diff --git a/python/_custom_build/backend.py b/python/_custom_build/backend.py
index 01745fcfa8..c58d3f84a7 100644
--- a/python/_custom_build/backend.py
+++ b/python/_custom_build/backend.py
@@ -21,8 +21,8 @@ def wrapper(config_settings=None):
         orig_list = getattr(_orig, func.__name__)(config_settings)
         cuda_suffix = os.getenv('RAPIDS_PY_WHEEL_CUDA_SUFFIX', default='')
         append_list = [
-            f"rmm{cuda_suffix}==23.2.*",
-            f"pylibraft{cuda_suffix}==23.2.*",
+            f"rmm{cuda_suffix}==23.4.*",
+            f"pylibraft{cuda_suffix}==23.4.*",
         ]
         return orig_list + append_list
 
diff --git a/python/setup.py b/python/setup.py
index d204f9e46c..9f11396726 100644
--- a/python/setup.py
+++ b/python/setup.py
@@ -127,7 +127,7 @@ def get_versions():
       description="cuML - RAPIDS ML Algorithms",
       url="https://github.com/rapidsai/cuml",
       author="NVIDIA Corporation",
-      license="Apache",
+      license="Apache 2.0",
       classifiers=[
           "Intended Audience :: Developers",
           "Programming Language :: Python",
@@ -147,10 +147,10 @@ def get_versions():
         "seaborn",
         "treelite==3.1.0",
         "treelite_runtime==3.1.0",
-        f"cudf{cuda_suffix}==23.2.*",
-        f"dask-cudf{cuda_suffix}==23.2.*",
-        f"pylibraft{cuda_suffix}==23.2.*",
-        f"raft-dask{cuda_suffix}==23.2.*",
+        f"cudf{cuda_suffix}==23.4.*",
+        f"dask-cudf{cuda_suffix}==23.4.*",
+        f"pylibraft{cuda_suffix}==23.4.*",
+        f"raft-dask{cuda_suffix}==23.4.*",
       ],
       extras_require={
           "test": [