Tutorials

1. Introduction To Deepchem

1. The Basic Tools of the Deep Life Sciences

2. Working With Datasets

3. An Introduction To MoleculeNet

4. Molecular Fingerprints

5. Creating Models with TensorFlow and PyTorch

6. Introduction to Graph Convolutions

7. Going Deeper on Molecular Featurizations

8. Working With Splitters

9. Advanced Model Training

10. Creating a high fidelity model from experimental data

11. Putting Multitask Learning to Work

12. Modeling Protein Ligand Interactions

13. Modeling Protein Ligand Interactions With Atomic Convolutions

14. Conditional Generative Adversarial Networks

15. Training a Generative Adversarial Network on MNIST

16. Advanced model training using hyperopt

17. Introduction to Gaussian Processes

18. PytorchLightning Integration

19. Compiling DeepChem Torch Models

2. Molecular Machine Learning

1. Molecular Fingerprints

2. Going Deeper on Molecular Featurizations

3. Learning Unsupervised Embeddings for Molecules

4. Synthetic Feasibility Scoring

5. Atomic Contributions for Molecules

6. Interactive Model Evaluation with Trident Chemwidgets

7. Transfer Learning With ChemBERTa Transformers

8. Training a Normalizing Flow on QM9

9. Large Scale Chemical Screens

10. Introduction to Molecular Attention Transformer

11. Generating molecules with MolGAN

12. Introduction to GROVER

13. Introduction to PROTACs

14. Druggablity Assessment with Fpocket and Machine Learning

3. Modeling Proteins

1. Protein Deep Learning

2. DeepChem AntibodyTutorial Simplified

3. Protein Structure Prediction with ESMFold

4. Introduction to ProtBERT

5. ProteinLM Tutorial0

4. Protein Ligand Modeling

1. Introduction to Binding Sites

2. Modeling Protein Ligand Interactions

3. Modeling Protein Ligand Interactions With Atomic Convolutions

4. DeepChemXAlphafold

5. UniProt Data Preprocessing for Binding Sites

5. Quantum Chemistry

1. Exploring Quantum Chemistry with GDB1k

2. DeepQMC tutorial

3. Training an Exchange Correlation Functional using Deepchem

6. Bioinformatics

1. Introduction to Bioinformatics

2. Multisequence Alignments

3. Scanpy

4. Deep probabilistic analysis of single-cell omics data

5. Cell Counting Tutorial

7. Material Science

1. Introduction To Material Science

8. Machine Learning Methods

1. Using Reinforcement Learning to Play Pong

2. Introduction to Model Interpretability

3. Uncertainty In Deep Learning

9. Deep Differential Equations

1. Physics Informed Neural Networks

2. Introducing JaxModel and PINNModel

3. About nODE Using Torchdiffeq in Deepchem

4. Differentiation Infrastructure in Deepchem

5. Ordinary Differential Equation Solving using deepchem

10. Equivariance

1. Introduction to Equivariance

11. Olfaction

1. Predict Multi Label Odor Descriptors using OpenPOM

12. Polymer Science

1. An Introduction to the Polymers and Their Representation

2. crystallization tendency regression

3. Introduction to Polymer SMILES

4. Understanding Weighted Directed Graphs for Polymer Implimentations

5. PolyBERT Introduction to Chemical Language Models for Fingerprint Generation

Introduction to the Molecular Attention Transformer. ¶

In this tutorial we will learn more about the Molecular Attention Transformer, or MAT. MAT is a model based on transformers, aimed towards performing molecular prediction tasks. MAT is easy to tune and performs quite well relative to other molecular prediction tasks. The weights from MAT are chemically interpretable, thus making the model quite useful.

Reference Paper: Molecular Attention Transformer, Maziarka et. al.

In this tutorial, we will explore how to train MAT, and predict hydration enthalpy values for molecules from the freesolv hydration enthalpy dataset with MAT.

Colab ¶

This tutorial and the rest in this sequence are designed to be done in Google colab. If you'd like to open this notebook in colab, you can use the following link.

In [ ]:

!pip install --pre deepchem

Import required modules ¶

In [1]:

import deepchem as dc
from deepchem.models.torch_models import MATModel
from deepchem.feat import MATFeaturizer
import matplotlib.pyplot as plt

wandb: WARNING W&B installed but not logged in.  Run `wandb login` or set the WANDB_API_KEY env variable.
wandb: WARNING W&B installed but not logged in.  Run `wandb login` or set the WANDB_API_KEY env variable.

