[](https://torchdrug.ai/)
<h1 align="center">
with
<a href="https://torchprotein.ai/">
<img src="asset/torchprotein_logo_tight.svg" alt="TorchProtein" style="height:26px" />
</a>
</h1>
[](https://colab.research.google.com/drive/1Tbnr1Fog_YjkqU1MOhcVLuxqZ4DC-c8-#forceEdit=true&sandboxMode=true)
[](https://github.com/DeepGraphLearning/torchdrug/blob/master/CONTRIBUTING.md)
[](https://github.com/DeepGraphLearning/torchdrug/blob/master/LICENSE)
[](https://pypi.org/project/torchdrug/)
[](https://twitter.com/DrugTorch)
[Docs] | [Tutorials] | [Benchmarks] | [Papers Implemented]
[Docs]: https://deepgraphlearning.github.io/torchdrug-site/docs
[Tutorials]: https://deepgraphlearning.github.io/torchdrug-site/docs/tutorials
[Benchmarks]: https://deepgraphlearning.github.io/torchdrug-site/docs/benchmark
[Papers Implemented]: https://deepgraphlearning.github.io/torchdrug-site/docs/paper
TorchDrug is a [PyTorch]-based machine learning toolbox designed for several purposes.
- Easy implementation of graph operations in a PyTorchic style with GPU support
- Being friendly to practitioners with minimal knowledge about drug discovery
- Rapid prototyping of machine learning research
[PyTorch]: https://pytorch.org/
Installation
------------
TorchDrug can be installed on either Linux, Windows or macOS. It is compatible with
3.7 <= Python <= 3.10 and PyTorch >= 1.8.0.
### From Conda ###
```bash
conda install torchdrug -c milagraph -c conda-forge -c pytorch -c pyg
```
### From Pip ###
```bash
pip install torch==1.9.0
pip install torch-scatter torch-cluster -f https://pytorch-geometric.com/whl/torch-1.9.0+cu102.html
pip install torchdrug
```
To install `torch-scatter` for other PyTorch or CUDA versions, please see the
instructions in https://github.com/rusty1s/pytorch_scatter
### From Source ###
```bash
git clone https://github.com/DeepGraphLearning/torchdrug
cd torchdrug
pip install -r requirements.txt
python setup.py install
```
### Windows (PowerShell) ###
We need to first install the build tools for Visual Studio. We then install the
following modules in PowerShell.
```powershell
Install-Module Pscx -AllowClobber
Install-Module VSSetup
```
Initialize Visual Studio in PowerShell with the following commands. We may setup
this for all PowerShell sessions by writing it to the PowerShell profile. Change
the library path according to your own case.
```powershell
Import-VisualStudioVars -Architecture x64
$env:LIB += ";C:\Program Files\Python37\libs"
```
### Apple Silicon (M1/M2 Chips) ###
We need PyTorch >= 1.13 to run TorchDrug on Apple silicon. For `torch-scatter` and
`torch-cluster`, they can be compiled from their sources. Note TorchDrug doesn't
support `mps` devices.
```bash
pip install torch==1.13.0
pip install git+https://github.com/rusty1s/pytorch_scatter.git
pip install git+https://github.com/rusty1s/pytorch_cluster.git
pip install torchdrug
```
Quick Start
-----------
TorchDrug is designed for humans and focused on graph structured data.
It enables easy implementation of graph operations in machine learning models.
All the operations in TorchDrug are backed by [PyTorch] framework, and support GPU
acceleration and auto differentiation.
```python
from torchdrug import data
edge_list = [[0, 1], [1, 2], [2, 3], [3, 4], [4, 5], [5, 0]]
graph = data.Graph(edge_list, num_node=6)
graph = graph.cuda()
# the subgraph induced by nodes 2, 3 & 4
subgraph = graph.subgraph([2, 3, 4])
```
Molecules are also supported in TorchDrug. You can get the desired molecule
properties without any domain knowledge.
```python
mol = data.Molecule.from_smiles("CCOC(=O)N", atom_feature="default", bond_feature="default")
print(mol.node_feature)
print(mol.atom_type)
print(mol.to_scaffold())
```
You may also register custom node, edge or graph attributes. They will be
automatically processed during indexing operations.
```python
with mol.edge():
mol.is_CC_bond = (mol.edge_list[:, :2] == td.CARBON).all(dim=-1)
sub_mol = mol.subgraph(mol.atom_type != td.NITROGEN)
print(sub_mol.is_CC_bond)
```
TorchDrug provides a wide range of common datasets and building blocks for drug
discovery. With minimal code, you can apply standard models to solve your own
problem.
```python
import torch
from torchdrug import datasets
dataset = datasets.Tox21()
dataset[0].visualize()
lengths = [int(0.8 * len(dataset)), int(0.1 * len(dataset))]
lengths += [len(dataset) - sum(lengths)]
train_set, valid_set, test_set = torch.utils.data.random_split(dataset, lengths)
```
```python
from torchdrug import models, tasks
model = models.GIN(dataset.node_feature_dim, hidden_dims=[256, 256, 256, 256])
task = tasks.PropertyPrediction(model, task=dataset.tasks)
```
Training and inference are accelerated by multiple CPUs or GPUs.
