We envision models that are pre-trained on a vast range of domain-relevant tasks to become key for molecule property prediction

Overview

HuggingMolecules

License

We envision models that are pre-trained on a vast range of domain-relevant tasks to become key for molecule property prediction. This repository aims to give easy access to state-of-the-art pre-trained models.

Quick tour

To quickly fine-tune a model on a dataset using the pytorch lightning package follow the below example based on the MAT model and the freesolv dataset:

from huggingmolecules import MatModel, MatFeaturizer

# The following import works only from the source code directory:
from experiments.src import TrainingModule, get_data_loaders

from torch.nn import MSELoss
from torch.optim import Adam

from pytorch_lightning import Trainer
from pytorch_lightning.metrics import MeanSquaredError

# Build and load the pre-trained model and the appropriate featurizer:
model = MatModel.from_pretrained('mat_masking_20M')
featurizer = MatFeaturizer.from_pretrained('mat_masking_20M')

# Build the pytorch lightning training module:
pl_module = TrainingModule(model,
                           loss_fn=MSELoss(),
                           metric_cls=MeanSquaredError,
                           optimizer=Adam(model.parameters()))

# Build the data loader for the freesolv dataset:
train_dataloader, _, _ = get_data_loaders(featurizer,
                                          batch_size=32,
                                          task_name='ADME',
                                          dataset_name='hydrationfreeenergy_freesolv')

# Build the pytorch lightning trainer and fine-tune the module on the train dataset:
trainer = Trainer(max_epochs=100)
trainer.fit(pl_module, train_dataloader=train_dataloader)

# Make the prediction for the batch of SMILES strings:
batch = featurizer(['C/C=C/C', '[C]=O'])
output = pl_module.model(batch)

Installation

Create your conda environment and install the rdkit package:

conda create -n huggingmolecules python=3.8.5
conda activate huggingmolecules
conda install -c conda-forge rdkit==2020.09.1

Then install huggingmolecules from the cloned directory:

conda activate huggingmolecules
pip install -e ./src

Project Structure

The project consists of two main modules: src/ and experiments/ modules:

  • The src/ module contains abstract interfaces for pre-trained models along with their implementations based on the pytorch library. This module makes configuring, downloading and running existing models easy and out-of-the-box.
  • The experiments/ module makes use of abstract interfaces defined in the src/ module and implements scripts based on the pytorch lightning package for running various experiments. This module makes training, benchmarking and hyper-tuning of models flawless and easily extensible.

Supported models architectures

Huggingmolecules currently provides the following models architectures:

For ease of benchmarking, we also include wrappers in the experiments/ module for three other models architectures:

The src/ module

The implementations of the models in the src/ module are divided into three modules: configuration, featurization and models module. The relation between these modules is shown on the following examples based on the MAT model:

Configuration examples

from huggingmolecules import MatConfig

# Build the config with default parameters values, 
# except 'd_model' parameter, which is set to 1200:
config = MatConfig(d_model=1200)

# Build the pre-defined config:
config = MatConfig.from_pretrained('mat_masking_20M')

# Build the pre-defined config with 'init_type' parameter set to 'normal':
config = MatConfig.from_pretrained('mat_masking_20M', init_type='normal')

# Save the pre-defined config with the previous modification:
config.save_to_cache('mat_masking_20M_normal.json')

# Restore the previously saved config:
config = MatConfig.from_pretrained('mat_masking_20M_normal.json')

Featurization examples

from huggingmolecules import MatConfig, MatFeaturizer

# Build the featurizer with pre-defined config:
config = MatConfig.from_pretrained('mat_masking_20M')
featurizer = MatFeaturizer(config)

# Build the featurizer in one line:
featurizer = MatFeaturizer.from_pretrained('mat_masking_20M')

# Encode (featurize) the batch of two SMILES strings: 
batch = featurizer(['C/C=C/C', '[C]=O'])

Models examples

from huggingmolecules import MatConfig, MatFeaturizer, MatModel

# Build the model with the pre-defined config:
config = MatConfig.from_pretrained('mat_masking_20M')
model = MatModel(config)

# Load the pre-trained weights 
# (which do not include the last layer of the model)
model.load_weights('mat_masking_20M')

# Build the model and load the pre-trained weights in one line:
model = MatModel.from_pretrained('mat_masking_20M')

# Encode (featurize) the batch of two SMILES strings: 
featurizer = MatFeaturizer.from_pretrained('mat_masking_20M')
batch = featurizer(['C/C=C/C', '[C]=O'])

# Feed the model with the encoded batch:
output = model(batch)

# Save the weights of the model (usually after the fine-tuning process):
model.save_weights('tuned_mat_masking_20M.pt')

# Load the previously saved weights
# (which now includes all layers of the model):
model.load_weights('tuned_mat_masking_20M.pt')

