MLJ Results display error

General Usage
1

Hi. I have run the first intro example in MLJ (in Jupyter), and it seems to have run OK, but the display of the results seems garbled:

evaluate(my_tree_model, X, y,
                resampling=CV(shuffle=true), measure=cross_entropy, verbosity=0)

┌───────────────────────────────────┬───────────────┬───────────────────────────
│ _.measure                         │ _.measurement │ _.per_fold               ⋯
├───────────────────────────────────┼───────────────┼───────────────────────────
│ \e[34mLogLoss{Float64} @298\e[39m │ 1.92          │ [1.44, 2.88, 1.44, 1.44, ⋯
└───────────────────────────────────┴───────────────┴───────────────────────────
                                                                1 column omitted
_.per_observation = [[[2.22e-16, 2.22e-16, ..., 2.22e-16], [2.22e-16, 2.22e-16, ..., 2.22e-16], [2.22e-16, 2.22e-16, ..., 2.22e-16], [2.22e-16, 2.22e-16, ..., 2.22e-16], [2.22e-16, 2.22e-16, ..., 2.22e-16], [2.22e-16, 2.22e-16, ..., 2.22e-16]]]
_.fitted_params_per_fold = [ … ]
_.report_per_fold = [ … ]

It seems there lots of ‘\e[34m’-type (meta-character?) expresssions. Is there a way to suppress these?

Thanks for any help.

2

These are colours. You can turn these off at the MLJ level using MLJ.color_off()

1 Like
3

Hiya, Thanks.

[misunderstood what you meant by ‘using’ at first]