Comments (5)
Dan, Alexis, thanks for bringing this up. Sorry for the DATA_PATH
-related issue -- on my machines, I usually keep all the data in a dedicated folder to which DATA_PATH
is pointing. I agree that for the examples such a set up is an overkill and shouldn't be a requirement.
Dan, thanks for the code snippets. Similarly to what you suggested, I will update the data utils to allow automatic data retrieving in case it is not found. It's a little bit trickier when you want to support both Python 2 and 3, since pickle formats differ. I will push the updates during the coming week.
You are more than welcome to open a PR with the setup instructions for mac.
from keras-gp.
Hi the following worked for me
- get actuator.mat from http://www.iau.dtu.dk/nnbook/systems.html
- make sysid directory in kgp/kgp/datasets/ ( i.e. kgp/kgp/datasets/sysid/)
- load the mat file in python and save it as actuator.pkl (I have to save 'u' as 'X' and 'p' as 'Y') in this directory kgp/datasets/sysid/
- set your environment variable DATA_PATH to point to kgp/kgp/datasets (full path)
from keras-gp.
Thanks! I wrote a little script to transform the mat file, but it is trivial:
import scipy.io as sio
import cPickle
data = sio.loadmat('actuator.mat')
X = data['u']
Y = data['p']
newdata = {}
newdata['X'] = X
newdata['Y'] = Y
f = file('actuator.pkl', 'wb')
cPickle.dump(newdata, f, protocol=cPickle.HIGHEST_PROTOCOL)
f.close()
but that (of course) just dumps to the local directory (so you have to mkdir the sysid folder structure first and use it there).
Better yet would be to use something like the above and also get the data automatically:
import urllib
urllib.urlretrieve("http://www.iau.dtu.dk/nnbook/Files/actuator.mat","actuator.mat")
and run this in the setup file (with constructing the directory folder structure).
Let me know if you'd think this would be helpful (I could write a PR for it).
Also, I am currently running the example on my mac (10.11.6 El Capitan), so if you'd like setup instructions for doing so, I could provide them in the PR.
Dan
from keras-gp.
Hey Maruan, I run the programs in example directory, the same error occurs in each program
"Cannot find DATA_PATH variable in the environment. DATA_PATH should be the folder that contains
kin40k/
directory with the data. Please export DATA_PATH before loading the data "
i cannot download any data from the example, could you tell where and how can i get those needed data?
from keras-gp.
@XYLee01, examples that use sysid data should run (they automatically retrieve and preprocess the data). If you want to run examples on kin40k
, the data loader assumes that you already have that dataset (and DATA_PATH
should correctly point to it). You can download kin40k
data from Andrew Wilson's page (also, follow the instructions there on how to preprocess the dataset).
Of course, it would be nice to have an automatic data loader for kin40k
that takes care of all preprocessing. I very much welcome contributions!
from keras-gp.
Related Issues (20)
- Recurrent structure HOT 2
- Gaussian Process as non-final layer HOT 2
- Training GP on outputs from the last hidden layer HOT 3
- Inference not Converging msgp_mlp_kin40k.py HOT 1
- Octave returned: error: Two few unique points. HOT 4
- Training Error (Model object has no attribute '_check_num_samples') HOT 1
- Missing actuator.mat file
- Grid Creation HOT 2
- Save and load trained keras-gp model
- missing startup.m in the gpml in the backend? HOT 1
- P-Dimensional GP input shape HOT 1
- How can I get one standard deviation of the predictive distributions? HOT 2
- A strange phenomenon when gp-lstm is applied to a simulated dataset HOT 2
- Code for Pytorch HOT 1
- A question about hyperparameters
- A question about hyperparameters 'd' of Materniso kernel
- ImportError: cannot import name '_to_list' from 'keras.engine.topology' HOT 4
- Dataset missing for actuator.mat HOT 1
- Statistics package not installed HOT 1
- TensorFlow error when running gp_lstm_actuator.py HOT 1
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 keras-gp.