pulsatrix 1.1.0
Loading...
Searching...
No Matches
grad_clipping.hpp
Go to the documentation of this file.
1
5#pragma once
6
8
9namespace pulsatrix {
10
27[[nodiscard]] float ClipGradNorm(Module& module, float max_norm, bool error_if_nonfinite = false);
28
29} // namespace pulsatrix
Base class for every layer type (LinearModule, Conv2DModule, activations, ...).
Definition module.hpp:78
Abstract base every layer subclasses – NVI forward(), pure-virtual LRP contract.
Definition acquisition_functions.hpp:16
float ClipGradNorm(Module &module, float max_norm, bool error_if_nonfinite=false)
Scales every trainable parameter's gradient so their global L2 norm is at most max_norm,...