Comments (6)
Hi, if you do not want to do the validation during training, just remove the validation code in the train.py:
val_err[0] = val(val_dataloader, model)
Besides, you should change the 'parse_arg_val.py'.
--anno_phase='train'
from vnl_monocular_depth_prediction.
Thank you for your response. I finished the problem. thanks to you..
now i faced another problem with test.. i tried to test with pre-trained data
python ./tools/test_any_images.py
--dataroot ./
--dataset any
--cfg_file lib/configs/resnext101_32x4d_nyudv2_class
--load_ckpt model/nyu_rawdata.pth
i downloaded the pretrain data and give it into the created model folder. But the result is strange. When i open it, all images becomes grey and has 50000 value. i tried it with your images in example, and the result are the same. Is there something wrong with my procedure?
from vnl_monocular_depth_prediction.
Sorry for late reply. Have you fixed this problem?
from vnl_monocular_depth_prediction.
Yes, i solved it already. Thank you..
but i have another problem,
i always out of Cuda memory while training. I am using GTX Titan X with 11 GB memory, it is obviously weird. i already try to reduce the batchsize and try the other batchsize, but it doesnt work. How i can solve this? where i need to change the parameter for this?
Also, can i ask another question?
i have a dataset with the same structure of train.mat.
i want to update/fine tune the nyu_rawdata.pth with my dataset.
Can i just load and train it with resume option?
from vnl_monocular_depth_prediction.
Hi, if out of memory, you can adjust the image size for training. Please see the config file under lib/core/config.py. You can adjust it to 256*256 or others.
Yes, you can. However, you should set the new learning rate for fine-tuning.
from vnl_monocular_depth_prediction.
Thank you for your response. I finished the problem. thanks to you..
now i faced another problem with test.. i tried to test with pre-trained data
python ./tools/test_any_images.py
--dataroot ./
--dataset any
--cfg_file lib/configs/resnext101_32x4d_nyudv2_class
--load_ckpt model/nyu_rawdata.pthi downloaded the pretrain data and give it into the created model folder. But the result is strange. When i open it, all images becomes grey and has 50000 value. i tried it with your images in example, and the result are the same. Is there something wrong with my procedure?
HI,aufaclav :
I have the same problem as you. When I use the model to predict the depth map of the picture, the results are all gray. How did you solve this problem? thank you
from vnl_monocular_depth_prediction.
Related Issues (20)
- Performance issue HOT 2
- Error while loading the model HOT 1
- How makew
- How make inference on single image? HOT 3
- Only can train 1 epoch? HOT 3
- Setting for training in ablation study HOT 1
- Some questions about surface normal estimation and robutness test HOT 3
- Might it be a small false figure reference of the paper uploaded on Arxiv? HOT 1
- how can I train with NYUD-V2 dataset HOT 1
- About the Camera Parameters HOT 2
- How can I generate the dense ground truth depth maps in KITTI? HOT 6
- How to generate a point cloud map? HOT 1
- Error when running train_kitti_metric.py HOT 1
- pretaind resnext101_32x4d.pth HOT 2
- abs_rel value
- yaml_cfg load error HOT 1
- Could you please provide the pretrain model of moblinenetv2? HOT 1
- The test_any_image file cannot correct output img and the test_nyu file output very bad quality image! HOT 2
- Questions about datasets
- How to understand the concept of convert depth to Point Cloud
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 vnl_monocular_depth_prediction.