diff --git a/examples/benchmarks/basic.sh b/examples/benchmarks/basic.sh
index 6b043c567..c5592970f 100644
--- a/examples/benchmarks/basic.sh
+++ b/examples/benchmarks/basic.sh
@@ -14,7 +14,7 @@ do
     echo "Running $SCENE"
 
     # train without eval
-    CUDA_VISIBLE_DEVICES=0 python simple_trainer.py default --eval_steps -1 --disable_viewer --data_factor $DATA_FACTOR \
+    CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py default --eval_steps -1 --disable_viewer --data_factor $DATA_FACTOR \
         --render_traj_path $RENDER_TRAJ_PATH \
         --data_dir data/360_v2/$SCENE/ \
         --result_dir $RESULT_DIR/$SCENE/
@@ -22,7 +22,7 @@ do
     # run eval and render
     for CKPT in $RESULT_DIR/$SCENE/ckpts/*;
     do
-        CUDA_VISIBLE_DEVICES=0 python simple_trainer.py default --disable_viewer --data_factor $DATA_FACTOR \
+        CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py default --disable_viewer --data_factor $DATA_FACTOR \
             --render_traj_path $RENDER_TRAJ_PATH \
             --data_dir data/360_v2/$SCENE/ \
             --result_dir $RESULT_DIR/$SCENE/ \
diff --git a/examples/benchmarks/basic_2dgs.sh b/examples/benchmarks/basic_2dgs.sh
index 04d3d8fd4..f683d4f95 100755
--- a/examples/benchmarks/basic_2dgs.sh
+++ b/examples/benchmarks/basic_2dgs.sh
@@ -13,7 +13,7 @@ do
     echo "Running $SCENE"
 
     # train without eval
-    CUDA_VISIBLE_DEVICES=0 python simple_trainer_2dgs.py --eval_steps -1 --disable_viewer --data_factor $DATA_FACTOR \
+    CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer_2dgs.py --eval_steps -1 --disable_viewer --data_factor $DATA_FACTOR \
         --model_type 2dgs \
         --data_dir data/360_v2/$SCENE/ \
         --result_dir $RESULT_DIR/$SCENE/
@@ -21,7 +21,7 @@ do
     # run eval and render
     for CKPT in $RESULT_DIR/$SCENE/ckpts/*;
     do
-        CUDA_VISIBLE_DEVICES=0 python simple_trainer_2dgs.py --disable_viewer --data_factor $DATA_FACTOR \
+        CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer_2dgs.py --disable_viewer --data_factor $DATA_FACTOR \
             --model_type 2dgs \
             --data_dir data/360_v2/$SCENE/ \
             --result_dir $RESULT_DIR/$SCENE/ \
diff --git a/examples/benchmarks/basic_4gpus.sh b/examples/benchmarks/basic_4gpus.sh
index 01b0c2c18..809905256 100644
--- a/examples/benchmarks/basic_4gpus.sh
+++ b/examples/benchmarks/basic_4gpus.sh
@@ -13,7 +13,7 @@ do
     echo "Running $SCENE"
 
     # train and eval at the last step (30000)
-    CUDA_VISIBLE_DEVICES=0,1,2,3 python simple_trainer.py default --eval_steps 30000 --disable_viewer --data_factor $DATA_FACTOR \
+    CUDA_VISIBLE_DEVICES=0,1,2,3 python examples/simple_trainer.py default --eval_steps 30000 --disable_viewer --data_factor $DATA_FACTOR \
         # 4 GPUs is effectively 4x batch size so we scale down the steps by 4x as well.
         # "--packed" reduces the data transfer between GPUs, which leads to faster training. 
         --steps_scaler 0.25 --packed \
diff --git a/examples/benchmarks/bilarf/mcmc_bilarf.sh b/examples/benchmarks/bilarf/mcmc_bilarf.sh
index 044628032..819707ea0 100644
--- a/examples/benchmarks/bilarf/mcmc_bilarf.sh
+++ b/examples/benchmarks/bilarf/mcmc_bilarf.sh
@@ -13,7 +13,7 @@ do
     echo "Running $SCENE"
 
     # train without eval
-    CUDA_VISIBLE_DEVICES=0 python simple_trainer.py mcmc --disable_viewer --data_factor $DATA_FACTOR \
+    CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py mcmc --disable_viewer --data_factor $DATA_FACTOR \
         --use_bilateral_grid \
         --render_traj_path $RENDER_TRAJ_PATH \
         --data_dir $SCENE_DIR/$SCENE/ \
diff --git a/examples/benchmarks/compression/mcmc.sh b/examples/benchmarks/compression/mcmc.sh
index 4c7165f3d..bc30808b4 100644
--- a/examples/benchmarks/compression/mcmc.sh
+++ b/examples/benchmarks/compression/mcmc.sh
@@ -30,13 +30,13 @@ do
     echo "Running $SCENE"
 
