From dcadf5141103e5f73e582426afac2f120aac4e30 Mon Sep 17 00:00:00 2001 From: Ikko Eltociear Ashimine Date: Mon, 6 May 2024 00:52:08 +0900 Subject: [PATCH] Update README.md --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 613bb82..ad54738 100644 --- a/README.md +++ b/README.md @@ -20,4 +20,4 @@ It needs more experiments. 2024-05-04 Update: @xiaol hinted that the constant initialization of `base_weight` parameters can be a problem on MNIST. For now I've changed both the `base_weight` and `spline_scaler` matrices to be initialized with `kaiming_uniform_`, following `nn.Linear`'s initialization. -It seems to work much much better on MNIST (~20% to ~97%), but I'm not sure if it's a good idea in general. +It seems to work much better on MNIST (~20% to ~97%), but I'm not sure if it's a good idea in general.