diff --git a/pyproject.toml b/pyproject.toml
index 46c7fc4602..74e1ccc190 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -37,7 +37,7 @@ triton = [
 ]
 
 huggingface = [
-    "unsloth_zoo>=2025.6.7",
+    "unsloth_zoo>=2025.6.8",
     "packaging",
     "tyro",
     "transformers>=4.51.3,!=4.47.0,!=4.52.0,!=4.52.1,!=4.52.2,!=4.52.3",
@@ -381,7 +381,7 @@ colab-ampere-torch220 = [
     "flash-attn>=2.6.3",
 ]
 colab-new = [
-    "unsloth_zoo>=2025.6.7",
+    "unsloth_zoo>=2025.6.8",
     "packaging",
     "tyro",
     "transformers>=4.51.3,!=4.47.0,!=4.52.0,!=4.52.1,!=4.52.2,!=4.52.3",
diff --git a/unsloth/models/vision.py b/unsloth/models/vision.py
index 39ff1fd2bb..53c5497424 100644
--- a/unsloth/models/vision.py
+++ b/unsloth/models/vision.py
@@ -440,12 +440,12 @@ class FastBaseModel:
             for jj, (name, module) in enumerate(model.named_modules()):
                 exec(custom_datatype)
             pass
-            # Clear deleted GPU items
-            for _ in range(3):
-                gc.collect()
-                if DEVICE_TYPE == "cuda":  torch.cuda.empty_cache()
-                elif DEVICE_TYPE == "xpu": torch.xpu.empty_cache()
-            pass
+        pass
+        # Clear deleted GPU items
+        for _ in range(3):
+            gc.collect()
+            if DEVICE_TYPE == "cuda":  torch.cuda.empty_cache()
+            elif DEVICE_TYPE == "xpu": torch.xpu.empty_cache()
         pass
 
         # Counteract saved tokenizers
@@ -562,7 +562,7 @@ class FastBaseModel:
         finetune_mlp_modules       = True,
         layers_to_transform        = None,
         layers_pattern             = None,
-        use_gradient_checkpointing = True,
+        use_gradient_checkpointing = "unsloth",
         random_state               = 3407,
         max_seq_length             = 2048, # not used anymore
         use_rslora                 = False,