# Load the previously saved weights, but without 
# the last layer of the model ('generator' in the case of the 'MatModel')
model.load_weights('tuned_mat_masking_20M.pt', excluded=['generator'])

# Build the model and load the previously saved weights:
config = MatConfig.from_pretrained('mat_masking_20M')
model = MatModel.from_pretrained('tuned_mat_masking_20M.pt',
                                 excluded=['generator'],
                                 config=config)

Running tests

To run base tests for src/ module, type:

pytest src/ --ignore=src/tests/downloading/

To additionally run tests for downloading module (which will download all models to your local computer and therefore may be slow), type:

pytest src/tests/downloading

The experiments/ module

Requirements

In addition to dependencies defined in the src/ module, the experiments/ module goes along with few others. To install them, run:

pip install -r experiments/requirements.txt

The following packages are crucial for functioning of the experiments/ module:

Neptune.ai

In addition, we recommend installing the neptune.ai package:

  1. Sign up to neptune.ai at https://neptune.ai/.

  2. Get your Neptune API token (see getting-started for help).

  3. Export your Neptune API token to NEPTUNE_API_TOKEN environment variable.

  4. Install neptune-client: pip install neptune-client.

  5. Enable neptune.ai in the experiments/configs/setup.gin file.

  6. Update neptune.project_name parameters in experiments/configs/bases/*.gin files.

Running scripts:

We recommend running experiments scripts from the source code. For the moment there are three scripts implemented:

  • experiments/scripts/train.py - for training with the pytorch lightning package
  • experiments/scripts/tune_hyper.py - for hyper-parameters tuning with the optuna package
  • experiments/scripts/benchmark.py - for benchmarking based on the hyper-parameters tuning (grid-search)

In general running scripts can be done with the following syntax:

python -m experiments.scripts. /
       -d  / 
       -m  /
       -b 

Then the script .py runs with functions/methods parameters values defined in the following gin-config files:

  1. experiments/configs/bases/.gin
  2. experiments/configs/datasets/.gin
  3. experiments/configs/models/.gin

If the binding flag -b is used, then bindings defined in overrides corresponding bindings defined in above gin-config files.

So for instance, to fine-tune the MAT model (pre-trained on masking_20M task) on the freesolv dataset using GPU 1, simply run:

python -m experiments.scripts.train /
       -d freesolv / 
       -m mat /
       -b model.pretrained_name=\"mat_masking_20M\"#train.gpus=[1]

or equivalently:

python -m experiments.scripts.train /
       -d freesolv / 
       -m mat /
       --model.pretrained_name mat_masking_20M /
       --train.gpus [1]

Local dataset

To use a local dataset, create an appropriate gin-config file in the experiments/configs/datasets directory and specify the data.data_path parameter within. For details see the get_data_split implementation.

Benchmarking

For the moment there is one benchmark available. It works as follows:

  • experiments/scripts/benchmark.py: on the given dataset we fine-tune the given model on 10 learning rates and 6 seeded data splits (60 fine-tunings in total). Then we choose that learning rate that minimizes an averaged (on 6 data splits) validation metric (metric computed on the validation dataset, e.g. RMSE). The result is the averaged value of test metric for the chosen learning rate.

Running a benchmark is essentially the same as running any other script from the experiments/ module. So for instance to benchmark the vanilla MAT model (without pre-training) on the Caco-2 dataset using GPU 0, simply run:

python -m experiments.scripts.benchmark /
       -d caco2 / 
       -m mat /
       --model.pretrained_name None /
       --train.gpus [0]

However, the above script will only perform 60 fine-tunings. It won't compute the final benchmark result. To do that wee need to run:

python -m experiments.scripts.benchmark --results_only /
       -d caco2 / 
       -m mat

The above script won't perform any fine-tuning, but will only compute the benchmark result. If we had neptune enabled in experiments/configs/setup.gin, all data necessary to compute the result will be fetched from the neptune server.

Benchmark results

We performed the benchmark described in Benchmarking as experiments/scripts/benchmark.py for various models architectures and pre-training tasks.

