temp save (56d66761) · Commits · 钟慕尧 / deepchem

deepchem/hyper/gaussian_process.py

+4 −2

Original line number	Diff line number	Diff line
		@@ -31,7 +31,7 @@ class GaussianProcessHyperparamOpt(HyperparamOpt):
		search_range=4,
		hp_invalid_list=[
		'seed', 'nb_epoch', 'penalty_type', 'dropouts',
		'bypass_dropouts', 'n_pair_feat'
		'bypass_dropouts', 'n_pair_feat', 'fit_transformers'
		],
		logdir=None):
		"""Perform hyperparams search using a gaussian process assumption
		@@ -97,7 +97,9 @@ class GaussianProcessHyperparamOpt(HyperparamOpt):
		if hp_list_class[i] is list]

		# Number of parameters
		n_param = len(hp_list_single + sum([hp[1] for hp in hp_list_multiple]))
		n_param = len(hp_list_single)
		if len(hp_list_multiple) > 0:
		n_param = n_param + sum([hp[1] for hp in hp_list_multiple])
		# Range of optimization
		param_range = []
		for hp in hp_list_single:

deepchem/molnet/run_benchmark.py

+4 −2

Original line number	Diff line number	Diff line
		@@ -14,6 +14,7 @@ import csv
		import numpy as np
		import tensorflow as tf
		import deepchem
		import pickle
		from deepchem.molnet.run_benchmark_models import benchmark_classification, benchmark_regression
		from deepchem.molnet.check_availability import CheckFeaturizer, CheckSplit
		from deepchem.molnet.preset_hyper_parameters import hps
		@@ -29,7 +30,7 @@ def run_benchmark(datasets,
		hyper_parameters=None,
		hyper_param_search=False,
		max_iter=20,
		search_range=4,
		search_range=2,
		test=False,
		reload=True,
		seed=123):
		@@ -221,7 +222,8 @@ def run_benchmark(datasets,
		output_line.extend(
		['time_for_running', time_finish_fitting - time_start_fitting])
		writer.writerow(output_line)

		with open(os.path.join(out_path, dataset + model + '.pkl'), 'w') as f:
		pickle.dump(hyper_parameters, f)

		#
		# Note by @XericZephyr. Reason why I spun off this function:

examples/benchmark.py

+26 −7

Original line number	Diff line number	Diff line
		@@ -88,23 +88,42 @@ else:
		seed = 123

		if len(splitters) == 0:
		splitters = ['index', 'random', 'scaffold']
		splitters = ['random']
		if len(models) == 0:
		models = [
		'tf', 'tf_robust', 'logreg', 'graphconv', 'tf_regression',
		'tf_regression_ft', 'graphconvreg'
		'tf', 'tf_robust', 'logreg', 'graphconv', 'irv', 'tf_regression',
		'tf_regression_ft', 'graphconvreg', 'weave', 'weave_regression', 'dtnn'
		]
		#irv, rf, rf_regression should be assigned manually
		if len(datasets) == 0:
		datasets = [
		'bace_c', 'bace_r', 'bbbp', 'clearance', 'clintox', 'delaney', 'hiv',
		'hopv', 'lipo', 'muv', 'pdbbind', 'ppb', 'qm7b', 'qm8', 'qm9', 'sampl',
		'sider', 'tox21', 'toxcast'
		'clintox', 'delaney', 'lipo', 'qm7b', 'qm8', 'sampl',
		'sider', 'tox21', 'toxcast', 'muv'
		]

		metrics = {
		'qm7': [dc.metrics.Metric(dc.metrics.mean_absolute_error, np.mean, mode='regression')],
		'qm7b': [dc.metrics.Metric(dc.metrics.mean_absolute_error, np.mean, mode='regression')],
		'qm8': [dc.metrics.Metric(dc.metrics.mean_absolute_error, np.mean, mode='regression')],
		'qm9': [dc.metrics.Metric(dc.metrics.mean_absolute_error, np.mean, mode='regression')],
		'delaney': [dc.metrics.Metric(dc.metrics.rms_score, np.mean, mode='regression')],
		'sampl': [dc.metrics.Metric(dc.metrics.rms_score, np.mean, mode='regression')],
		'lipo': [dc.metrics.Metric(dc.metrics.rms_score, np.mean, mode='regression')],
		'pdbbind': [dc.metrics.Metric(dc.metrics.rms_score, np.mean, mode='regression')],
		'pcba': [dc.metrics.Metric(dc.metrics.prc_auc_score, np.mean, mode='classification')],
		'muv': [dc.metrics.Metric(dc.metrics.prc_auc_score, np.mean, mode='classification')],
		'hiv': [dc.metrics.Metric(dc.metrics.roc_auc_score, np.mean, mode='classification')],
		'tox21': [dc.metrics.Metric(dc.metrics.roc_auc_score, np.mean, mode='classification')],
		'toxcast': [dc.metrics.Metric(dc.metrics.roc_auc_score, np.mean, mode='classification')],
		'sider': [dc.metrics.Metric(dc.metrics.roc_auc_score, np.mean, mode='classification')],
		'clintox': [dc.metrics.Metric(dc.metrics.roc_auc_score, np.mean, mode='classification')],
		'bace_c': [dc.metrics.Metric(dc.metrics.roc_auc_score, np.mean, mode='classification')],
		'bbbp': [dc.metrics.Metric(dc.metrics.roc_auc_score, np.mean, mode='classification')]
		}
		for dataset in datasets:
		for split in splitters:
		for model in models:
		np.random.seed(seed)
		dc.molnet.run_benchmark(
		[dataset], str(model), split=split, test=test, seed=seed)
		[dataset], str(model), split=split, metric=metrics[dataset],
		hyper_param_search=True, test=False, seed=seed)

Admin message