diff --git a/requirements-cuda-legacy.txt b/requirements-cuda-legacy.txt
index c0e849699..20b579be9 100644
--- a/requirements-cuda-legacy.txt
+++ b/requirements-cuda-legacy.txt
@@ -4,11 +4,11 @@
 
 # pytorch
 --extra-index-url https://download.pytorch.org/whl/cu126
-torch==2.12.0+cu126
-torchvision==0.27.0+cu126
+torch==2.13.0+cu126
+torchvision==0.28.0+cu126
 onnxruntime-gpu==1.23.2
 nvidia-nccl-cu12==2.29.3; sys_platform == "linux"
-triton-windows==3.7.0.post26; sys_platform == "win32"
+triton-windows==3.7.1.post27; sys_platform == "win32"
 
 # optimizers
 bitsandbytes==0.49.1 # bitsandbytes for 8-bit optimizers and weight quantization
diff --git a/requirements-cuda.txt b/requirements-cuda.txt
index 5297caf00..b37b87515 100644
--- a/requirements-cuda.txt
+++ b/requirements-cuda.txt
@@ -1,10 +1,10 @@
 # pytorch
 --extra-index-url https://download.pytorch.org/whl/cu130
-torch==2.12.0+cu130
-torchvision==0.27.0+cu130
+torch==2.13.0+cu130
+torchvision==0.28.0+cu130
 onnxruntime-gpu==1.23.2
 nvidia-nccl-cu13==2.29.7; sys_platform == "linux"
-triton-windows==3.7.0.post26; sys_platform == "win32"
+triton-windows==3.7.1.post27; sys_platform == "win32"
 
 # optimizers
 bitsandbytes==0.49.1 # bitsandbytes for 8-bit optimizers and weight quantization
diff --git a/requirements-default.txt b/requirements-default.txt
index 0bf5ee76d..187471362 100644
--- a/requirements-default.txt
+++ b/requirements-default.txt
@@ -1,6 +1,6 @@
 # pytorch
-torch==2.12.0
-torchvision==0.27.0
+torch==2.13.0
+torchvision==0.28.0
 onnxruntime==1.23.2
 
 # optimizers
diff --git a/requirements-rocm.txt b/requirements-rocm.txt
index 116ccd444..a5ee19b9d 100644
--- a/requirements-rocm.txt
+++ b/requirements-rocm.txt
@@ -3,8 +3,8 @@
 
 # pytorch
 --extra-index-url https://download.pytorch.org/whl/rocm7.2
-torch==2.12.0+rocm7.2; sys_platform == "linux"
-torchvision==0.27.0+rocm7.2; sys_platform == "linux"
+torch==2.13.0+rocm7.2; sys_platform == "linux"
+torchvision==0.28.0+rocm7.2; sys_platform == "linux"
 onnxruntime==1.23.2
 
 # optimizers