Hi, I've been running an experiment in which I'm testing various preprocessing methods, so I'm sort of looping around creating new datasets and networks.
To build the networks, I'm currently using Bayesian search, and I hit an error after running my script for a while.
```
File "/ms/aircraft_package/model/_model.py", line 42, in fit
net = bay_search.learn(ds_train)
ValueError: vector::reserve
```
The bay_search is an object of class pysmile.learning.BayesianSearch().
I think I might be running low on memory with my dataset, and this is why it is happening. It works fine and finishes when I'm running it on a subset of data. But I'm not sure if its just my laptop not having enough memory or there are any other problems.
ValueError: vector::reserve
-
shooltz[BayesFusion]
- Site Admin
- Posts: 1399
- Joined: Mon Nov 26, 2007 5:51 pm
Re: ValueError: vector::reserve
The error is caused by C++ exception caught by code binding between core SMILE library and PySMILE. It's a memory issue. Are you running 32-bit or 64-bit Python?
Re: ValueError: vector::reserve
Python 64-bit version. I'm pretty sure that I'm the reason why this error happened.
I was playing around with setting a lot of background knowledge on a very large network. Sometimes when the networks managed to finish training and save, their size varied from 100 mb to up to 2 gb. So I guess it had to store multiple of such networks like this in the memory, and this is what caused the error. I'm just gonna give up on this idea as I had way too many forced connection from BK.
I was playing around with setting a lot of background knowledge on a very large network. Sometimes when the networks managed to finish training and save, their size varied from 100 mb to up to 2 gb. So I guess it had to store multiple of such networks like this in the memory, and this is what caused the error. I'm just gonna give up on this idea as I had way too many forced connection from BK.