Comments (4)
Hi, thanks for your attention! I guess the problem is in the version of python and pytorch. Do you use python 2.7 and pytorch 0.3?
from person-reid-triplet-loss-baseline.
Hi
Thanks for your reply.
Problem fixed now I got similar results as reported in this repo with more epochs of training. But it's still strange that since the precision are both around 99.99% and I don't know why the results are so different. Anyway thanks for your work.
from person-reid-triplet-loss-baseline.
Glad to hear your results. The number of possible triplets for the training set is huge. So training longer helps the model to see more triplets, which may be useful, although training precision has reached 99%.
In this code, a total of 300 epochs seems satisfying, which is better than 150 epochs. How many epochs did you train that gives you 66% CMC1?
from person-reid-triplet-loss-baseline.
Can't remember. I think I just used the default 150 epochs. I just realized that I stopped and the whole training process in my first try. So I probably have tested with an checkpoint instead of the final trained model. Then it should be my problem. Thanks for your reply and are you on Wechat? If so please add me 382968762
from person-reid-triplet-loss-baseline.
Related Issues (20)
- How can I refer your model in my paper HOT 1
- the different with open-reid HOT 1
- 关于一个epoch中的step数目 HOT 2
- Performance is worse than yours
- 关于d_ap的值 HOT 5
- Error: when i try to run train.py (no module named tri_loss. dataset ) HOT 1
- except Exception, msg: SyntaxError: invalid syntax
- 为什么triloss用的是MarginRankingLoss而不是TripletMarginLoss
- dist_mat[is_pos].contiguous().view(N, -1), 1, keepdim=True) seems wrong HOT 1
- dist_mat[is_pos].contiguous().view(N, -1), 1, keepdim=True) HOT 2
- 关于实现和paper的区别,请指点一下 HOT 5
- 你好,您的cuda和cudnn选的是什么版本呢 HOT 1
- cannot reproduce from pre-trained model on martket 1501 stride 1 HOT 3
- 关于实验性能请教 HOT 1
- Rank 1, 5, 10 are high, but mAP is low HOT 1
- 损失函数是不是可以更直接呢 HOT 2
- how to save my im_paths and features? HOT 1
- Hi how to train your code with different pre-trained models HOT 4
- 热图
- hard triplet convergence
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 person-reid-triplet-loss-baseline.