yancy lindquist

Hi,

Just for fun I am seeing if I can train a model to predict the next step in a smoke sim. I figured if that would work, I could put it in a solver sop, and have a ml smoke solver. It works surprisingly well for predicting the first next step, but when I feed the result back to the MlRegressionInference the second time, nothing happens.

Maybe someone out there can tell me the fix? I have attached my file.

I have installed and am using tensorflow in houdini. I am calling it in python nodes in tops and lops. It is working, but I am getting warnings that it will only use my cpu. I want to reinstall it to try to get a fix, and I had a few questions that maybe someone can help me with.

-Is there an official set of steps to install tensorflow for houdini?

-Would the "Python Virtual Environment" node in tops expose tensorflow to the other contexts like lops?

Thanks for any help.
Yancy

Hi,

Maybe I am doing something wrong, but when I import either a single float or double array usd attribute from lops using usdImportData, all my floats are rounded down to ints.

ex. 3.67 comes in as 3.

I can see them in the usd with decimal places. When I write them to a cvs spreadsheet, it confirms the issue. Image attached.


Any thoughts?

Cheers,
Yancy