Resource limit reached

New comment
 

Hello,

I've built a framework to train models and export them to ONNX files. However, I'm running into a resource limit error during compilation.
It seems that the limit is set to 128 MB, which is quite low for a medium sized model.

Has anyone found a way to get around this limitation?

Thanks a lot!


 
Thibauld Charles Ghislain Robin:

I've built a framework to train models and export them to ONNX files. However, I'm running into a resource limit error during compilation.

It seems that the limit is set to 128 MB, which is quite low for a medium sized model.

Has anyone found a way to get around this limitation?

Use files in the common folder (external for sources/program), or split large files into multiple resources.
 
Stanislav Korotky #:
Use files in the common folder (external for sources/program)...
How is the common folder related to a resource size limitation ?
 
Alain Verleyen #:
How is the common folder related to a resource size limitation ?
He can load the model directly from a file (without the size limit), not from an embedded resource. This also will allow to update the model without recompilation.
 
Stanislav Korotky #:
He can load the model directly from a file (without the size limit), not from an embedded resource. This also will allow to update the model without recompilation.
Ok but this is not specific to using the common folder, he could do the same from MQL5\Files.
 
Alain Verleyen #:
Ok but this is not specific to using the common folder, he could do the same from MQL5\Files.
The common folder is much more convenient in the sense, that it seamlessly supports testing and optimization on the local agents. A specific instance's folder (MQL5/Files) would require to tweak the source code a bit (back and forth(!)) in order to support functionality in the tester.
New comment