Hi Gabriel,
It seems that Cloudy tells you that you are not using one of its own grains files. I'm not sure why it mentions the graphite while you are using a midified silicate. Is this the only grain file you are using in this model? Is the file you compile still have some ghost of aprevious data in it that could be graphite based?
This is actually more a Cloudy question than a pyCloudy one ;-) You could ask in the Cloudy group: https://cloudyastrophysics.groups.io/
Hope it helps,
Christophe