diff --git a/.github/workflows/examples.yaml b/.github/workflows/examples.yaml
index 1c9af43..1c42367 100644
--- a/.github/workflows/examples.yaml
+++ b/.github/workflows/examples.yaml
@@ -22,16 +22,16 @@ jobs:
         - name: Run default example on CPU.
           run: |
             source .venv/bin/activate
-            CUDA_VISIBLE_DEVICES="" python -m distributed_shampoo.examples.default_cifar10_example --optimizer-type DISTRIBUTED_SHAMPOO --precondition-frequency 30 --grafting-type ADAM --use-bias-correction --use-decoupled-weight-decay --use-merge-dims --epochs 1 --batch-size 1024
+            CUDA_VISIBLE_DEVICES="" python -m distributed_shampoo.examples.default_cifar10_example --optimizer-type DISTRIBUTED_SHAMPOO --precondition-frequency 30 --grafting-type ADAM --use-bias-correction --use-decoupled-weight-decay --use-merge-dims --epochs 1 --batch-size 1024 --momentum=0.9
         - name: Run default example on GPU.
           run: |
             source .venv/bin/activate
-            python -m distributed_shampoo.examples.default_cifar10_example --optimizer-type DISTRIBUTED_SHAMPOO --precondition-frequency 30 --grafting-type ADAM --use-bias-correction --use-decoupled-weight-decay --use-merge-dims --epochs 1 --batch-size 1024
+            python -m distributed_shampoo.examples.default_cifar10_example --optimizer-type DISTRIBUTED_SHAMPOO --precondition-frequency 30 --grafting-type ADAM --use-bias-correction --use-decoupled-weight-decay --use-merge-dims --epochs 1 --batch-size 1024 --momentum=0.9
         - name: Run DDP example on CPU.
           run: |
             source .venv/bin/activate
-            CUDA_VISIBLE_DEVICES="" torchrun --standalone --nnodes=1 --nproc_per_node=2 -m distributed_shampoo.examples.ddp_cifar10_example --optimizer-type DISTRIBUTED_SHAMPOO --precondition-frequency 15 --grafting-type ADAM --use-bias-correction --use-decoupled-weight-decay --use-merge-dims --epochs 1 --local-batch-size 1024 --backend gloo
+            CUDA_VISIBLE_DEVICES="" torchrun --standalone --nnodes=1 --nproc_per_node=2 -m distributed_shampoo.examples.ddp_cifar10_example --optimizer-type DISTRIBUTED_SHAMPOO --precondition-frequency 15 --grafting-type ADAM --use-bias-correction --use-decoupled-weight-decay --use-merge-dims --epochs 1 --local-batch-size 1024 --backend gloo --momentum=0.9
         - name: Run DDP example on GPU.
           run: |
             source .venv/bin/activate
-            torchrun --standalone --nnodes=1 --nproc_per_node=1 -m distributed_shampoo.examples.ddp_cifar10_example --optimizer-type DISTRIBUTED_SHAMPOO --precondition-frequency 30 --grafting-type ADAM --use-bias-correction --use-decoupled-weight-decay --use-merge-dims --epochs 1 --local-batch-size 1024
+            torchrun --standalone --nnodes=1 --nproc_per_node=1 -m distributed_shampoo.examples.ddp_cifar10_example --optimizer-type DISTRIBUTED_SHAMPOO --precondition-frequency 30 --grafting-type ADAM --use-bias-correction --use-decoupled-weight-decay --use-merge-dims --epochs 1 --local-batch-size 1024 --momentum=0.9