Summary

We report mean/median ranks of tested models across all datasets (both regression and classification ones). For detailed results see Regression and Classification sections.

model mean rank rank std
MAT 200k 5.6 3.5
MAT 2M 5.3 3.4
MAT 20M 4.1 2.2
GROVER Base 3.8 2.7
GROVER Large 3.6 2.4
ChemBERTa 7.4 2.8
MolBERT 5.9 2.9
D-MPNN 6.3 2.3
D-MPNN 2d 6.4 2.0
D-MPNN mc 5.3 2.1

Regression

As the metric we used MAE for QM7 and RMSE for the rest of datasets.

model FreeSolv Caco-2 Clearance QM7 Mean rank
MAT 200k 0.913 ± 0.196 0.405 ± 0.030 0.649 ± 0.341 87.578 ± 15.375 5.25
MAT 2M 0.898 ± 0.165 0.471 ± 0.070 0.655 ± 0.327 81.557 ± 5.088 6.75
MAT 20M 0.854 ± 0.197 0.432 ± 0.034 0.640 ± 0.335 81.797 ± 4.176 5.0
Grover Base 0.917 ± 0.195 0.419 ± 0.029 0.629 ± 0.335 62.266 ± 3.578 3.25
Grover Large 0.950 ± 0.202 0.414 ± 0.041 0.627 ± 0.340 64.941 ± 3.616 2.5
ChemBERTa 1.218 ± 0.245 0.430 ± 0.013 0.647 ± 0.314 177.242 ± 1.819 8.0
MolBERT 1.027 ± 0.244 0.483 ± 0.056 0.633 ± 0.332 177.117 ± 1.799 8.0
Chemprop 1.061 ± 0.168 0.446 ± 0.064 0.628 ± 0.339 74.831 ± 4.792 5.5
Chemprop 2d 1 1.038 ± 0.235 0.454 ± 0.049 0.628 ± 0.336 77.912 ± 10.231 6.0
Chemprop mc 2 0.995 ± 0.136 0.438 ± 0.053 0.627 ± 0.337 75.575 ± 4.683 4.25

1 chemprop with additional rdkit_2d_normalized features generator
2 chemprop with additional morgan_count features generator

Classification

We used ROC AUC as the metric.

model HIA Bioavailability PPBR Tox21 (NR-AR) BBBP Mean rank
MAT 200k 0.943 ± 0.015 0.660 ± 0.052 0.896 ± 0.027 0.775 ± 0.035 0.709 ± 0.022 5.8
MAT 2M 0.941 ± 0.013 0.712 ± 0.076 0.905 ± 0.019 0.779 ± 0.056 0.713 ± 0.022 4.2
MAT 20M 0.935 ± 0.017 0.732 ± 0.082 0.891 ± 0.019 0.779 ± 0.056 0.735 ± 0.006 3.4
Grover Base 0.931 ± 0.021 0.750 ± 0.037 0.901 ± 0.036 0.750 ± 0.085 0.735 ± 0.006 4.0
Grover Large 0.932 ± 0.023 0.747 ± 0.062 0.901 ± 0.033 0.757 ± 0.057 0.757 ± 0.057 4.2
ChemBERTa 0.923 ± 0.032 0.666 ± 0.041 0.869 ± 0.032 0.779 ± 0.044 0.717 ± 0.009 7.0
MolBERT 0.942 ± 0.011 0.737 ± 0.085 0.889 ± 0.039 0.761 ± 0.058 0.742 ± 0.020 4.6
Chemprop 0.924 ± 0.069 0.724 ± 0.064 0.847 ± 0.052 0.766 ± 0.040 0.726 ± 0.008 7.0
Chemprop 2d 0.923 ± 0.015 0.712 ± 0.067 0.874 ± 0.030 0.775 ± 0.041 0.724 ± 0.006 6.8
Chemprop mc 0.924 ± 0.082 0.740 ± 0.060 0.869 ± 0.033 0.772 ± 0.041 0.722 ± 0.008 6.2
Owner
GMUM
Group of Machine Learning Research, Jagiellonian University
GMUM
This is the code of using DQN to play Sekiro .

Update for using DQN to play sekiro 2021.2.2(English Version) This is the code of using DQN to play Sekiro . I am very glad to tell that I have writen

144 Dec 25, 2022
本步态识别系统主要基于GaitSet模型进行实现

本步态识别系统主要基于GaitSet模型进行实现。在尝试部署本系统之前,建立理解GaitSet模型的网络结构、训练和推理方法。 系统的实现效果如视频所示: 演示视频 由于模型较大,部分模型文件存储在百度云盘。 链接提取码:33mb 具体部署过程 1.下载代码 2.安装requirements.txt

16 Oct 22, 2022
The Pytorch code of "Joint Distribution Matters: Deep Brownian Distance Covariance for Few-Shot Classification", CVPR 2022 (Oral).

