Comments (6)
Hi, thanks for trying out the code. I think maybe the problem is your GPU memory is small, so you encounter the out of memory error. Try reduce the batch size during test.
Specifically, reduce batch_size here and here. Since for each video, we will have 250 frames in total (25 sampled frames, and 10-crop augmentation, so 25 x 10 = 250). Your batch size should be multiples of 5. I suggest try 25 first. If doesn't work, try 10 or even 5. But be warned that the test time will be much longer due to small batch size.
from hidden-two-stream.
Hi, thanks for your advice.
However, the same error still occurs even though I reduced the batch_size
to 1.
Do you have an idea where caffe_gpu_memcpy
is called or what else could cause the problem?
from hidden-two-stream.
Hi, I am sorry I don't have such experience. Usually most people encounter cudasuccess 2 vs 0, which is the out of memory error. I searched online for your problem at here, here and here. I don't quite understand their answers, but it seems that the computation is too much for your card (or the threads management is out of order). But I think 1060 is a pretty decent card, it shouldn't happen such problems.
Are you using multiple cards? And can you reduce the size of the network and try again? I mean, you only use 2 layers for motionnet, and 2 layers for temporal stream. You don't need to train the model long, you just get such a model and test it. If the evaluation script goes well, then it means it is the computation load is too much. Otherwise, we can focus on something else.
from hidden-two-stream.
Hi, thanks for all your help again.
Today I tried it on another pc with a GeForce GTX 960 with 2GB and its seems to be working. I have no idea why it didn't work before. However, I am happy as long as it is working now :)
from hidden-two-stream.
Great to hear it is working now, cheers.
from hidden-two-stream.
Hi,
As a part of my thesis I am implementing this project. sadly i am new to this technology i am getting errors during training. could you please share detailed implementation steps with me. thank you
from hidden-two-stream.
Related Issues (20)
- About training question HOT 3
- question about memory HOT 5
- About Flow loss HOT 2
- when runing demo_hidden.py HOT 4
- the accuracy I get is 69% HOT 1
- Training Motion Net HOT 2
- spatial model HOT 3
- I want to ask a question about training and test networks HOT 2
- about the supervised and unsupervised learning MotionNet HOT 1
- Error while running demo_hidden.py HOT 6
- looking forward the pytorch version HOT 1
- Provide Dockerfile setup
- out of memory while testing HOT 2
- where is the spatial stream cnn? HOT 5
- about compile HOT 6
- Could you please release the pytorch version? HOT 1
- Comparison on THUMOS14
- dataset problem HOT 1
- Seperate training on MotionNet HOT 1
- Data loading for MotionNet training
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 hidden-two-stream.