[Add] Support HOTA evaluation for MOTChallenge. #417
Conversation
| # Use 'track' as the default tracker | ||
| TRACKERS_TO_EVAL=['track'], | ||
| # Option values: ['pedestrian'] | ||
| CLASSES_TO_EVAL=['pedestrian'], |
There was a problem hiding this comment.
We have self.classes, is CLASSES_TO_EVAL related to this item?
| # Option values: ['pedestrian'] | ||
| CLASSES_TO_EVAL=['pedestrian'], | ||
| # Option Values: 'MOT17', 'MOT16', 'MOT20', 'MOT15' | ||
| BENCHMARK=self.img_prefix.rsplit('/', 2)[-2], |
There was a problem hiding this comment.
This is risky, write a function to convert lower-case names, and raise errors if there are no specific benchmarks in their names.
| 'to manually install trackeval') | ||
|
|
||
| seqmap = osp.join(self.img_prefix, 'videoseq.txt') | ||
| if not osp.exists(seqmap): |
There was a problem hiding this comment.
Add document instructions to indicate dataset preparations.
Codecov Report
@@ Coverage Diff @@
## master #417 +/- ##
==========================================
+ Coverage 70.89% 70.98% +0.09%
==========================================
Files 117 117
Lines 6249 6296 +47
Branches 1198 1208 +10
==========================================
+ Hits 4430 4469 +39
- Misses 1458 1459 +1
- Partials 361 368 +7
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
| 'to manually install trackeval') | ||
|
|
||
| # get video name sequence for HOTA evaluation | ||
| seqmap = osp.join(self.img_prefix, 'videoseq.txt') |
There was a problem hiding this comment.
Can we online generate this txt ?
|
change CI and install.md |
docs/zh_cn/install.md
Outdated
| @@ -51,7 +51,16 @@ | |||
|
|
|||
| 如果不是安装预构建的包，而是从源码中构建 PyTorch，则可以使用更多的 CUDA 版本，例如 CUDA 9.0。 | |||
|
|
|||
| 3. 安装 mmcv-full，我们推荐您安装以下预构建包： | |||
| 3. 为VOT数据集测试评估安装额外库（可选） | |||
There was a problem hiding this comment.
typically, we need add a blank space between Chinese and English
GT9505
changed the title
No description provided.