Comments (5)
Another way I see is that the loss is calculated per timestep so at time t and then loss.backward() is done at that time frame.
from mono3d.
Actually, we can also concatenate the output per timeframe and compute the loss at the end of the batch. Is this how loss should be calculated?
from mono3d.
Hi @mcsuy1998 ,
Total loss formulation is shown in Eq. 3. We compute the total loss based on a batch of sequence, and then backward the loss.
from mono3d.
I see. I wanted to ask how to train the video model. I know you initialized the video model with a pretrained image model. Is the learning rate of the video model the same as the image model? You also mentioned that video model uses extra 1x1 convolutions weights. Did you use a different learning rate for these additional weights?
from mono3d.
- the learning rate was initially set to 0.0001 and further reduced by a factor of 3.33 when the loss plateaus (known as ReduceLRonPlateau in PyTorch)
- We didn't adopt such a strategy. But I think it is worth a try to set a lower learning rate for the additional weights.
from mono3d.
Related Issues (6)
- Path Issue? HOT 1
- Training code HOT 1
- Training for Video HOT 1
- CNS Module HOT 1
- CNS module in image version model
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 mono3d.