megnet.layers package¶
Subpackages¶
Module contents¶
Megnet layers implementations. This subpackage includes
Graph convolution layers
Readout layers
-
class
CrystalGraphLayer
(*args, **kwargs)[source]¶ Bases:
megnet.layers.graph.base.GraphNetworkLayer
The CGCNN graph implementation as described in the paper
Xie et al. PHYSICAL REVIEW LETTERS 120, 145301 (2018)
-
call
(inputs, mask=None)¶ the logic of the layer, returns the final graph
-
compute_output_shape
(input_shape)[source]¶ compute static output shapes, returns list of tuple shapes
-
phi_v
(b_e_p, inputs)[source]¶ update the atom attributes by the results from previous step b_e_p and all the inputs returns v_p.
- Parameters
activation (str) – Default: None. The activation function used for each sub-neural network. Examples include ‘relu’, ‘softmax’, ‘tanh’, ‘sigmoid’ and etc.
use_bias (bool) – Default: True. Whether to use the bias term in the neural network.
kernel_initializer (str) – Default: ‘glorot_uniform’. Initialization function for the layer kernel weights,
bias_initializer (str) – Default: ‘zeros’
activity_regularizer (str) – Default: None. The regularization function for the output
kernel_constraint (str) – Default: None. Keras constraint for kernel values
bias_constraint (str) – Default: None .Keras constraint for bias values
kwargs (dictionary) – additional keyword args
-
build
(input_shapes)[source]¶ Build the weights for the layer :param input_shapes: the shapes of all input tensors :type input_shapes: sequence of tuple
-
compute_output_shape
(input_shape)[source]¶ Compute output shapes from input shapes :param input_shape: input shapes :type input_shape: sequence of tuple
Returns: sequence of tuples output shapes
-
get_config
()[source]¶ Part of keras layer interface, where the signature is converted into a dict
- Returns
configurational dictionary
-
phi_e
(inputs)[source]¶ Edge update function :param inputs: :type inputs: tuple of tensor
- Returns
output tensor
-
phi_u
(b_e_p, b_v_p, inputs)[source]¶ - Parameters
b_e_p (tf.Tensor) – edge/bond to global aggregated tensor
b_v_p (tf.Tensor) – node/atom to global aggregated tensor
inputs (Sequence) – list or tuple for the graph inputs
- Returns
updated globa/state attributes
-
phi_v
(b_ei_p, inputs)[source]¶ Node update function :param b_ei_p: edge aggregated tensor :type b_ei_p: tensor :param inputs: other graph inputs :type inputs: tuple of tensors
Returns: updated node tensor
-
rho_e_u
(e_p, inputs)[source]¶ aggregate edge to state :param e_p: edge tensor :type e_p: tensor :param inputs: other graph input tensors :type inputs: tuple of tensors
Returns: edge aggregated tensor for states
-
-
class
GaussianExpansion
(*args, **kwargs)[source]¶ Bases:
tensorflow.python.keras.engine.base_layer.Layer
Simple Gaussian expansion. A vector of distance [d1, d2, d3, …, dn] is expanded to a matrix of shape [n, m], where m is the number of Gaussian basis centers
- Parameters
centers (np.ndarray) – Gaussian basis centers
width (float) – width of the Gaussian basis
**kwargs –
-
build
(input_shape)[source]¶ build the layer :param input_shape: tuple of int for the input shape :type input_shape: tuple
-
class
InteractionLayer
(*args, **kwargs)[source]¶ Bases:
megnet.layers.graph.base.GraphNetworkLayer
The Continuous filter InteractionLayer in Schnet
Schütt et al. SchNet: A continuous-filter convolutional neural network for modeling quantum interactions
-
call
(inputs, mask=None)¶ the logic of the layer, returns the final graph
-
compute_output_shape
(input_shape)[source]¶ compute static output shapes, returns list of tuple shapes
-
phi_v
(b_e_p, inputs)[source]¶ update the atom attributes by the results from previous step b_e_p and all the inputs returns v_p.
