[Fix]modify dict keys of sot tracking outputs #223
Conversation
JingweiZhang12
changed the title
|
please fix the lint error firstly. |
mmtrack/datasets/lasot_dataset.py
Outdated
| @@ -75,7 +75,8 @@ def evaluate(self, results, metric=['track'], logger=None): | |||
| inds.append(len(self.data_infos)) | |||
|
|
|||
| track_results = [ | |||
There was a problem hiding this comment.
Change track_results to track_bboxes.
Please also fix the unittest errors.
There was a problem hiding this comment.
we should change the code to :
track_bboxes = [
results['track_results'][inds[i]:inds[i + 1]][:4]
for i in range(num_vids)
]
There was a problem hiding this comment.
results['track_results'] is the list[list[nparray]] format
tests/test_data/test_dataset.py
Outdated
| @@ -452,7 +452,7 @@ def test_lasot_evaluation(): | |||
| x1, y1, x2, y2 = result.split(',') | |||
| track_results.append(np.array([int(x1), int(y1), int(x2), int(y2)])) | |||
There was a problem hiding this comment.
The track_results need append a score to be consistent with the above code.
Codecov Report
@@ Coverage Diff @@
## master #223 +/- ##
==========================================
- Coverage 69.81% 69.37% -0.45%
==========================================
Files 87 87
Lines 4496 4496
Branches 877 877
==========================================
- Hits 3139 3119 -20
- Misses 1060 1071 +11
- Partials 297 306 +9
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
mmtrack/models/sot/siamrpn.py
Outdated
| @@ -262,7 +263,7 @@ def simple_test(self, img, img_metas, gt_bboxes, **kwargs): | |||
| self.memo.bbox = bbox_xyxy_to_cxcywh(gt_bboxes) | |||
| self.memo.z_feat, self.memo.avg_channel = self.init( | |||
| img, self.memo.bbox) | |||
| best_score = None | |||
| best_score = -1 | |||
There was a problem hiding this comment.
-1. We need be careful about the type of variables
tools/test.py
Outdated
| for key in [ | ||
| 'interval', 'tmpdir', 'start', 'gpu_collect', 'save_best', | ||
| 'rule', 'by_epoch' | ||
| ]: |
There was a problem hiding this comment.
Too ugly here. we can modify as following:
eval_hook_args = [ 'interval', 'tmpdir', 'start', 'gpu_collect', 'save_best' , 'rule', 'by_epoch']
for key in eval_hook_args:
tests/test_data/test_dataset.py
Outdated
| @@ -450,9 +450,9 @@ def test_lasot_evaluation(): | |||
| track_results = [] | |||
| for result in results: | |||
| x1, y1, x2, y2 = result.split(',') | |||
| track_results.append(np.array([int(x1), int(y1), int(x2), int(y2)])) | |||
| track_results.append(np.array([int(x1), int(y1), int(x2), int(y2), 0])) | |||
replace "score" and "bbox" to "track_results" in dict keys of sot tracking outputs