Molecule Featurization using MATFeaturizer ¶

MATFeaturizer is the featurizer intended to be used with the Molecular Attention Transformer, or MAT in short. MATFeaturizer takes a smile string or molecule as input and returns a MATEncoding dataclass object, which contains 3 numpy arrays: Node features matrix, adjacency matrix and distance matrix.

In [2]:

featurizer = dc.feat.MATFeaturizer()
# Let us now take an example array of smile strings and featurize it.
smile_string = ["CCC"]
output = featurizer.featurize(smile_string)
print(type(output[0]))
print(output[0].node_features)
print(output[0].adjacency_matrix)
print(output[0].distance_matrix)

<class 'deepchem.feat.molecule_featurizers.mat_featurizer.MATEncoding'>
[[1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
  0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]
 [0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.
  0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]
 [0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.
  0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]
 [0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 1. 0. 0. 0.
  0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]]
[[0. 0. 0. 0.]
 [0. 0. 0. 1.]
 [0. 0. 0. 1.]
 [0. 1. 1. 0.]]
[[1.e+06 1.e+06 1.e+06 1.e+06]
 [1.e+06 0.e+00 2.e+00 1.e+00]
 [1.e+06 2.e+00 0.e+00 1.e+00]
 [1.e+06 1.e+00 1.e+00 0.e+00]]

Getting the Freesolv Hydration enthalpy dataset ¶

We will now acquire the Freesolv Hydration enthalpy dataset from MoleculeNet. If it already exists in the directory, the file will be used. Else, deepchem will automatically download the dataset from its AWS bucket.

In [3]:

tasks, dataset, transformers = dc.molnet.load_freesolv()
train_dataset, val_dataset, test_dataset = dataset
train_smiles = train_dataset.ids
val_smiles = val_dataset.ids

In [4]:

train_dataset

Out[4]:

<DiskDataset X.shape: (513,), y.shape: (513, 1), w.shape: (513, 1), ids: ['CCCCNCCCC' 'CCOC=O' 'CCCCCCCCC' ... 'COC' 'CCCCCCCCBr'
 'CCCc1ccc(c(c1)OC)O'], task_names: ['y']>

Training the model ¶

Now that we have acquired the dataset to be used and made the necessary imports, we will be invoking the Molecular Attention Transformer in deepchem, called MATModel, and we will be training it. We will be using default parameters for the purposes of this tutorial, however they can be changed anytime according to the user's preferences.

In [5]:

device = 'cpu'
model = MATModel(device = device)

In [6]:

losses, val_losses = [], []

In [7]:

%%time
max_epochs = 10

for epoch in range(max_epochs):
    loss = model.fit(train_dataset, nb_epoch = 1, max_checkpoints_to_keep = 1, all_losses = losses)
    metric = dc.metrics.Metric(dc.metrics.score_function.rms_score)
    val_losses.append(model.evaluate(val_dataset, metrics = [metric])['rms_score']**2)

# The warnings are not relevant to this tutorial thus we can safely skip them.

/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(

CPU times: user 20min 48s, sys: 10.7 s, total: 20min 58s
Wall time: 2min 41s

In [8]:

f, ax = plt.subplots()
ax.scatter(range(len(losses)), losses, label='train loss')
ax.scatter(range(len(val_losses)), val_losses, label='val loss')
plt.legend(loc='upper right');

No description has been provided for this image

Testing the model ¶

Optimally, MAT should be trained for a lot more epochs with a GPU. Due to computational constraints, we train this model for very few epochs in this tutorial. Let us now see how to predict the hydration enthalpy values for molecues now with MAT.

In [9]:

# We will be predicting the enthalpy value for the smile string we featurized earlier in the MATFeaturizer section.
model.predict_on_batch(output)

/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:165: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  node_features = torch.tensor(data[0]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:166: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  adjacency_matrix = torch.tensor(data[1]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/mat.py:167: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(data[2]).float()
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:171: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  torch.sum(torch.tensor(adj_matrix), dim=-1).unsqueeze(2) + eps)
/home/atreyamaj/Desktop/deepchem/deepchem/models/torch_models/layers.py:178: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
  distance_matrix = torch.tensor(distance_matrix).squeeze().masked_fill(

Out[9]:

array([[-0.32668447]], dtype=float32)