Comments (5)
@Tetsujinfr Hi, the weights checkpoints has been updatated.
Sorry for the delay baecause of Chinese Lunar New Year.
from mcmot-bytetrack.
Hi @CaptainEven
so I managed to get the code running but I have some strange results.
On the test13 video, the results I get are not so good I think, with a lot of IDs where the ID changes for one to another while there is no occlusion happening.
2022_04_14_02_51_10_low.mp4
Here is another video where the results are completely off:
2022_04_14_02_55_28_low.mp4
Here is my test command (I did move the test demo file at the repo root and changed the cfg path accordingly):
python3 demo_mcmot.py --reid True -f exps/example/mot/yolox_tiny_track_c5_darknet.py --ckpt weights/latest_ckpt.pth.tar --path /home/tetsfr/Videos/video1.mp4
What am I doing wrong?
thanks for your help
from mcmot-bytetrack.
hi @CaptainEven
could you give a bit more guidance on how to run the demo with REID on the 5 classes pretrained model ? thanks
Hi, Can you run the demo correctly now?
If so, could you please give me some guidance?
from mcmot-bytetrack.
@Zhangyongtao123 no unfortunately I did not push further. Hopefully @CaptainEven can give us a bit morr guidance on how to reproduce the 5 classes results or highlight the piece of code which needs to be tweaked to make sure the 5 classes are recognized properly on frames in the wild.
from mcmot-bytetrack.
Hi @CaptainEven
so I managed to get the code running but I have some strange results. On the test13 video, the results I get are not so good I think, with a lot of IDs where the ID changes for one to another while there is no occlusion happening.
2022_04_14_02_51_10_low.mp4
Here is another video where the results are completely off:2022_04_14_02_55_28_low.mp4
Here is my test command (I did move the test demo file at the repo root and changed the cfg path accordingly):python3 demo_mcmot.py --reid True -f exps/example/mot/yolox_tiny_track_c5_darknet.py --ckpt weights/latest_ckpt.pth.tar --path /home/tetsfr/Videos/video1.mp4
What am I doing wrong? thanks for your help
I guess your setting is “tracker = byte” and "reid = true", you should note that based bytetracker set the detection thresh is 0.001, so maybe a lot of False positives will be apparent, so you should adjust this parameter such as 0.5.
from mcmot-bytetrack.
Related Issues (20)
- Training on Visdrone dataset HOT 1
- basic info to run the mcmot bytetrack HOT 2
- Run demo_mcmot.py IndexError: index 1 is out of bounds for dimension 0 with size 1 HOT 3
- Failed to run demo_mcmot.py, Unexpected key(s) in state_dict: "mtl_loss.sigma". HOT 2
- Still got error "Unexpected key(s) in state_dict: "mtl_loss.sigma", after change branch to DarknetCfg, and use the latest checkpoint HOT 5
- About C5 dataset, searched everywhere, but still cannot find where to download the c5 dataset. HOT 1
- Trouble downloading weight (checkpoint) HOT 1
- max_id_dict.npz HOT 1
- 跟踪其他物体? HOT 2
- Checkpoint
- Can you give a demo of how to run the code? HOT 1
- New position estimation of object without detection
- How can I train with custom datasets? HOT 7
- Using detection confidence from other detector
- No class_id in the output HOT 2
- "CUDA out of memory" on dataset with 300 classes. HOT 1
- IndexError in multitracker.py
- Reid with ocsort? HOT 1
- How to train my own dataset?
- no module name yolox_tiny_track_c5_darknet
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from mcmot-bytetrack.