     # train without eval
-    CUDA_VISIBLE_DEVICES=0 python simple_trainer.py mcmc --eval_steps -1 --disable_viewer --data_factor $DATA_FACTOR \
+    CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py mcmc --eval_steps -1 --disable_viewer --data_factor $DATA_FACTOR \
         --strategy.cap-max $CAP_MAX \
         --data_dir data/360_v2/$SCENE/ \
         --result_dir $RESULT_DIR/$SCENE/
 
     # eval: use vgg for lpips to align with other benchmarks
-    CUDA_VISIBLE_DEVICES=0 python simple_trainer.py mcmc --disable_viewer --data_factor $DATA_FACTOR \
+    CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py mcmc --disable_viewer --data_factor $DATA_FACTOR \
         --strategy.cap-max $CAP_MAX \
         --data_dir data/360_v2/$SCENE/ \
         --result_dir $RESULT_DIR/$SCENE/ \
diff --git a/examples/benchmarks/compression/mcmc_tt.sh b/examples/benchmarks/compression/mcmc_tt.sh
index 054920929..0ff2eff24 100644
--- a/examples/benchmarks/compression/mcmc_tt.sh
+++ b/examples/benchmarks/compression/mcmc_tt.sh
@@ -23,13 +23,13 @@ do
     echo "Running $SCENE"
 
     # train without eval
-    CUDA_VISIBLE_DEVICES=0 python simple_trainer.py mcmc --eval_steps -1 --disable_viewer --data_factor 1 \
+    CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py mcmc --eval_steps -1 --disable_viewer --data_factor 1 \
         --strategy.cap-max $CAP_MAX \
         --data_dir $SCENE_DIR/$SCENE/ \
         --result_dir $RESULT_DIR/$SCENE/
 
     # eval: use vgg for lpips to align with other benchmarks
-    CUDA_VISIBLE_DEVICES=0 python simple_trainer.py mcmc --disable_viewer --data_factor 1 \
+    CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py mcmc --disable_viewer --data_factor 1 \
         --strategy.cap-max $CAP_MAX \
         --data_dir $SCENE_DIR/$SCENE/ \
         --result_dir $RESULT_DIR/$SCENE/ \
diff --git a/examples/benchmarks/fisheye/mcmc_zipnerf.sh b/examples/benchmarks/fisheye/mcmc_zipnerf.sh
index bd0e57f44..6290df9b5 100644
--- a/examples/benchmarks/fisheye/mcmc_zipnerf.sh
+++ b/examples/benchmarks/fisheye/mcmc_zipnerf.sh
@@ -14,7 +14,7 @@ do
     echo "Running $SCENE"
 
     # train and eval
-    CUDA_VISIBLE_DEVICES=0 python simple_trainer.py mcmc --disable_viewer --data_factor $DATA_FACTOR \
+    CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py mcmc --disable_viewer --data_factor $DATA_FACTOR \
         --strategy.cap-max $CAP_MAX \
         --opacity_reg 0.001 \
         --use_bilateral_grid \
diff --git a/examples/benchmarks/fisheye/mcmc_zipnerf_undistorted.sh b/examples/benchmarks/fisheye/mcmc_zipnerf_undistorted.sh
index 012c685cf..812845a30 100644
--- a/examples/benchmarks/fisheye/mcmc_zipnerf_undistorted.sh
+++ b/examples/benchmarks/fisheye/mcmc_zipnerf_undistorted.sh
@@ -14,7 +14,7 @@ do
     echo "Running $SCENE"
 
     # train and eval
-    CUDA_VISIBLE_DEVICES=0 python simple_trainer.py mcmc --disable_viewer --data_factor $DATA_FACTOR \
+    CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py mcmc --disable_viewer --data_factor $DATA_FACTOR \
         --strategy.cap-max $CAP_MAX \
         --opacity_reg 0.001 \
         --use_bilateral_grid \
diff --git a/examples/benchmarks/mcmc.sh b/examples/benchmarks/mcmc.sh
index 0eaa5c8bb..dd9ebac09 100644
--- a/examples/benchmarks/mcmc.sh
+++ b/examples/benchmarks/mcmc.sh
@@ -16,7 +16,7 @@ do
     echo "Running $SCENE"
 
     # train without eval
-    CUDA_VISIBLE_DEVICES=0 python simple_trainer.py mcmc --eval_steps -1 --disable_viewer --data_factor $DATA_FACTOR \
+    CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py mcmc --eval_steps -1 --disable_viewer --data_factor $DATA_FACTOR \
         --strategy.cap-max $CAP_MAX \
         --render_traj_path $RENDER_TRAJ_PATH \
         --data_dir $SCENE_DIR/$SCENE/ \
@@ -25,7 +25,7 @@ do
     # run eval and render
     for CKPT in $RESULT_DIR/$SCENE/ckpts/*;
     do
-        CUDA_VISIBLE_DEVICES=0 python simple_trainer.py mcmc --disable_viewer --data_factor $DATA_FACTOR \
+        CUDA_VISIBLE_DEVICES=0 python examples/simple_trainer.py mcmc --disable_viewer --data_factor $DATA_FACTOR \
             --strategy.cap-max $CAP_MAX \
             --render_traj_path $RENDER_TRAJ_PATH \
             --data_dir $SCENE_DIR/$SCENE/ \