Loading README.md +11 −7 Original line number Diff line number Diff line Loading @@ -347,9 +347,12 @@ Scaffold splitting | |Random forest |Scaffold |0.956 |0.201 | | |NN regression |Scaffold |0.897 |0.208 | | |Graphconv regression|Scaffold |0.783 |0.068 | |chembl |MT-NN regression |Index |0.443 |0.427 | | |MT-NN regression |Random |0.464 |0.434 | | |MT-NN regression |Scaffold |0.484 |0.361 | |chembl |MT-NN regression |Index |0.828 |0.565 | | |Graphconv regression|Index |0.192 |0.293 | | |MT-NN regression |Random |0.829 |0.562 | | |Graphconv regression|Random |0.198 |0.271 | | |MT-NN regression |Scaffold |0.843 |0.430 | | |Graphconv regression|Scaffold |0.231 |0.294 | |clearance |Random forest |Index |0.953 |0.244 | | |NN regression |Index |0.884 |0.211 | | |Graphconv regression|Index |0.696 |0.230 | Loading Loading @@ -388,11 +391,11 @@ Scaffold splitting | |NN regression |Scaffold |0.831 |0.302 | | |Graphconv regression|Scaffold |0.882 |0.593 | |nci |MT-NN regression |Index |0.690 |0.062 | | |Graphconv regression|Index |0.123 |0.048 | | |MT-NN regression |Random |0.168 |0.085 | | |Graphconv regression|Index |0.123 |0.053 | | |MT-NN regression |Random |0.698 |0.117 | | |Graphconv regression|Random |0.117 |0.076 | | |MT-NN regression |Scaffold |0.180 |0.052 | | |Graphconv regression|Scaffold |0.131 |0.046 | | |MT-NN regression |Scaffold |0.692 |0.036 | | |Graphconv regression|Scaffold |0.131 |0.036 | |pdbbind(core) |Random forest |Random |0.969 |0.445 | | |NN regression |Random |0.973 |0.494 | |pdbbind(refined)|Random forest |Random |0.963 |0.511 | Loading Loading @@ -528,6 +531,7 @@ Time needed for benchmark test(~20h in total) | |Graphconv regression|10 |110 | | |Random forest |10 |50 | |chembl |MT-NN regression |200 |9000 | | |Graphconv regression|250 |1800 | |clearance |NN regression |10 |20 | | |Graphconv regression|10 |60 | | |Random forest |10 |10 | Loading deepchem/molnet/load_function/kaggle_datasets.py +9 −14 Original line number Diff line number Diff line Loading @@ -42,7 +42,6 @@ def get_transformers(train_dataset): # Set shard size low to avoid memory problems. def gen_kaggle(KAGGLE_tasks, raw_train_dir, train_dir, valid_dir, test_dir, Loading Loading @@ -94,9 +93,11 @@ def gen_kaggle(KAGGLE_tasks, remove_missing_entries(valid_dataset) remove_missing_entries(test_dataset) print("Shuffling order of train dataset.") train_dataset.sparse_shuffle() print("Transforming datasets with transformers.") transformers = get_transformers(train_dataset) raw_train_dataset = train_dataset for transformer in transformers: print("Performing transformations with %s" % transformer.__class__.__name__) Loading @@ -105,11 +106,7 @@ def gen_kaggle(KAGGLE_tasks, valid_dataset = transformer.transform(valid_dataset) test_dataset = transformer.transform(test_dataset) print("Shuffling order of train dataset.") train_dataset.sparse_shuffle() print("Moving directories") raw_train_dataset.move(raw_train_dir) train_dataset.move(train_dir) valid_dataset.move(valid_dir) test_dataset.move(test_dir) Loading @@ -119,7 +116,7 @@ def gen_kaggle(KAGGLE_tasks, print("TIMING: KAGGLE fitting took %0.3f s" % (time2 - time1)) ############################################################## TIMING return (raw_train_dataset, train_dataset, valid_dataset, test_dataset) return train_dataset, valid_dataset, test_dataset def load_kaggle(shard_size=2000, featurizer=None, split=None): Loading @@ -134,24 +131,22 @@ def load_kaggle(shard_size=2000, featurizer=None, split=None): data_dir = "/tmp" data_dir = os.path.join(data_dir, "kaggle") raw_train_dir = os.path.join(data_dir, "raw_train_dir") train_dir = os.path.join(data_dir, "train_dir") valid_dir = os.path.join(data_dir, "valid_dir") test_dir = os.path.join(data_dir, "test_dir") if (os.path.exists(raw_train_dir) and os.path.exists(train_dir) and os.path.exists(valid_dir) and os.path.exists(test_dir)): if (os.path.exists(train_dir) and os.path.exists(valid_dir) and os.path.exists(test_dir)): print("Reloading existing datasets") raw_train_dataset = deepchem.data.DiskDataset(raw_train_dir) train_dataset = deepchem.data.DiskDataset(train_dir) valid_dataset = deepchem.data.DiskDataset(valid_dir) test_dataset = deepchem.data.DiskDataset(test_dir) else: print("Featurizing datasets") (raw_train_dataset, train_dataset, valid_dataset, test_dataset) = \ gen_kaggle(KAGGLE_tasks, raw_train_dir, train_dir, valid_dir, test_dir, data_dir, train_dataset, valid_dataset, test_dataset = \ gen_kaggle(KAGGLE_tasks, train_dir, valid_dir, test_dir, data_dir, shard_size=shard_size) transformers = get_transformers(raw_train_dataset) transformers = get_transformers(train_dataset) return KAGGLE_tasks, (train_dataset, valid_dataset, test_dataset), transformers Loading
