New basis-function-based methods module#155
Conversation
…ran jupyter notebooks
| "output_type": "stream", | ||
| "text": [ | ||
| "100%|██████████| 13/13 [01:56<00:00, 8.96s/it]\n" | ||
| "100%|██████████| 14/14 [01:57<00:00, 8.39s/it]\n" |
There was a problem hiding this comment.
Possibly got faster due to #154. I'm seeing polydiff takes 30-35 seconds to optimize now. It was formerly nearly a minute.
|
|
||
| def test_simulate(): | ||
| return | ||
| # def test_simulate(): |
There was a problem hiding this comment.
Having real tests in here that don't do anything is odd, because they get run. Yet I don't want to remove the stubs, as a reminder that we should maybe have some tests for these. But in the case of simulations, it's really just "Do they all run okay?", to which the answer is a known yes unless we futz with them significantly, and in the case of evaluations the metrics and plotting functions are defined by hand with intentionality. In both cases the more obvious place to see a failure would be in the notebooks, not in a unit test, although we can probably unit test something for safety.
1 participant
To further address #138.
rbfdiffin optimization