Comments (1)
pengsongyou
commented on June 7, 2024
Hi @star9988rr,
Thanks for your wonderful work! I use a 40G A100 to train ScanNet with a batch size of 8, but the training time is about 2 days. Is this right?
I believe that is correct if you try to train for 100 epochs. However, we noticed that after 50-60 epochs one can already obtain the best performance.
Besides, when I increase the batch size to 16 (the learning rate is also doubled), I only get 41.9 (46.0 in paper) mIoU on ScanNet. Have you tried with the batch size of 16?
I have not tried that, but probably you should not double your learning rate. You can try again with a batch size of 16 with the same learning rate.
from openscene.
Related Issues (20)
- Confusion about sinput of model HOT 1
- MinkowskiEngine HOT 1
- Misalignment in unprojected RGBD images HOT 5
- num_rand_file_per_scene HOT 6
- Eval on train dataset error HOT 3
- Question about labelset HOT 1
- Question about 3D point cloud visualization HOT 1
- Create a multi-view fusion feature using median method HOT 1
- Question about the labeling of scannet 2d data HOT 1
- Loading images
- Hi I want to do segmentation only on 3D point cloud how can I use the scripts to do so for a demo and which dataset HOT 1
- Checkpoint with input color HOT 2
- meet error when training on 1 GPU
- Issue running evaluation on AWS: version `GLIBC_2.33' not found HOT 4
- Should also add libxcb-cursor0 as the dependency for the Qt6.50 HOT 1
- Guidance on how to eval on my own dataset
- How to conduct open-vocabulary 3D search? HOT 1
- NaN or Inf found in input tensor
- About the details in demo HOT 1
- One file is missing in the nuscenes_3d val directory. HOT 3
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 openscene.