diff --git a/dockerfiles/hermes-local-image-server.py b/dockerfiles/hermes-local-image-server.py
index d9555ec24..868907479 100644
--- a/dockerfiles/hermes-local-image-server.py
+++ b/dockerfiles/hermes-local-image-server.py
@@ -289,8 +289,6 @@ def _render_in_process(payload: dict[str, Any]) -> dict[str, Any]:
             pipe.enable_model_cpu_offload()
         else:
             raise ValueError("HERMES_LOCAL_IMAGE_OFFLOAD_MODE must be sequential or model")
-        pipe.enable_vae_slicing()
-        pipe.enable_vae_tiling()
         _set_state(phase="rendering")
         arguments: dict[str, Any] = {
             "prompt": prompt,
diff --git a/services/hermes/local-image-deployment.yaml b/services/hermes/local-image-deployment.yaml
index 1167eee95..3dc69e2d3 100644
--- a/services/hermes/local-image-deployment.yaml
+++ b/services/hermes/local-image-deployment.yaml
@@ -43,7 +43,7 @@ spec:
             sizeLimit: 2Gi
       containers:
         - name: local-image
-          image: registry.bstein.dev/bstein/hermes-local-image@sha256:a937f02fe1e85df9d86c0edb1972de6c020fd8ef4f768809624cbebdc521bca1
+          image: registry.bstein.dev/bstein/hermes-local-image@sha256:d6257f49a60244fb1e30733848a187fa932acebfdb82a2560d4e760fa018b645
           imagePullPolicy: IfNotPresent
           ports:
             - name: local-image
diff --git a/testing/tests/test_hermes_chat_quality.py b/testing/tests/test_hermes_chat_quality.py
index 3d1a7d885..4cfdb360e 100644
--- a/testing/tests/test_hermes_chat_quality.py
+++ b/testing/tests/test_hermes_chat_quality.py
@@ -767,8 +767,6 @@ def test_local_flux_uses_low_vram_offload_without_reducing_resolution():
     assert 'OFFLOAD_MODE = os.environ.get(' in source
     assert '"HERMES_LOCAL_IMAGE_OFFLOAD_MODE", "sequential"' in source
     assert "pipe.enable_sequential_cpu_offload()" in source
-    assert "pipe.enable_vae_slicing()" in source
-    assert "pipe.enable_vae_tiling()" in source
     assert '"square": (1024, 1024)' in source