Updated code example #5064
Updated code example #5064
Conversation
|
AppVeyor and Travis failed on python 3.4, but it passed in my python34 environment. Any suggestion on what might be wrong or what to check? |
|
The test failures are caused by #5045 and only visible on scipy 0.16. Don't worry about them. |
|
You could have just pushed again into #4964 by the way. |
| >>> y_true_categorical = np.array(["spam", "ham", "ham", "spam", "ham", "spam"]) | ||
| >>> y_prob = np.array([0.1, 0.9, 0.8, 0.3, 0.5, 0.7]) | ||
| >>> brier_score_loss(y_true, y_prob) | ||
| 0.14833333333333332 |
There was a problem hiding this comment.
you should probably use ellipsis instead of all the digits.
|
This says that it is still an open issue. I was going to work on it, but I cannot see any actual problem:
|
|
The tests are failing in this PR. I have restarted travis to get an update on the status (as that was a long time ago), but we cannot merge a PR with failing test. That's for sure a problem. |
|
OK, based on the travis error, the pull request needs to be rebased on master. |
|
fixed in #6841 |
#4804