This can be seamlessly switched in TorchDrug by just a line of code.
```python
from torchdrug import core
# Single CPU / Multiple CPUs / Distributed CPUs
solver = core.Engine(task, train_set, valid_set, test_set, optimizer)
# Single GPU
solver = core.Engine(task, train_set, valid_set, test_set, optimizer, gpus=[0])
# Multiple GPUs
solver = core.Engine(task, train_set, valid_set, test_set, optimizer, gpus=[0, 1, 2, 3])
# Distributed GPUs
solver = core.Engine(task, train_set, valid_set, test_set, optimizer, gpus=[0, 1, 2, 3, 0, 1, 2, 3])
```
Experiments can be easily tracked and managed through [Weights & Biases platform].
```python
solver = core.Engine(task, train_set, valid_set, test_set, optimizer, logger="wandb")
```
[Weights & Biases platform]: https://wandb.ai/
Contributing
------------
Everyone is welcome to contribute to the development of TorchDrug.
Please refer to [contributing guidelines](CONTRIBUTING.md) for more details.
License
-------
TorchDrug is released under [Apache-2.0 License](LICENSE).
Raw data
{
"_id": null,
"home_page": "https://torchdrug.ai/",
"name": "torchdrug",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.7,<3.11",
"maintainer_email": "",
"keywords": "deep-learning,pytorch,drug-discovery",
"author": "TorchDrug Team",
"author_email": "",
"download_url": "",
"platform": null,
"description": "[](https://torchdrug.ai/)\n<h1 align=\"center\">\n with\n <a href=\"https://torchprotein.ai/\">\n <img src=\"asset/torchprotein_logo_tight.svg\" alt=\"TorchProtein\" style=\"height:26px\" />\n </a>\n</h1>\n\n[](https://colab.research.google.com/drive/1Tbnr1Fog_YjkqU1MOhcVLuxqZ4DC-c8-#forceEdit=true&sandboxMode=true)\n[](https://github.com/DeepGraphLearning/torchdrug/blob/master/CONTRIBUTING.md)\n[](https://github.com/DeepGraphLearning/torchdrug/blob/master/LICENSE)\n[](https://pypi.org/project/torchdrug/)\n[](https://twitter.com/DrugTorch)\n\n[Docs] | [Tutorials] | [Benchmarks] | [Papers Implemented]\n\n[Docs]: https://deepgraphlearning.github.io/torchdrug-site/docs\n[Tutorials]: https://deepgraphlearning.github.io/torchdrug-site/docs/tutorials\n[Benchmarks]: https://deepgraphlearning.github.io/torchdrug-site/docs/benchmark\n[Papers Implemented]: https://deepgraphlearning.github.io/torchdrug-site/docs/paper\n\nTorchDrug is a [PyTorch]-based machine learning toolbox designed for several purposes.\n\n- Easy implementation of graph operations in a PyTorchic style with GPU support\n- Being friendly to practitioners with minimal knowledge about drug discovery\n- Rapid prototyping of machine learning research\n\n[PyTorch]: https://pytorch.org/\n\nInstallation\n------------\n\nTorchDrug can be installed on either Linux, Windows or macOS. It is compatible with\n3.7 <= Python <= 3.10 and PyTorch >= 1.8.0.\n\n### From Conda ###\n\n```bash\nconda install torchdrug -c milagraph -c conda-forge -c pytorch -c pyg\n```\n\n### From Pip ###\n\n```bash\npip install torch==1.9.0\npip install torch-scatter torch-cluster -f https://pytorch-geometric.com/whl/torch-1.9.0+cu102.html\npip install torchdrug\n```\n\nTo install `torch-scatter` for other PyTorch or CUDA versions, please see the\ninstructions in https://github.com/rusty1s/pytorch_scatter\n\n### From Source ###\n\n```bash\ngit clone https://github.com/DeepGraphLearning/torchdrug\ncd torchdrug\npip install -r requirements.txt\npython setup.py install\n```\n\n### Windows (PowerShell) ###\n\nWe need to first install the build tools for Visual Studio. We then install the\nfollowing modules in PowerShell.\n\n```powershell\nInstall-Module Pscx -AllowClobber\nInstall-Module VSSetup\n```\n\nInitialize Visual Studio in PowerShell with the following commands. We may setup\nthis for all PowerShell sessions by writing it to the PowerShell profile. Change\nthe library path according to your own case.\n\n```powershell\nImport-VisualStudioVars -Architecture x64\n$env:LIB += \";C:\\Program Files\\Python37\\libs\"\n```\n\n### Apple Silicon (M1/M2 Chips) ###\n\nWe need PyTorch >= 1.13 to run TorchDrug on Apple silicon. For `torch-scatter` and\n`torch-cluster`, they can be compiled from their sources. Note TorchDrug doesn't\nsupport `mps` devices.