https://github.com/L1aoXingyu/reid_baseline_gluon
SOTA results for reid baseline model (Gluon implementation)
https://github.com/L1aoXingyu/reid_baseline_gluon
gluon mxnet-gluon-interface reid
Last synced: about 12 hours ago
JSON representation
SOTA results for reid baseline model (Gluon implementation)
- Host: GitHub
- URL: https://github.com/L1aoXingyu/reid_baseline_gluon
- Owner: L1aoXingyu
- Created: 2018-08-03T09:09:19.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2018-08-06T02:12:38.000Z (over 6 years ago)
- Last Synced: 2025-04-12T16:13:53.813Z (5 days ago)
- Topics: gluon, mxnet-gluon-interface, reid
- Language: Python
- Homepage:
- Size: 23.4 KB
- Stars: 13
- Watchers: 1
- Forks: 7
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- Awesome-MXNet - reid_baseline_gluon
README
# ReID_baseline
Baseline model (with bottleneck) for person ReID (using softmax and triplet loss).
We support
- multi-GPU training
- easy dataset preparation
- end-to-end training and evaluation
## Get Started
1. `cd` to folder where you want to download this repo
2. Run `git clone https://github.com/L1aoXingyu/reid_baseline_gluon.git`
3. Install dependencies:
- [mxnet 1.3.1](http://mxnet.incubator.apache.org/versions/1.2.1/install/index.html)
```
pip install --pre mxnet-cu90
```
- tensorflow (for tensorboard)
- [MXBoard](https://github.com/awslabs/mxboard)
4. Prepare dataset
Create a directory to store reid datasets under this repo via
```bash
cd reid_baseline
mkdir data
```
1. Download dataset to `data/` from http://www.liangzheng.org/Project/project_reid.html
2. Extract dataset and rename to `market1501`. The data structure would like:
```
market1501/
bounding_box_test/
bounding_box_train/
```
5. Prepare pretrained model if you don't have
```python
from mxnet import gluon
gluon.model_zoo.vision.resnet50_v1(pretrained=True)
```
Then it will automatically download model in `~.mxnet/models/`, you should set this path in `config.py`
## Train
You can run
```bash
bash scripts/train_triplet_softmax.sh
```
in `reid_baseline` folder if you want to train with softmax and triplet loss. You can find others train scripts in `scripts`.
## Results
| loss | rank1 | map |
| --- | --| ---|
| softmax | 87.1% | 67.8% |
| triplet | 88.2% | 73.7% |
|triplet + softmax | 90.4% | 76.4% |
I find the mxnet.gluon results are a little bit lower than [pytorch](https://github.com/L1aoXingyu/reid_baseline) results, and I cannot get the reason. I would appreciate that if anyone can help me.