mathurinm

Output of benchopt --version is not updated if pip install . is not rerun

6

This is a bit counter intuitive IMO since I ran `pip install -e .` and thus did not think I had to rerun `pip install` ping @qb3

benchopt generate-results produces html giving a File not found

4

``` (base) ➜ lasso git:(flash) ✗ pwd /home/mathurin/workspace/benchmarks/lasso (base) ➜ lasso git:(flash) ✗ benchopt generate-results -b . Rendering benchmark: . Processing outputs/benchopt_run_2021-11-05_08h30m08.csv Processing outputs/benchopt_run_2021-11-05_08h30m38.csv Processing outputs/benchopt_run_2021-11-05_08h34m48.csv Solver sklearn did not...

some solvers run for a long time and make the plot upper xlimit too large

9

eg  should this be addressed at the solver/bench level, or at the plot level ?

tabmat has no attribute __version__

3

I find it convenient to be able to check directly inside a python shell. ```python In [1]: import tabmat In [2]: tabmat.__version__ --------------------------------------------------------------------------- AttributeError Traceback (most recent call last) in...

It is not used in the Lasso case, and for the logreg case we can avoid this design with the declaration of inv_lc_j inside the various if/else of the cd...

NOMRG ENH: add pyproject.toml to fix wheel building and install from github

2

fixes #122

ENH use Xw - y instead of y - Xw as dual point

1

This is the classic Fenchel-Rockafellar primal dual link. It will be easier to check the maths, notably in the enet case.

ENH: slow solver on large scale problems with majority of features screened

1

Finance: ```python import time import libsvmdata import numpy as np from numpy.linalg import norm from celer import Lasso X, y = libsvmdata.fetch_libsvm("finance", min_nnz=3) alpha_max = norm(X.T @ y, ord=np.inf) /...

warnings in Logisticregression with prox newton

7

Running pytest: ```python celer/tests/test_logreg.py::test_LogisticRegression[True] /home/mathurin/workspace/celer/celer/homotopy.py:311: ConvergenceWarning: Objective did not converge: duality gap: 0.841892524980679, tolerance: 0.005545177444479563. Increasing `tol` may make the solver faster without affecting the results much. Fitting data with...

BUG: installing celer with pip from github first fails, then succeeds

2

To reproduce in a newly created environement `test` where pip in installed: ``` (test) ➜ benchOpt git:(pip2conda) ✗ which pip /home/mathurin/miniconda3/envs/test/bin/pip (test) ➜ benchOpt git:(pip2conda) ✗ pip install git+https://github.com/mathurinm/celer.git Collecting...