gistfile1.txt

import numpy as np
import deepchem as dc

datasets = ['muv']
models = ['weave']
metrics = [dc.metrics.Metric(dc.metrics.roc_auc_score, np.mean),dc.metrics.Metric(dc.metrics.prc_auc_score, np.mean)]

for model in models:
    print("RUNNING:",model)
    dc.molnet.run_benchmark(datasets,model,test=True,metric=metrics)


datasets = ['qm9']
models = ['graphconvreg']

metrics = [
    dc.metrics.Metric(dc.metrics.rms_score, np.mean),
    dc.metrics.Metric(dc.metrics.mean_squared_error, np.mean),
    dc.metrics.Metric(dc.metrics.mean_absolute_error, np.mean),
    dc.metrics.Metric(dc.metrics.pearson_r2_score, np.mean),
    dc.metrics.Metric(dc.metrics.r2_score, np.mean)
]

for model in models:
    print("RUNNING:",model)
    dc.molnet.run_benchmark(datasets,model,test=True,metric=metrics)

MUV

gcnn

	MoleculeNet MTL, run_benchmark	Regenerated MTL, run_benchmark	Regenerated MTL, examples
MUV-466	--	0.005	0.008
MUV-548	--	0.062	0.002
MUV-600	--	0.010	0.005
MUV-644	--	0.011	0.009
MUV-652	--	0.001	0.074
MUV-689	--	0.014	0.002
MUV-692	--	0.001	0.006
MUV-712	--	0.001	0.065
MUV-713	--	0.064	0.008
MUV-733	--	0.012	0.004
MUV-737	--	0.005	0.027
MUV-810	--	0.008	0.006
MUV-832	--	0.003	0.027
MUV-846	--	0.167	0.024
MUV-852	--	0.056	0.060
MUV-858	--	0.009	0.006
MUV-859	--	0.004	0.002
average	0.046	0.026	0.020

weave

	MoleculeNet MTL, run_benchmark	Regenerated MTL, run_benchmark	Regenerated MTL, examples
MUV-466	--	0.501	0.000
MUV-548	--	0.500	0.000
MUV-600	--	0.501	0.003
MUV-644	--	0.501	0.003
MUV-652	--	0.501	0.001
MUV-689	--	0.502	0.001
MUV-692	--	0.501	0.006
MUV-712	--	0.500	0.005
MUV-713	--	0.501	0.002
MUV-733	--	0.502	0.004
MUV-737	--	0.501	0.002
MUV-810	--	0.501	0.001
MUV-832	--	0.501	0.003
MUV-846	--	0.500	0.003
MUV-852	--	0.501	0.004
MUV-858	--	0.502	0.002
MUV-859	--	0.501	0.002
average	0.109	0.501	0.003