- Parameters
activation (str) – Default: None. The activation function used for each sub-neural network. Examples include ‘relu’, ‘softmax’, ‘tanh’, ‘sigmoid’ and etc.
use_bias (bool) – Default: True. Whether to use the bias term in the neural network.
kernel_initializer (str) – Default: ‘glorot_uniform’. Initialization function for the layer kernel weights,
bias_initializer (str) – Default: ‘zeros’
activity_regularizer (str) – Default: None. The regularization function for the output
kernel_constraint (str) – Default: None. Keras constraint for kernel values
bias_constraint (str) – Default: None .Keras constraint for bias values
-
build
(input_shapes)[source]¶ Build the weights for the layer :param input_shapes: the shapes of all input tensors :type input_shapes: sequence of tuple
-
compute_output_shape
(input_shape)[source]¶ Compute output shapes from input shapes :param input_shape: input shapes :type input_shape: sequence of tuple
Returns: sequence of tuples output shapes
-
get_config
()[source]¶ Part of keras layer interface, where the signature is converted into a dict
- Returns
configurational dictionary
-
phi_e
(inputs)[source]¶ Edge update function :param inputs: :type inputs: tuple of tensor
- Returns
output tensor
-
phi_u
(b_e_p, b_v_p, inputs)[source]¶ - Parameters
b_e_p (tf.Tensor) – edge/bond to global aggregated tensor
b_v_p (tf.Tensor) – node/atom to global aggregated tensor
inputs (Sequence) – list or tuple for the graph inputs
- Returns
updated globa/state attributes
-
phi_v
(b_ei_p, inputs)[source]¶ Node update function :param b_ei_p: edge aggregated tensor :type b_ei_p: tensor :param inputs: other graph inputs :type inputs: tuple of tensors
Returns: updated node tensor
-
rho_e_u
(e_p, inputs)[source]¶ aggregate edge to state :param e_p: edge tensor :type e_p: tensor :param inputs: other graph input tensors :type inputs: tuple of tensors
Returns: edge aggregated tensor for states
-
-
class
LinearWithIndex
(*args, **kwargs)[source]¶ Bases:
tensorflow.python.keras.engine.base_layer.Layer
Sum or average the node/edge attributes to get a structure-level vector
- Parameters
mode – (str) ‘mean’, ‘sum’, ‘max’, ‘mean’ or ‘prod’
**kwargs –
-
build
(input_shape)[source]¶ Build tensors :param input_shape: input shapes :type input_shape: sequence of tuple
-
call
(inputs, mask=None)[source]¶ Main logic :param inputs: input tensors :type inputs: tuple of tensor :param mask: mask tensor :type mask: tensor
Returns: output tensor
-
class
MEGNetLayer
(*args, **kwargs)[source]¶ Bases:
megnet.layers.graph.base.GraphNetworkLayer
The MEGNet graph implementation as described in the paper
Chen, Chi; Ye, Weike Ye; Zuo, Yunxing; Zheng, Chen; Ong, Shyue Ping. Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals, 2018, arXiv preprint. [arXiv:1812.05055](https://arxiv.org/abs/1812.05055) .. method:: call(inputs, mask=None)
the logic of the layer, returns the final graph
-
compute_output_shape
(input_shape)[source]¶ compute static output shapes, returns list of tuple shapes
-
phi_v
(b_e_p, inputs)[source]¶ update the atom attributes by the results from previous step b_e_p and all the inputs returns v_p.
- Parameters
units_v (list of integers) – the hidden layer sizes for node update neural network
units_e (list of integers) – the hidden layer sizes for edge update neural network
units_u (list of integers) – the hidden layer sizes for state update neural network
pool_method (str) – ‘mean’ or ‘sum’, determines how information is gathered to nodes from neighboring edges
activation (str) – Default: None. The activation function used for each sub-neural network. Examples include ‘relu’, ‘softmax’, ‘tanh’, ‘sigmoid’ and etc.
use_bias (bool) – Default: True. Whether to use the bias term in the neural network.
