Requirements¶

Hard requirements¶

DeepChem officially supports Python 3.6 through 3.7 and requires these packages on any condition.

joblib
NumPy
pandas
scikit-learn
SciPy
TensorFlow
- deepchem>=2.4.0 depends on TensorFlow v2 (2.3.x)
- deepchem<2.4.0 depends on TensorFlow v1 (>=1.14)

DeepChem has a number of “soft” requirements.

Package name	Version	Location where this package is used (dc: deepchem)
BioPython	latest	`dc.utlis.genomics_utils`
Deep Graph Library	0.5.x	`dc.feat.graph_data`, `dc.models.torch_models`
DGL-LifeSci	0.2.x	`dc.models.torch_models`
HuggingFace Transformers	Not Testing	`dc.feat.smiles_tokenizer`
HuggingFace Tokenizers	latest	`dc.feat.HuggingFaceVocabularyBuilder`
LightGBM	latest	`dc.models.gbdt_models`
matminer	latest	`dc.feat.materials_featurizers`
MDTraj	latest	`dc.utils.pdbqt_utils`
Mol2vec	latest	`dc.utils.molecule_featurizers`
Mordred	latest	`dc.utils.molecule_featurizers`
NetworkX	latest	`dc.utils.rdkit_utils`
OpenAI Gym	Not Testing	`dc.rl`
OpenMM	latest	`dc.utils.rdkit_utils`
PDBFixer	latest	`dc.utils.rdkit_utils`
Pillow	latest	`dc.data.data_loader`, `dc.trans.transformers`
PubChemPy	latest	`dc.feat.molecule_featurizers`
pyGPGO	latest	`dc.hyper.gaussian_process`
Pymatgen	latest	`dc.feat.materials_featurizers`
PyTorch	1.6.0	`dc.data.datasets`
PyTorch Geometric	1.6.x (with PyTorch 1.6.0)	`dc.feat.graph_data` `dc.models.torch_models`
RDKit	latest	Many modules (we recommend you to install)
simdna	latest	`dc.metrics.genomic_metrics`, `dc.molnet.dnasim`
Tensorflow Probability	0.11.x	`dc.rl`
Weights & Biases	Not Testing	`dc.models.keras_model`, `dc.models.callbacks`
XGBoost	latest	`dc.models.gbdt_models`
Tensorflow Addons	latest	`dc.models.optimizers`