def _regress_out_chunk(X, y):
"""
Performs a data_cunk.shape[1] number of local linear regressions,
replacing the data in the original chunk w/ the regressed result.
Parameters
----------
X : cupy.ndarray of shape (n_cells, 3)
Matrix of regressors
y : cupy.sparse.spmatrix of shape (n_cells,)
Sparse matrix containing a single column of the cellxgene matrix
Returns
-------
dense_mat : cupy.ndarray of shape (n_cells,)
Adjusted column
"""
y_d = y.todense()
lr = LinearRegression(fit_intercept=False, output_type="cupy")
lr.fit(X, y_d, convert_dtype=True)
return y_d.reshape(y_d.shape[0], ) - lr.predict(X).reshape(y_d.shape[0])
def _regress_out_chunk(X, y):
"""
Performs a data_cunk.shape[1] number of local linear regressions,
replacing the data in the original chunk w/ the regressed result.
"""
output = []
lr = LinearRegression(fit_intercept=False)
lr.fit(X, y, convert_dtype=True)
return y.reshape(y.shape[0], ) - lr.predict(X).reshape(y.shape[0])
def _regress_out_chunk(data_chunk, regressors):
"""
Performs a data_cunk.shape[1] number of local linear regressions,
replacing the data in the original chunk w/ the regressed result.
"""
output = []
for col in range(data_chunk.shape[1]):
y = data_chunk[:, col]
X = regressors
lr = LinearRegression(fit_intercept=False)
lr.fit(X, y, convert_dtype=True)
mu = lr.predict(X)
data_chunk[:, col] = y - mu
return data_chunk
model = regression.AutoSklearnRegressor(
time_left_for_this_task=alg.task_time,
per_run_time_limit=alg.run_time,
resampling_strategy=alg.sampling_strategy,
resampling_strategy_arguments={'folds': alg.folds}
)
else:
model = regression.AutoSklearnRegressor(
time_left_for_this_task=alg.task_time,
per_run_time_limit=alg.run_time
)
warn_not_gpu_support(alg)
elif alg.name == 'LinearRegression':
if NVIDIA_RAPIDS_ENABLED:
from cuml.linear_model import LinearRegression
model = LinearRegression(**alg.input_variables.__dict__)
else:
from sklearn.linear_model import LinearRegression
model = LinearRegression(**alg.input_variables.__dict__)
elif alg.name == 'SupportVectorRegression':
if NVIDIA_RAPIDS_ENABLED:
from cuml.svm import SVR
else:
from sklearn.svm import SVR
model = SVR(**alg.input_variables.__dict__)
elif alg.name == 'BayesianRidgeRegression':
from sklearn.linear_model import BayesianRidge
model = BayesianRidge(**alg.input_variables.__dict__)
warn_not_gpu_support(alg)
elif alg.name == 'AdaBoost' and alg.type == 'regression':
from sklearn.ensemble import AdaBoostRegressor
def function(self):
self.out_1.val = LinearRegression()
def lr_gpu():
lr_gpu = LRCUML(fit_intercept=True, normalize=False, algorithm="eig")
res = lr_gpu.fit(X, y)