From 7372db87899c5dbeb3f04f8c7d1500c0897f2c23 Mon Sep 17 00:00:00 2001 From: Fabian Isensee Date: Tue, 21 May 2024 12:10:10 +0200 Subject: [PATCH] fix #2110 --- documentation/resenc_presets.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/documentation/resenc_presets.md b/documentation/resenc_presets.md index 4c27a61a4..7acbe221d 100644 --- a/documentation/resenc_presets.md +++ b/documentation/resenc_presets.md @@ -72,7 +72,7 @@ The presets differ from `ResEncUNetPlanner` in two ways: - They set new default values for `gpu_memory_target_in_gb` to target the respective VRAM consumptions - They remove the batch size cap of 0.05 (= previously one batch could not cover mode pixels than 5% of the entire dataset, not it can be arbitrarily large) -The preset are merely there ot make life easier, and to provide standardized configurations people can benchmark with. +The presets are merely there to make life easier, and to provide standardized configurations people can benchmark with. You can easily adapt the GPU memory target to match your GPU, and to scale beyond 40GB of GPU memory. Here is an example for how to scale to 80GB VRAM on Dataset003_Liver: