Sam Coveney comments

Results 86 comments of


                                            Sam Coveney

BF: Allow passing kwargs in `fit` method, by moving parallelization kwargs elsewhere

can I also ask: why is this solution here "brittle"? Does this solution allow for each parallel worker to receive a _different_ value for the same kwarg, as would be...

BF: Allow passing kwargs in `fit` method, by moving parallelization kwargs elsewhere

Thanks for the work. I'm afraid I can't tell from the PR easily, whether this makes it possible to pass a different keyword to every parallel process. Can you tell...

BF: Allow passing kwargs in `fit` method, by moving parallelization kwargs elsewhere

Much thanks for the work on this so far @arokem and @skoudoro

NF: New workflow: dipy_correct_bvecs.

> ``` > > import numpy as np > data = np.loadtxt("~/.dipy/stanford_hardi/HARDI150.bvec") > > print("Original bvecs data",data[0,25:29]) > data[0,25:29] *= -1 > print("Changed bvecs data",data[0,25:29]) > > np.savetxt("flipped.bvec", data) >...

iter_fit_tensor returns {} instead of None causing KeyError in nlls_fit_tensor when return_leverages=False

Do you mind if I look into this first?

iter_fit_tensor returns {} instead of None causing KeyError in nlls_fit_tensor when return_leverages=False

So I can't quite see how this is coming about on first look, because the `ols_fit_tensor` function looks like it would in fact return `None` for leverages if `return_leverages=False` :...