diff --git a/docker/ml-gpu.yml b/docker/ml-gpu.yml
index 3f90d1d5da..5a3bfbda5e 100644
--- a/docker/ml-gpu.yml
+++ b/docker/ml-gpu.yml
@@ -22,7 +22,7 @@ services:
     # https://github.com/triton-inference-server/server/blob/main/docs/user_guide/model_management.md
     # for more information
     entrypoint: "/opt/nvidia/nvidia_entrypoint.sh tritonserver --model-repository=/models --model-control-mode=explicit --load-model=*"
-    mem_limit: 15g
+    mem_limit: 20g
     runtime: nvidia
     deploy:
       resources:
diff --git a/docker/ml.yml b/docker/ml.yml
index b7dca98531..c348138b70 100644
--- a/docker/ml.yml
+++ b/docker/ml.yml
@@ -14,7 +14,7 @@ services:
     # https://github.com/triton-inference-server/server/blob/main/docs/user_guide/model_management.md
     # for more information
     entrypoint: "tritonserver --model-repository=/models --model-control-mode=explicit --load-model=*"
-    mem_limit: 15g
+    mem_limit: 20g
 
   fasttext:
     restart: $RESTART_POLICY