
act='relu', | ||
) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
model = MODNetModel(targets, | |
weights, | |
num_classes = None # only needed for classification, e.g. num_classes = {'is_metal':2} | |
num_neurons=[[64],[32],[16],[16]], | |
n_feat=300, | |
act='relu', | |
) | |
model = MODNetModel( | |
targets, | |
weights, | |
num_classes = None # only needed for classification, e.g. num_classes = {'is_metal':2} | |
num_neurons=[[64],[32],[16],[16]], | |
n_feat=300, | |
act='relu', | |
) |
README.md
Outdated
model.fit(data, | ||
val_fraction = 0.0, | ||
val_data = None, | ||
val_key = None, | ||
lr=0.001, | ||
epochs = 200, | ||
batch_size = 128, | ||
xscale='minmax', | ||
loss='mse', | ||
) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
model.fit(data, | |
val_fraction = 0.0, | |
val_data = None, | |
val_key = None, | |
lr=0.001, | |
epochs = 200, | |
batch_size = 128, | |
xscale='minmax', | |
loss='mse', | |
) | |
model.fit( | |
data, | |
val_fraction = 0.0, | |
val_data = None, | |
val_key = None, | |
lr=0.001, | |
epochs = 200, | |
batch_size = 128, | |
xscale='minmax', | |
loss='mse', | |
) |
- install - getting-started - support
Modnet documentation automatically generated by sphinx apidoc.
To be discussed: