Loading README.md +17 −6 Original line number Diff line number Diff line Loading @@ -205,11 +205,22 @@ different subclasses of ``Featurizer`` for convenience: ### Performances |Dataset |Model |Train score/ROC-AUC|Valid score/ROC-AUC|Time(loading)/s |Time(running)/s| |-----------|--------------------|-------------------|-------------------|----------------|---------------| |tox21 |tensorflow(MT-DNN) |0.987 |0.800 |35 |36 | |muv |tensorflow(MT-DNN) |0.979 |0.660 |414 |255 | |pcba |tensorflow(MT-DNN) |0.949 |0.791 |1765 |7209 | |sider |tensorflow(MT-DNN) |0.864 |0.627 |10 |63 | |toxcast |tensorflow(MT-DNN) |0.944 |0.697 |75 |2374 | |tox21 |logistic regression |0.910 |0.759 |30 |30 | | |tensorflow(MT-NN) |0.987 |0.800 | | | | |graph convolution |0.930 |0.819 | | | |muv |logistic regression |0.910 |0.744 |400 |800 | | |tensorflow(MT-NN) |0.980 |0.710 | | | | |graph convolution |0.881 |0.832 | | | |pcba |logistic regression |0.741 |0.719 |1800 |7200 | | |tensorflow(MT-NN) |0.949 |0.791 | | | | |graph convolution |0.866 |0.836 | | | |sider |logistic regression |0.900 |0.620 |20 |60 | | |tensorflow(MT-NN) |0.931 |0.647 | | | | |graph convolution |0.845 |0.646 | | | |toxcast |logistic regression |0.762 |0.622 |80 |2400 | | |tensorflow(MT-NN) |0.926 |0.705 | | | | |graph convolution |0.906 |0.725 | | | ## Contributing to DeepChem Loading examples/benchmark.py +1 −1 Original line number Diff line number Diff line Loading @@ -253,7 +253,7 @@ def benchmark_train_and_valid(base_dir,train_dataset,valid_dataset,tasks, sess = tf.Session(graph=g) K.set_session(sess) with g.as_default(): graph_model = dc.models.SequentialGraphModel(n_features) graph_model = dc.nn.SequentialGraph(n_features) graph_model.add(dc.nn.GraphConv(int(n_filters), activation='relu')) graph_model.add(dc.nn.BatchNormalization(epsilon=1e-5, mode=1)) graph_model.add(dc.nn.GraphPool()) Loading examples/best_results.csv +10 −1 Original line number Diff line number Diff line tox21,logreg,train,0.910,valid,0.759,learning_rate,0.004~0.008,penalty(l1 or l2),0.3~0.6 sider,logreg,train,0.900,valid,0.620,learning_rate,0.004~0.008,penalty(l2),0.4~1 muv,logreg,train,0.910,valid,0.744,learning_rate,0.002~0.004,penalty(l2),0.4~0.7 toxcast,logreg,train,0.762,valid,0.622,learning_rate,0.004~0.1,penalty(l2),0.2~0.6 sider,tf,train,0.931,valid,0.647,learning_rate,0.0003~0.003,dropouts,0.2~0.3,layer_sizes,1000~1200 tox21,tf,train,0.996,valid,0.763,learning_rate,0.001~0.003,dropouts,0.4~0.5,layer_sizes,1000~1200 toxcast,tf,train,0.944,valid,0.699,learning_rate,0.0004,dropouts,0.4,layer_sizes,500 toxcast,tf,train,0.926,valid,0.705,learning_rate,0.0008~0.0012,dropouts,0.3~0.5,layer_sizes,1200 muv,tf,train,0.980,valid,0.710,learning_rate,0.0008~0.0013,dropouts,0.25~0.4,layer_sizes,1200 pcba,tf,train,0.970,valid,0.779,learning_rate,0.0008,dropouts,0.35,layer_sizes,1200 pcba,graphconv,train,0.866,valid,0.836,learning_rate,0.0001,learning_rate_decay,1500,layer_structure,2layers&64filters muv,graphconv,train,0.881,valid,0.832,learning_rate,0.0004~0.0008,learning_rate_decay,1500,layer_structure,2layers&64~128filters tox21,graphconv,train,0.930,valid,0.819,learning_rate,0.0003~0.002,learning_rate_decay,1000~2000,layer_structure,2layers&128filters sider,graphconv,train,0.845,valid,0.646,learning_rate,0.0004~0.002,,,layer_structure,2layers&64filters toxcast,graphconv,train,0.906,valid,0.725,learning_rate,0.0004,,,layer_structure,2layers&96filters Loading
README.md +17 −6 Original line number Diff line number Diff line Loading @@ -205,11 +205,22 @@ different subclasses of ``Featurizer`` for convenience: ### Performances |Dataset |Model |Train score/ROC-AUC|Valid score/ROC-AUC|Time(loading)/s |Time(running)/s| |-----------|--------------------|-------------------|-------------------|----------------|---------------| |tox21 |tensorflow(MT-DNN) |0.987 |0.800 |35 |36 | |muv |tensorflow(MT-DNN) |0.979 |0.660 |414 |255 | |pcba |tensorflow(MT-DNN) |0.949 |0.791 |1765 |7209 | |sider |tensorflow(MT-DNN) |0.864 |0.627 |10 |63 | |toxcast |tensorflow(MT-DNN) |0.944 |0.697 |75 |2374 | |tox21 |logistic regression |0.910 |0.759 |30 |30 | | |tensorflow(MT-NN) |0.987 |0.800 | | | | |graph convolution |0.930 |0.819 | | | |muv |logistic regression |0.910 |0.744 |400 |800 | | |tensorflow(MT-NN) |0.980 |0.710 | | | | |graph convolution |0.881 |0.832 | | | |pcba |logistic regression |0.741 |0.719 |1800 |7200 | | |tensorflow(MT-NN) |0.949 |0.791 | | | | |graph convolution |0.866 |0.836 | | | |sider |logistic regression |0.900 |0.620 |20 |60 | | |tensorflow(MT-NN) |0.931 |0.647 | | | | |graph convolution |0.845 |0.646 | | | |toxcast |logistic regression |0.762 |0.622 |80 |2400 | | |tensorflow(MT-NN) |0.926 |0.705 | | | | |graph convolution |0.906 |0.725 | | | ## Contributing to DeepChem Loading
examples/benchmark.py +1 −1 Original line number Diff line number Diff line Loading @@ -253,7 +253,7 @@ def benchmark_train_and_valid(base_dir,train_dataset,valid_dataset,tasks, sess = tf.Session(graph=g) K.set_session(sess) with g.as_default(): graph_model = dc.models.SequentialGraphModel(n_features) graph_model = dc.nn.SequentialGraph(n_features) graph_model.add(dc.nn.GraphConv(int(n_filters), activation='relu')) graph_model.add(dc.nn.BatchNormalization(epsilon=1e-5, mode=1)) graph_model.add(dc.nn.GraphPool()) Loading
examples/best_results.csv +10 −1 Original line number Diff line number Diff line tox21,logreg,train,0.910,valid,0.759,learning_rate,0.004~0.008,penalty(l1 or l2),0.3~0.6 sider,logreg,train,0.900,valid,0.620,learning_rate,0.004~0.008,penalty(l2),0.4~1 muv,logreg,train,0.910,valid,0.744,learning_rate,0.002~0.004,penalty(l2),0.4~0.7 toxcast,logreg,train,0.762,valid,0.622,learning_rate,0.004~0.1,penalty(l2),0.2~0.6 sider,tf,train,0.931,valid,0.647,learning_rate,0.0003~0.003,dropouts,0.2~0.3,layer_sizes,1000~1200 tox21,tf,train,0.996,valid,0.763,learning_rate,0.001~0.003,dropouts,0.4~0.5,layer_sizes,1000~1200 toxcast,tf,train,0.944,valid,0.699,learning_rate,0.0004,dropouts,0.4,layer_sizes,500 toxcast,tf,train,0.926,valid,0.705,learning_rate,0.0008~0.0012,dropouts,0.3~0.5,layer_sizes,1200 muv,tf,train,0.980,valid,0.710,learning_rate,0.0008~0.0013,dropouts,0.25~0.4,layer_sizes,1200 pcba,tf,train,0.970,valid,0.779,learning_rate,0.0008,dropouts,0.35,layer_sizes,1200 pcba,graphconv,train,0.866,valid,0.836,learning_rate,0.0001,learning_rate_decay,1500,layer_structure,2layers&64filters muv,graphconv,train,0.881,valid,0.832,learning_rate,0.0004~0.0008,learning_rate_decay,1500,layer_structure,2layers&64~128filters tox21,graphconv,train,0.930,valid,0.819,learning_rate,0.0003~0.002,learning_rate_decay,1000~2000,layer_structure,2layers&128filters sider,graphconv,train,0.845,valid,0.646,learning_rate,0.0004~0.002,,,layer_structure,2layers&64filters toxcast,graphconv,train,0.906,valid,0.725,learning_rate,0.0004,,,layer_structure,2layers&96filters
