diff --git a/CHANGELOG.md b/CHANGELOG.md
index fd9bd8a5..35a06b60 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -61,6 +61,9 @@
     scipy sparse matrices in scipy 1.14.
 * Fix `methods/geneformer` erroring with a 404 on every dictionary and model file: upstream reorganised the repo for
     Geneformer V2 and dropped the gc95M assets from `main`, so pin the downloads to the last revision that has them.
+* Give `methods/fadvi` a `gpu` label. Without one it was scheduled on the CPU partition with no GPU attached, so it
+    trained on the CPU and hit its walltime on every dataset.
+* Give `methods/scalex` a `midcpu` label instead of `lowcpu`. It is CPU-bound only because its engine has no CUDA.
 
 * Split Scanorama into two methods/scores
     - Split Scanorama into embedding (integrate) and count-correction (correct) modes, instead of running both together. 
diff --git a/src/methods/fadvi/config.vsh.yaml b/src/methods/fadvi/config.vsh.yaml
index 80a8e7e0..dcfaa34c 100644
--- a/src/methods/fadvi/config.vsh.yaml
+++ b/src/methods/fadvi/config.vsh.yaml
@@ -97,4 +97,4 @@ runners:
   # Allows turning the component into a Nextflow module / pipeline.
   - type: nextflow
     directives:
-      label: [midtime,midmem,midcpu]
+      label: [hightime, midmem, midcpu, gpu]
diff --git a/src/methods/scalex/config.vsh.yaml b/src/methods/scalex/config.vsh.yaml
index 59c1aed4..fb890d6b 100644
--- a/src/methods/scalex/config.vsh.yaml
+++ b/src/methods/scalex/config.vsh.yaml
@@ -33,4 +33,4 @@ runners:
   - type: executable
   - type: nextflow
     directives:
-      label: [lowmem, lowcpu, hightime]
+      label: [lowmem, midcpu, hightime]