Official implementation of "Evaluating Scalable Bayesian Deep Learning Methods for Robust Computer Vision", CVPR Workshops 2020.
Home Page: http://www.fregu856.com
License: MIT License
Hello @fregu856,
I'm trying to run the evaluating_bdl/depthCompletion/mcdropout_train.py from your docker image and, as you can see, the code detects the training and validation images, but it's throwing the following error in line 75. Do you know what could be?
root@03559b6308a7:~# python evaluating_bdl/depthCompletion/mcdropout_train.py
DatasetKITTIAugmentation - num unique examples: 85898
DatasetKITTIAugmentation - num examples: 171796
DatasetKITTIVal - num examples: 1000
MaskedL2Gauss
model_mcdropout.py
Traceback (most recent call last):
File "evaluating_bdl/depthCompletion/mcdropout_train.py", line 75, in <module>
for i_iter, batch in enumerate(train_loader):
File "/opt/conda/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 637, in __next__
return self._process_next_batch(batch)
File "/opt/conda/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 658, in _process_next_batch
raise batch.exc_type(batch.exc_msg)
TypeError: Traceback (most recent call last):
File "/opt/conda/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 138, in _worker_loop
samples = collate_fn([dataset[i] for i in batch_indices])
File "/opt/conda/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 138, in <listcomp>
samples = collate_fn([dataset[i] for i in batch_indices])
File "/root/evaluating_bdl/depthCompletion/datasets.py", line 97, in __getitem__
sparse = sparse[(img_h - new_img_h):img_h, int(img_w/2.0 - new_img_w/2.0):int(img_w/2.0 + new_img_w/2.0)] # (shape: (352, 1216))
TypeError: 'NoneType' object is not subscriptable
Thanks for sharing. I noticed that the SGLD implementation here yields much better result i.e. lower KL-Div than when I tried using SGLD implementation from some of previous SGLD papers in the same toy classification experiment (despite that the latter could produce a good classification accuracy but the KL-Div was rather high and the probability distribution plot didn't resemble the HMC reference).
I wonder where the chosen prior came from as I find it hard to get detailed information about setting prior over network parameters in practice from previous SGLD publications (I could find very little clue both from the papers and the source code).
And what is the range for valid value of parameter alpha in the prior formula?
When computing the AUCE, do you consider area above the curve as positive or negative?
nvm, I think any deviation from p should be positive.
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 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.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
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.