kernel_initializer (str) – Default: ‘glorot_uniform’. Initialization function for the layer kernel weights,
bias_initializer (str) – Default: ‘zeros’
activity_regularizer (str) – Default: None. The regularization function for the output
kernel_constraint (str) – Default: None. Keras constraint for kernel values
bias_constraint (str) – Default: None .Keras constraint for bias values
-
build
(input_shapes)[source]¶ Build the weights for the layer :param input_shapes: the shapes of all input tensors :type input_shapes: sequence of tuple
-
compute_output_shape
(input_shape)[source]¶ Compute output shapes from input shapes :param input_shape: input shapes :type input_shape: sequence of tuple
Returns: sequence of tuples output shapes
-
get_config
()[source]¶ Part of keras layer interface, where the signature is converted into a dict
- Returns
configurational dictionary
-
phi_e
(inputs)[source]¶ Edge update function :param inputs: :type inputs: tuple of tensor
- Returns
output tensor
-
phi_u
(b_e_p, b_v_p, inputs)[source]¶ - Parameters
b_e_p (tf.Tensor) – edge/bond to global aggregated tensor
b_v_p (tf.Tensor) – node/atom to global aggregated tensor
inputs (Sequence) – list or tuple for the graph inputs
- Returns
updated globa/state attributes
-
phi_v
(b_ei_p, inputs)[source]¶ Node update function :param b_ei_p: edge aggregated tensor :type b_ei_p: tensor :param inputs: other graph inputs :type inputs: tuple of tensors
Returns: updated node tensor
-
rho_e_u
(e_p, inputs)[source]¶ aggregate edge to state :param e_p: edge tensor :type e_p: tensor :param inputs: other graph input tensors :type inputs: tuple of tensors
Returns: edge aggregated tensor for states
-
-
class
Set2Set
(*args, **kwargs)[source]¶ Bases:
tensorflow.python.keras.engine.base_layer.Layer
For a set of vectors, the set2set neural network maps it to a single vector. The order invariance is acheived by a attention mechanism. See Vinyals, Oriol, Samy Bengio, and Manjunath Kudlur. “Order matters: Sequence to sequence for sets.” arXiv preprint arXiv:1511.06391 (2015).
- Parameters
T – (int) recurrent step
n_hidden – (int) number of hidden units
activation – (str or object) activation function
activation_lstm – (str or object) activation function for lstm
recurrent_activation – (str or object) activation function for recurrent step
kernel_initializer – (str or object) initializer for kernel weights
recurrent_initializer – (str or object) initializer for recurrent weights
bias_initializer – (str or object) initializer for biases
use_bias – (bool) whether to use biases
unit_forget_bias – (bool) whether to use basis in forget gate
kernel_regularizer – (str or object) regularizer for kernel weights
recurrent_regularizer – (str or object) regularizer for recurrent weights
bias_regularizer – (str or object) regularizer for biases
kernel_constraint – (str or object) constraint for kernel weights
recurrent_constraint – (str or object) constraint for recurrent weights
bias_constraint – (str or object) constraint for biases
kwargs – other inputs for keras Layer class
-
build
(input_shape)[source]¶ Build tensors :param input_shape: input shapes :type input_shape: sequence of tuple
-
call
(inputs, mask=None)[source]¶ Main logic :param inputs: input tensors :type inputs: tuple of tensor :param mask: mask tensor :type mask: tensor
Returns: output tensor
-
keras_layer_deserialize
(config, custom_objects=None)¶ Instantiates a layer from a config dictionary.
- Parameters
config – dict of the form {‘class_name’: str, ‘config’: dict}
custom_objects – dict mapping class names (or function names) of custom (non-Keras) objects to class/functions
- Returns
Layer instance (may be Model, Sequential, Network, Layer…)
-
mean_squared_error_with_scale
(y_true, y_pred, scale=10000)[source]¶ Keras default log for tracking progress shows two decimal points, here we multiply the mse by a factor to fully show the loss in progress bar
- Parameters
y_true – (tensor) training y
y_pred – (tensor) predicted y
scale – (int or float) factor to multiply with mse
- Returns
scaled mse (float)