Comments (6)
I added trainer.module.save(checkpoint_directory, iterations)
in Line 319 of main.py so that you can store weights of stage 3.
You got the correct ReID results, but not generation results. It seems to me that something is wrong in generate_data.py. I will rerun the code to figure out the problem.
from gcl.
Another question: Did you also get bad quality images in Tensorboard by runningtensorboard --logdir logs/market_init_JVTC_unsupervised/
?
from gcl.
Another question: Did you also get bad quality images in Tensorboard by running
tensorboard --logdir logs/market_init_JVTC_unsupervised/
?
Actually I tried to use the tensorboard, but I was using ssh remote server and docker. I am now trying to make tensorboard work, but I'm still working on it.
-> I could not see the results of your tensorboard yet. So, I was trying to save the results in the folder by using examples/generate_data.py!
Actually I couldn't resume the model.best.pth weights(which is the Best Re-ID result) into the model in examples/generate_data.py.
So, do I only need to load the the best Re-ID result weights of id_net.pt, generator.pt and discrimminator.pt by using your code ?
model_1 = create_model(args)
trainer = DGNet_Trainer(config, model_1, args.idnet_fix).cuda()
iterations = trainer.resume(checkpoint_path, hyperparameters=config)
It seems to me that something is wrong in generate_data.py. I will rerun the code to figure out the problem.
-> Thank you so much.
from gcl.
Actually, this is the image saved during training in stage3 when re-ID score is the best. It seems okay. I think the problem in examples/generate_data.py, the meshes are not properly brought in with the original images.
from gcl.
Yes, you are right. It was the problem of incoherent data-dir and mesh-dir format in Line 240 of generate_data.py. I just updated them. Please use the new generate_data.py.
For reconstruction, use python examples/generate_data.py --mode recon
For new views, use python examples/generate_data.py --mode nv --degree 45
(degree could be 45, 90, 135, 180, 225, 270, 315)
from gcl.
Thank you. It is now working!
from gcl.
Related Issues (17)
- FileNotFoundError: [Errno 2] No such file or directory: '/data/zundeng/download/Market/train/render/0516_c5s1_143070_01.jpg' HOT 9
- About Generative module and GAN loss HOT 1
- torch version issue. HOT 2
- when i run the file train HOT 1
- when i run the file train_stage3_market.sh,it errors. HOT 3
- load the pretrained identity enconder HOT 1
- run the file train_stage3_market.sh, IndexError:list index out of range HOT 2
- when I run "sh train_stage2_market.sh",one problem happened HOT 1
- 是否可以提供预训练模型 HOT 1
- The fid in Eq(2) HOT 1
- How do I get test result in the paper? HOT 1
- About 3D Mesh generator. HOT 11
- How did you warm up identity encoder?? HOT 3
- Some questions in reproducing the code in windows OS HOT 5
- Feature memory initialization issue. HOT 1
- 训练问题
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 gcl.