mirrors
/
fast-reid
mirror of https://github.com/JDAI-CV/fast-reid.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

updata readme

pull/25/head
sherlock 2018-10-18 19:00:52 +08:00
parent 66d2ce8bf6
commit 5cc38c7e48
2 changed files with 8 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
README.md
View File

@ -45,14 +45,17 @@ in `reid_baseline` folder if you want to train with softmax and triplet loss. Yo
## Results
**network architecture**
ResNet50 -> global avg pooling -> BN(freeze beta) -> softmax (triplet)
<img src='https://ws3.sinaimg.cn/large/006tNbRwly1fvh3ekjh12j315k0j4q58.jpg' width='500'>
| config | Market1501 | CUHK03 |
| --- | --| ---|
| --- | --| ---| --- |
| bs(32) size(384,128) softmax | 92.2 (78.5) | |
| bs(64) size(384,128) softmax | 92.5 (79.6) | |
| bs(32) size(256,128) softmax | 92.0 (78.4) | |
| bs(64) size(256,128) softmax | 91.7 (78.3) | |
| triplet | 88.8% | 74.8% |
|triplet + softmax | 92.0% | 78.1% |
| bs(128) size(256,128) softmax | 91.2 (77.4) | |
| triplet(p=32,k=4) size(256,128) | 88.3 (73.8) | |
| triplet(p=16,k=4)+softmax size(384,128) | 93.1 (82.0) | |
| triplet(p=24,k=4)+softmax size(384,128) | 91.7 (79.0) | |

1
scripts/test.sh 100644
View File

@ -0,0 +1 @@
#!/usr/bin/env bash