Comments (7)
done, similar performance, and I notice that you uploaded the training code, close issue.
from lite-mono.
Hi, I will upload the pre-trained weights soon and let you know.
from lite-mono.
Thanks.
from lite-mono.
Hi @FangGet , I have uploaded the pre-trained weights. A tip for training, you need to set the weight_decay
to 1e-2
and drop_path
to 0.2
to mitigate overfitting. For the latest Lite-Mono-8M, I use a larger drop path rate, which is 0.4.
Please let me know if you have more questions.
from lite-mono.
Thanks @noahzn .
With AdamW optimizer, weight_decay=1e-2, drop_path=0.2 learning_rate=1e-4, pretrained Imagenet, I got the results:
name | abs_rel | sq_rel | rmse | rmse_log | a1 | a2 | a3 |
---|---|---|---|---|---|---|---|
paper | 0.107 | 0.765 | 4.561 | 0.183 | 0.886 | 0.963 | 0.983 |
own_trained | 0.108 | 0.798 | 4.658 | 0.187 | 0.885 | 0.961 | 0.982 |
Gap still exists, waiting for your training detail(maybe different learning rate for depth and pose network?).
from lite-mono.
Hi! Glad to see that you trained the network by yourself! Some comments:
- I used the initial learning rate
1e-4
for both the DepthNet and the PoseNet of Lite-Mono. Then, the learning rates decrease to5e-6
and1e-5
, respectively. - The best results are usually not achieved in the last epoch due to training fluctuations.
- I used the cosine annealing scheduler to reduce learning rates.
- The drop path rate introduces uncertainty into the training. The same results cannot be got unless the randomness of the training is fixed, by setting fixed seeds.
- I have just done several rounds of training after reading your comment, and found that some training achieved better results in 2 or 3 metrics than the results reported in the paper. But also one round of training couldn't converge.
- You can use Tensorboard to check the training process.
tensorboard --logdir ./tmp/your_saved_model_folder
.
from lite-mono.
Thanks.
- useful, I will try that.
- I also get best results not from last epoch.
- Cosine annealing has also been tried, but with a much smaller eta lr. I will try it with tips. 1.
- 4.5.6. understand.
thanks again.
from lite-mono.
Related Issues (20)
- Hi,I've got some new questions HOT 9
- Some question about the result HOT 4
- The edges of the image are foggy and blurry during the training process HOT 4
- Inexplicable "No such file or directory: 'Lite-Mono-main\\kitti_data\\\2011_09_26/2011_09_26_drive_0002_sync\\\image_02/data\\- 000000001.png'" HOT 2
- the CPU utilization has been very high, but the GPU utilization has been very low HOT 2
- training with Nuscenes dataset HOT 39
- About indicator sq_rel HOT 6
- Test outputs is black HOT 8
- I read your paper “For models trained from scratch an initial learning rate of 5e−4 with a cosine learning rate schedule [26] is adopted” But how should I implement it in the code? HOT 4
- How the reproduction of the results of the model achieves the results in the paper HOT 6
- Hello, my reproduction code is exactly the same as the code uploaded by the author of the paper, CUDA11.0, PyTorch 1.7.1. Please check your code and replace your CUDA and pytorch version.
- model reproduction HOT 11
- Hello, I have some questions regarding the model training with the KITTI dataset HOT 4
- Pre-training Weights HOT 1
- about pretrain acc HOT 1
- the results of training are black pictures HOT 5
- test code for pose HOT 2
- Subsequent issues with gt_depths.npz HOT 4
- sorry to bothering you HOT 1
- Output resolution problem HOT 5
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 lite-mono.