README.md +11 −7 Original line number Diff line number Diff line Loading @@ -347,9 +347,12 @@ Scaffold splitting | |Random forest |Scaffold |0.956 |0.201 | | |NN regression |Scaffold |0.897 |0.208 | | |Graphconv regression|Scaffold |0.783 |0.068 | |chembl |MT-NN regression |Index |0.443 |0.427 | | |MT-NN regression |Random |0.464 |0.434 | | |MT-NN regression |Scaffold |0.484 |0.361 | |chembl |MT-NN regression |Index |0.828 |0.565 | | |Graphconv regression|Index |0.192 |0.293 | | |MT-NN regression |Random |0.829 |0.562 | | |Graphconv regression|Random |0.198 |0.271 | | |MT-NN regression |Scaffold |0.843 |0.430 | | |Graphconv regression|Scaffold |0.231 |0.294 | |clearance |Random forest |Index |0.953 |0.244 | | |NN regression |Index |0.884 |0.211 | | |Graphconv regression|Index |0.696 |0.230 | Loading Loading @@ -388,11 +391,11 @@ Scaffold splitting | |NN regression |Scaffold |0.831 |0.302 | | |Graphconv regression|Scaffold |0.882 |0.593 | |nci |MT-NN regression |Index |0.690 |0.062 | | |Graphconv regression|Index |0.123 |0.048 | | |MT-NN regression |Random |0.168 |0.085 | | |Graphconv regression|Index |0.123 |0.053 | | |MT-NN regression |Random |0.698 |0.117 | | |Graphconv regression|Random |0.117 |0.076 | | |MT-NN regression |Scaffold |0.180 |0.052 | | |Graphconv regression|Scaffold |0.131 |0.046 | | |MT-NN regression |Scaffold |0.692 |0.036 | | |Graphconv regression|Scaffold |0.131 |0.036 | |pdbbind(core) |Random forest |Random |0.969 |0.445 | | |NN regression |Random |0.973 |0.494 | |pdbbind(refined)|Random forest |Random |0.963 |0.511 | Loading Loading @@ -528,6 +531,7 @@ Time needed for benchmark test(~20h in total) | |Graphconv regression|10 |110 | | |Random forest |10 |50 | |chembl |MT-NN regression |200 |9000 | | |Graphconv regression|250 |1800 | |clearance |NN regression |10 |20 | | |Graphconv regression|10 |60 | | |Random forest |10 |10 | Loading
deepchem/molnet/load_function/kaggle_datasets.py +9 −14 Original line number Diff line number Diff line Loading @@ -42,7 +42,6 @@ def get_transformers(train_dataset): # Set shard size low to avoid memory problems. def gen_kaggle(KAGGLE_tasks, raw_train_dir, train_dir, valid_dir, test_dir, Loading Loading @@ -94,9 +93,11 @@ def gen_kaggle(KAGGLE_tasks, remove_missing_entries(valid_dataset) remove_missing_entries(test_dataset) print("Shuffling order of train dataset.") train_dataset.sparse_shuffle() print("Transforming datasets with transformers.") transformers = get_transformers(train_dataset) raw_train_dataset = train_dataset for transformer in transformers: print("Performing transformations with %s" % transformer.__class__.__name__) Loading @@ -105,11 +106,7 @@ def gen_kaggle(KAGGLE_tasks, valid_dataset = transformer.transform(valid_dataset) test_dataset = transformer.transform(test_dataset) print("Shuffling order of train dataset.") train_dataset.sparse_shuffle() print("Moving directories") raw_train_dataset.move(raw_train_dir) train_dataset.move(train_dir) valid_dataset.move(valid_dir) test_dataset.move(test_dir) Loading @@ -119,7 +116,7 @@ def gen_kaggle(KAGGLE_tasks, print("TIMING: KAGGLE fitting took %0.3f s" % (time2 - time1)) ############################################################## TIMING return (raw_train_dataset, train_dataset, valid_dataset, test_dataset) return train_dataset, valid_dataset, test_dataset def load_kaggle(shard_size=2000, featurizer=None, split=None): Loading @@ -134,24 +131,22 @@ def load_kaggle(shard_size=2000, featurizer=None, split=None): data_dir = "/tmp" data_dir = os.path.join(data_dir, "kaggle") raw_train_dir = os.path.join(data_dir, "raw_train_dir") train_dir = os.path.join(data_dir, "train_dir") valid_dir = os.path.join(data_dir, "valid_dir") test_dir = os.path.join(data_dir, "test_dir") if (os.path.exists(raw_train_dir) and os.path.exists(train_dir) and os.path.exists(valid_dir) and os.path.exists(test_dir)): if (os.path.exists(train_dir) and os.path.exists(valid_dir) and os.path.exists(test_dir)): print("Reloading existing datasets") raw_train_dataset = deepchem.data.DiskDataset(raw_train_dir) train_dataset = deepchem.data.DiskDataset(train_dir) valid_dataset = deepchem.data.DiskDataset(valid_dir) test_dataset = deepchem.data.DiskDataset(test_dir) else: print("Featurizing datasets") (raw_train_dataset, train_dataset, valid_dataset, test_dataset) = \ gen_kaggle(KAGGLE_tasks, raw_train_dir, train_dir, valid_dir, test_dir, data_dir, train_dataset, valid_dataset, test_dataset = \ gen_kaggle(KAGGLE_tasks, train_dir, valid_dir, test_dir, data_dir, shard_size=shard_size) transformers = get_transformers(raw_train_dataset) transformers = get_transformers(train_dataset) return KAGGLE_tasks, (train_dataset, valid_dataset, test_dataset), transformers
