File size: 4,134 Bytes
ed7fafe 6969d48 ed7fafe 0c93937 ed7fafe 0c93937 ed7fafe 0c93937 ed7fafe df356fe ed7fafe 0c93937 ed7fafe 0c93937 ed7fafe 5076126 ed7fafe a5bbc1e ed7fafe |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 |
import gradio as gr
from time import time
from scipy import sparse
from scipy import linalg
from sklearn.datasets import make_regression
from sklearn.linear_model import Lasso
def load_dataset():
X, y = make_regression(n_samples=200, n_features=5000, random_state=0)
# create a copy of X in sparse format
X_sp = sparse.coo_matrix(X)
return X,X_sp,y
def compare_lasso_dense():
alpha_dense = 1
alpha_sparse = 0.1
sparse_lasso = Lasso(alpha= alpha_sparse, fit_intercept=False, max_iter=1000)
dense_lasso = Lasso(alpha=alpha_dense, fit_intercept=False, max_iter=1000)
t0 = time()
sparse_lasso.fit(X_sp, y)
# print(f"Sparse Lasso done in {(time() - t0):.3f}s")
elapse1 = time() - t0
t1 = time()
dense_lasso.fit(X, y)
# print(f"Dense Lasso done in {(time() - t0):.3f}s")
elapse2 = time() - t1
# compare the regression coefficients
coeff_diff = linalg.norm(sparse_lasso.coef_ - dense_lasso.coef_)
# print(f"Distance between coefficients : {coeff_diff:.2e}")
return f"Sparse Lasso done in {(elapse1):.3f}s\t\n" + f"Dense Lasso done in {(elapse2):.3f}s\t\n" + f"Distance between coefficients : {coeff_diff:.2e}\t\n"
def compare_lasso_sparse():
# make a copy of the previous data
Xs = X.copy()
# make Xs sparse by replacing the values lower than 2.5 with 0s
Xs[Xs < 2.5] = 0.0
# create a copy of Xs in sparse format
Xs_sp = sparse.coo_matrix(Xs)
Xs_sp = Xs_sp.tocsc()
# compute the proportion of non-zero coefficient in the data matrix
print(f"Matrix density : {(Xs_sp.nnz / float(X.size) * 100):.3f}%")
matrix_density = Xs_sp.nnz / float(X.size) * 100
alpha_dense = 1
alpha_sparse = 0.1
sparse_lasso = Lasso(alpha= alpha_sparse, fit_intercept=False, max_iter=1000)
dense_lasso = Lasso(alpha=alpha_dense, fit_intercept=False, max_iter=1000)
t0 = time()
sparse_lasso.fit(Xs_sp, y)
print(f"Sparse Lasso done in {(time() - t0):.3f}s")
elapses1 = time() - t0
t1 = time()
dense_lasso.fit(Xs, y)
print(f"Dense Lasso done in {(time() - t1):.3f}s")
elapses2 = time() - t1
# compare the regression coefficients
coeff_diff = linalg.norm(sparse_lasso.coef_ - dense_lasso.coef_)
print(f"Distance between coefficients : {coeff_diff:.2e}")
return f"Matrix density : {(Xs_sp.nnz / float(X.size) * 100):.3f}%\t\n"+ f"Sparse Lasso done in {(elapses1):.3f}s\t\n" + f"Dense Lasso done in {(elapses2):.3f}s\t\n" + f"Distance between coefficients : {coeff_diff:.2e}\t\n"
X,X_sp,y = load_dataset()
# compare_lasso_dense(X,X_sp,y)
# compare_lasso_sparse(X,X_sp,y)
title = " Lasso on Dense and Sparse data "
info = '''**Comparing the two Lasso implementations on Dense data**
We create a linear regression problem that is suitable for the Lasso, that is to say, with more features than samples.
We then store the data matrix in both dense (the usual) and sparse format, and train a Lasso on each. We compute the
runtime of both and check that they learned the same model by
computing the Euclidean norm of the difference between the coefficients they learned.
Because the data is dense, we expect better runtime with a dense data format.
'''
info2='''***Comparing the two Lasso implementations on Sparse data***
We make the previous problem sparse by replacing all small values with 0
and run the same comparisons as above. Because the data is now sparse,
we expect the implementation that uses the sparse data format to be faster.
'''
conclusion = '''**Conclusion**
We show that linear_model.Lasso provides the same results for dense and sparse data and that in the case of sparse data the speed is improved**.
'''
with gr.Blocks() as demo:
gr.Markdown(f"# {title}")
gr.Markdown(info)
txt_3 = gr.Textbox(value="", label="Dense Lasso comparison")
btn = gr.Button(value="Dense Lasso comparison")
btn.click(compare_lasso_dense, outputs=[txt_3])
gr.Markdown(info2)
txt_4 = gr.Textbox(value="", label="Sparse Lasso comparison")
btn = gr.Button(value="Sparse Lasso comparison")
btn.click(compare_lasso_sparse, outputs=[txt_4])
gr.Markdown(conclusion)
if __name__ == "__main__":
demo.launch()
|