Class DenseLinearRelease
java.lang.Object
smile.llm.quant.DenseLinearRelease
Releases placeholder dense
LinearLayer shells after quantized ops
are installed, so empty GPU weight storage does not inflate the KV budget.-
Method Summary
Modifier and TypeMethodDescriptionstatic voidunregisterAndClose(MemorySegment parentModule, String name, LinearOp op) UnregistersnamefromparentModule(if present) and closes the dense layer so LibTorch can free its parameter storage.
-
Method Details
-
unregisterAndClose
UnregistersnamefromparentModule(if present) and closes the dense layer so LibTorch can free its parameter storage.- Parameters:
parentModule- parent PyTorch module handle, ornull.name- registered submodule name, ornull.op- linear op to close when it is aLinearLayer.
-