FIX: bugs in evaluator #590
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…eased the robustness of data.utils.data_preparation, which can divide the dataset into two parts (train, test) or three parts (train, valid, test).
…dle empty dataset now.
FEA: Add config['benchmark_filename'] to load pre-split dataset.
FEA: Increased the robustness of trainer.evaluate && bug fix in GeneralFullDataLoader.
FIX: optimize the update_attentive_A function in KGAT
|
@guijiql Add some test cases in the |
tsotfsk
suggested changes
Dec 22, 2020
recbole/evaluator/proxy_evaluator.py
Outdated
| for metrics, evaluator in metric_eval_bind: | ||
| used_metrics = list(metrics_set.intersection(set(metrics.keys()))) | ||
| used_metrics = [metric for metric in metrics_list if metric in metrics.keys()] |
There was a problem hiding this comment.
in metrics.keys() -> in metrics
recbole/evaluator/metrics.py
Outdated
| all_with_pos = np.any(pos_len_list == 0) | ||
| all_with_neg = np.any(neg_len_list == 0) | ||
| non_zero_idx = np.full(len(user_len_list), True, dtype=np.bool) | ||
| if all_with_pos: |
There was a problem hiding this comment.
What's the meaning of this? why does all_with_pos mean np.any(pos_len_list == 0) , if pos_len_list = array([1,2,3]), then np.any(pos_len_list == 0) is False, so all_with_pos is False?
tsotfsk
suggested changes
Dec 22, 2020
tests/metrics/test_rank_metrics.py
Outdated
| from recbole.evaluator.metrics import metrics_dict | ||
|
|
||
|
|
||
| class TestCases(object): |
There was a problem hiding this comment.
Maybe, pos_rank_sum can‘t ensure your metric is right, please test the function evaluator.collect to ensure your pos_rank_sum is right. The reason to add the test is that it requires complex logic to calculate it.
tsotfsk
approved these changes
Dec 24, 2020
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
1.Individual Evaluator can't raise NotImplementedError when used with eval_setting is full.
2.metrics may be disordered in log information.
3.GAUC will be calculated by error when there are items with same predicted scores for one user.