\n\n```bash\npip install torch==1.13.0\npip install git+https://github.com/rusty1s/pytorch_scatter.git\npip install git+https://github.com/rusty1s/pytorch_cluster.git\npip install torchdrug\n```\n\nQuick Start\n-----------\n\nTorchDrug is designed for humans and focused on graph structured data.\nIt enables easy implementation of graph operations in machine learning models.\nAll the operations in TorchDrug are backed by [PyTorch] framework, and support GPU\nacceleration and auto differentiation.\n\n```python\nfrom torchdrug import data\n\nedge_list = [[0, 1], [1, 2], [2, 3], [3, 4], [4, 5], [5, 0]]\ngraph = data.Graph(edge_list, num_node=6)\ngraph = graph.cuda()\n# the subgraph induced by nodes 2, 3 & 4\nsubgraph = graph.subgraph([2, 3, 4])\n```\n\nMolecules are also supported in TorchDrug. You can get the desired molecule\nproperties without any domain knowledge.\n\n```python\nmol = data.Molecule.from_smiles(\"CCOC(=O)N\", atom_feature=\"default\", bond_feature=\"default\")\nprint(mol.node_feature)\nprint(mol.atom_type)\nprint(mol.to_scaffold())\n```\n\nYou may also register custom node, edge or graph attributes. They will be\nautomatically processed during indexing operations.\n\n```python\nwith mol.edge():\n\tmol.is_CC_bond = (mol.edge_list[:, :2] == td.CARBON).all(dim=-1)\nsub_mol = mol.subgraph(mol.atom_type != td.NITROGEN)\nprint(sub_mol.is_CC_bond)\n```\n\nTorchDrug provides a wide range of common datasets and building blocks for drug\ndiscovery. With minimal code, you can apply standard models to solve your own\nproblem.\n\n```python\nimport torch\nfrom torchdrug import datasets\n\ndataset = datasets.Tox21()\ndataset[0].visualize()\nlengths = [int(0.8 * len(dataset)), int(0.1 * len(dataset))]\nlengths += [len(dataset) - sum(lengths)]\ntrain_set, valid_set, test_set = torch.utils.data.random_split(dataset, lengths)\n```\n\n```python\nfrom torchdrug import models, tasks\n\nmodel = models.GIN(dataset.node_feature_dim, hidden_dims=[256, 256, 256, 256])\ntask = tasks.PropertyPrediction(model, task=dataset.tasks)\n```\n\nTraining and inference are accelerated by multiple CPUs or GPUs.\nThis can be seamlessly switched in TorchDrug by just a line of code.\n```python\nfrom torchdrug import core\n\n# Single CPU / Multiple CPUs / Distributed CPUs\nsolver = core.Engine(task, train_set, valid_set, test_set, optimizer)\n# Single GPU\nsolver = core.Engine(task, train_set, valid_set, test_set, optimizer, gpus=[0])\n# Multiple GPUs\nsolver = core.Engine(task, train_set, valid_set, test_set, optimizer, gpus=[0, 1, 2, 3])\n# Distributed GPUs\nsolver = core.Engine(task, train_set, valid_set, test_set, optimizer, gpus=[0, 1, 2, 3, 0, 1, 2, 3])\n```\n\nExperiments can be easily tracked and managed through [Weights & Biases platform].\n```python\nsolver = core.Engine(task, train_set, valid_set, test_set, optimizer, logger=\"wandb\")\n```\n\n[Weights & Biases platform]: https://wandb.ai/\n\nContributing\n------------\n\nEveryone is welcome to contribute to the development of TorchDrug.\nPlease refer to [contributing guidelines](CONTRIBUTING.md) for more details.\n\nLicense\n-------\n\nTorchDrug is released under [Apache-2.0 License](LICENSE).\n",
"bugtrack_url": null,
"license": "Apache-2.0",
"summary": "A powerful and flexible machine learning platform for drug discovery",
"version": "0.2.1",
"project_urls": {
"Homepage": "https://torchdrug.ai/"
},
"split_keywords": [
"deep-learning",
"pytorch",
"drug-discovery"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "ea295e1a14d7765106f38b4962c9d2c05210fe7b3ba14985d9c4e9b9b6a59ef0",
"md5": "5a2f6aeafbc6a249715d56d36dede828",
"sha256": "f416dbbfb3fc697770e9ee7fa7769611fad060a4749ee49499d87583cfb58ee2"
},
"downloads": -1,
"filename": "torchdrug-0.2.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "5a2f6aeafbc6a249715d56d36dede828",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.7,<3.11",
"size": 268505,
"upload_time": "2023-07-16T22:07:36",
"upload_time_iso_8601": "2023-07-16T22:07:36.768874Z",
"url": "https://files.pythonhosted.org/packages/ea/29/5e1a14d7765106f38b4962c9d2c05210fe7b3ba14985d9c4e9b9b6a59ef0/torchdrug-0.2.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-07-16 22:07:36",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "torchdrug"
}
JSON