DeepBDC for few-shot learning        Introduction In this repo, we provide the implementation of the following paper: "Joint Distribution Matters: Dee

FeiLong 116 Dec 19, 2022
Code for Deep Single-image Portrait Image Relighting

Deep Single-Image Portrait Relighting [Project Page] Hao Zhou, Sunil Hadap, Kalyan Sunkavalli, David W. Jacobs. In ICCV, 2019 Overview Test script for

438 Jan 05, 2023
A tight inclusion function for continuous collision detection

Tight-Inclusion Continuous Collision Detection A conservative Continuous Collision Detection (CCD) method with support for minimum separation. You can

Continuous Collision Detection 89 Jan 01, 2023
Learning to trade under the reinforcement learning framework

Trading Using Q-Learning In this project, I will present an adaptive learning model to trade a single stock under the reinforcement learning framework

Uirá Caiado 470 Nov 28, 2022
Code accompanying the paper "Wasserstein GAN"

Wasserstein GAN Code accompanying the paper "Wasserstein GAN" A few notes The first time running on the LSUN dataset it can take a long time (up to an

3.1k Jan 01, 2023
[ICCV 2021] Official Pytorch implementation for Discriminative Region-based Multi-Label Zero-Shot Learning SOTA results on NUS-WIDE and OpenImages

Discriminative Region-based Multi-Label Zero-Shot Learning (ICCV 2021) [arXiv][Project page coming soon] Sanath Narayan*, Akshita Gupta*, Salman Kh

Akshita Gupta 54 Nov 21, 2022
This repository is for our EMNLP 2021 paper "Automated Generation of Accurate & Fluent Medical X-ray Reports"

Introduction: X-Ray Report Generation This repository is for our EMNLP 2021 paper "Automated Generation of Accurate & Fluent Medical X-ray Reports". O

no name 36 Dec 16, 2022
Simple implementation of Mobile-Former on Pytorch

Simple-implementation-of-Mobile-Former At present, only the model but no trained. There may be some bug in the code, and some details may be different

Acheung 103 Dec 31, 2022
Re-implement CycleGAN in Tensorlayer

CycleGAN_Tensorlayer Re-implement CycleGAN in TensorLayer Original CycleGAN Improved CycleGAN with resize-convolution Prerequisites: TensorLayer Tenso

89 Aug 15, 2022
Artstation-Artistic-face-HQ Dataset (AAHQ)

Artstation-Artistic-face-HQ Dataset (AAHQ) Artstation-Artistic-face-HQ (AAHQ) is a high-quality image dataset of artistic-face images. It is proposed

onion 105 Dec 16, 2022
Official implementation of the paper "AAVAE: Augmentation-AugmentedVariational Autoencoders"

AAVAE Official implementation of the paper "AAVAE: Augmentation-AugmentedVariational Autoencoders" Abstract Recent methods for self-supervised learnin

Grid AI Labs 48 Dec 12, 2022
pip install python-office

🍬 python for office 👉 http://www.python4office.cn/ 👈 🌎 English Documentation 📚 简介 Python-office 是一个 Python 自动化办公第三方库,能解决大部分自动化办公的问题。而且每个功能只需一行代码,

程序员晚枫 272 Dec 29, 2022
Repository for the paper : Meta-FDMixup: Cross-Domain Few-Shot Learning Guided byLabeled Target Data

1 Meta-FDMIxup Repository for the paper : Meta-FDMixup: Cross-Domain Few-Shot Learning Guided byLabeled Target Data. (ACM MM 2021) paper News! the rep

Fu Yuqian 44 Nov 18, 2022
3D-CariGAN: An End-to-End Solution to 3D Caricature Generation from Normal Face Photos

3D-CariGAN: An End-to-End Solution to 3D Caricature Generation from Normal Face Photos This repository contains the source code and dataset for the pa

54 Oct 09, 2022
(to be released) [NeurIPS'21] Transformers Generalize DeepSets and Can be Extended to Graphs and Hypergraphs

Higher-Order Transformers Kim J, Oh S, Hong S, Transformers Generalize DeepSets and Can be Extended to Graphs and Hypergraphs, NeurIPS 2021. [arxiv] W

Jinwoo Kim 44 Dec 28, 2022
Kroomsa: A search engine for the curious

Kroomsa A search engine for the curious. It is a search algorithm designed to en

Wingify 7 Jun 20, 2022
My tensorflow implementation of "A neural conversational model", a Deep learning based chatbot

Deep Q&A Table of Contents Presentation Installation Running Chatbot Web interface Results Pretrained model Improvements Upgrade Presentation This wor

Conchylicultor 2.9k Dec 28, 2022
Official implementation for TTT++: When Does Self-supervised Test-time Training Fail or Thrive

TTT++ This is an official implementation for TTT++: When Does Self-supervised Test-time Training Fail or Thrive? TL;DR: Online Feature Alignment + Str

VITA lab at EPFL 39 Dec 25, 2022