def test_01(self):
if os.path.exists('./experiments_data') is False:
os.mkdir('./experiments_data')
if os.path.exists('./datasets/image') is False:
z = zipfile.ZipFile("./datasets/image.zip", "r")
seq_x = []
label_y = []
for filename in z.namelist():
z.extract(filename, './datasets')
cur_dataset = expdata_generator(self.expdata_id)
cur_dataset.get_exp_data(sel_task='diagnose',
data_root='./datasets/image')
def test_02_rf(self):
cur_dataset = expdata_generator(self.expdata_id)
cur_dataset.load_exp_data()
expmodel_id = 'test.randomforest'
clf = RandomForest(expmodel_id=expmodel_id)
clf.fit(cur_dataset.train, cur_dataset.valid)
clf.load_model()
clf.inference(cur_dataset.test)
pred_results = clf.get_results()
assert np.shape(pred_results['hat_y']) == np.shape(pred_results['y'])
assert True not in np.isnan(pred_results['hat_y']).tolist()
assert True not in np.isnan(pred_results['hat_y'] * 0).tolist()
def test_02_basiccnn_gpu(self):
cur_dataset = expdata_generator(self.expdata_id)
cur_dataset.load_exp_data()
expmodel_id = 'test.basicnn.gpu'
clf = BasicCNN(expmodel_id=expmodel_id,
n_batchsize=20,
use_gpu=True,
n_epoch=6)
clf.fit(cur_dataset.train, cur_dataset.valid)
clf.load_model()
clf.inference(cur_dataset.test)
pred_results = clf.get_results()
assert np.shape(pred_results['hat_y']) == np.shape(pred_results['y'])
assert True not in np.isnan(pred_results['hat_y']).tolist()
assert True not in np.isnan(pred_results['hat_y'] * 0).tolist()
def test_02_dblstm_ws_cpu(self):
cur_dataset = expdata_generator(self.expdata_id)
cur_dataset.load_exp_data()
expmodel_id = 'test.dblstm_ws.cpu'
clf = DBLSTM_WS(expmodel_id=expmodel_id,
n_batchsize=20,
use_gpu=False,
n_epoch=3)
clf.fit(cur_dataset.train, cur_dataset.valid)
clf.load_model()
clf.inference(cur_dataset.test)
pred_results = clf.get_results()
assert np.shape(pred_results['hat_y']) == np.shape(pred_results['y'])
assert True not in np.isnan(pred_results['hat_y']).tolist()
assert True not in np.isnan(pred_results['hat_y'] * 0).tolist()
def test_02_typicalcnn_gpu(self):
cur_dataset = expdata_generator(self.expdata_id)
cur_dataset.load_exp_data()
expmodel_id = 'test.typicalcnn'
clf = TypicalCNN(expmodel_id=expmodel_id,
cnn_name='resnet18',
n_batchsize=20,
use_gpu=False,
n_epoch=6)
clf.fit(cur_dataset.train, cur_dataset.valid)
clf.load_model()
clf.inference(cur_dataset.test)
pred_results = clf.get_results()
assert np.shape(pred_results['hat_y']) == np.shape(pred_results['y'])
assert True not in np.isnan(pred_results['hat_y']).tolist()
assert True not in np.isnan(pred_results['hat_y'] * 0).tolist()
#from pyhealth.models.ecg.dblstm_ws import DBLSTM_WS as model #from pyhealth.models.ecg.denseconv import DenseConv as model #from pyhealth.models.ecg.deepres1d import DeepRES1D as model #from pyhealth.models.ecg.sdaelstm import SDAELSTM as model #from pyhealth.models.ecg.mina import MINA as model #from pyhealth.models.ecg.rcrnet import RCRNet as model #from pyhealth.models.ecg.rf import RandomForest as model #from pyhealth.models.ecg.xgboost import XGBoost as model from pyhealth.evaluation.evaluator import func data_dir = os.path.join(root_dir, 'datasets', 'ecg') expdata_id = '2020.1104.data.diagnose.ecg' # set up the datasets cur_dataset = expdata_generator(expdata_id, root_dir=root_dir) cur_dataset.get_exp_data(sel_task='diagnose', data_root=data_dir) cur_dataset.load_exp_data() cur_dataset.show_data() # initialize the model for training expmodel_id = '2020.1104.ecg.diagnose.' clf = model(expmodel_id=expmodel_id, n_epoch=10, use_gpu=True) clf.fit(cur_dataset.train, cur_dataset.valid) # load the best model for inference clf.load_model() clf.inference(cur_dataset.test) results = clf.get_results() print(results)
