diff --git a/TPTBox/core/nii_wrapper.py b/TPTBox/core/nii_wrapper.py
index 89cd685..93c62fa 100755
--- a/TPTBox/core/nii_wrapper.py
+++ b/TPTBox/core/nii_wrapper.py
@@ -667,7 +667,7 @@ def astype(self,dtype,order:Literal["C","F","A","K"] ='K', casting:Literal["no",
             return c
         else:
             return self.get_array().astype(dtype,order=order,casting=casting, subok=subok,copy=copy)
-    def reorient(self:Self, axcodes_to: AX_CODES|None = ("P", "I", "R"), verbose:logging=False, inplace=False)-> Self:
+    def reorient(self:Self, axcodes_to: AX_CODES|str|None = ("P", "I", "R"), verbose:logging=False, inplace=False)-> Self:
         """
         Reorients the input Nifti image to the desired orientation, specified by the axis codes.
 
@@ -1863,7 +1863,7 @@ def save(self,file:str|Path,make_parents=True,verbose:logging=True, dtype = None
             self.set_dtype_("smallest_uint")
             arr = self.get_array() if not self.seg else self.get_seg_array()
 
-
+        self.header.set_data_dtype(arr.dtype)
         out = Nifti1Image(arr, self.affine,self.header)#,dtype=arr.dtype)
         if dtype is not None:
             out.set_data_dtype(dtype)
@@ -1929,7 +1929,7 @@ def save_nrrd(self:Self, file: str | Path|bids_files.BIDS_FILE,make_parents=True
         # Save NRRD file
 
         log.print(f"Saveing {file}",verbose=verbose,ltype=Log_Type.SAVE,end='\r')
-        nrrd.write(file, data=data, header=header,**args)
+        nrrd.write(str(file), data=data, header=header,**args) # nrrd only acepts strings...
         log.print(f"Save {file} as {header['type']}",verbose=verbose,ltype=Log_Type.SAVE)
 
     def __str__(self) -> str:
diff --git a/TPTBox/segmentation/VibeSeg/inference_nnunet.py b/TPTBox/segmentation/VibeSeg/inference_nnunet.py
index 889c4e4..fac4ec6 100644
--- a/TPTBox/segmentation/VibeSeg/inference_nnunet.py
+++ b/TPTBox/segmentation/VibeSeg/inference_nnunet.py
@@ -72,7 +72,10 @@ def run_inference_on_file(
     if out_file is not None and Path(out_file).exists() and not override:
         return out_file, None
 
-    from TPTBox.segmentation.nnUnet_utils.inference_api import load_inf_model, run_inference
+    from TPTBox.segmentation.nnUnet_utils.inference_api import (
+        load_inf_model,
+        run_inference,
+    )
 
     if isinstance(idx, int):
         download_weights(idx, model_path)
@@ -102,7 +105,7 @@ def run_inference_on_file(
             if "model_expected_orientation" in ds_info2:
                 ds_info["orientation"] = ds_info2["model_expected_orientation"]
             if "resolution_range" in ds_info2:
-                ds_info["spacing"] = ds_info2["resolution_range"]
+                ds_info["resolution_range"] = ds_info2["resolution_range"]
 
     nnunet = load_inf_model(
         nnunet_path,
@@ -127,7 +130,10 @@ def run_inference_on_file(
 
     try:
         zoom_old = ds_info.get("spacing")
+        if idx not in [527] and zoom_old is not None:
+            zoom_old = zoom_old[::-1]
 
+        zoom_old = ds_info.get("resolution_range", zoom_old)
         if zoom_old is None:
             zoom = plans_info["configurations"]["3d_fullres"]["spacing"]
             if all(zoom[0] == z for z in zoom):
diff --git a/TPTBox/segmentation/nnUnet_utils/predictor.py b/TPTBox/segmentation/nnUnet_utils/predictor.py
index be37fb4..a4f2314 100755
--- a/TPTBox/segmentation/nnUnet_utils/predictor.py
+++ b/TPTBox/segmentation/nnUnet_utils/predictor.py
@@ -523,9 +523,8 @@ def check_mem(shape):
                         j = np.argmax(s)
                         if s[j] == 1:
                             device = "cpu"
-                            print("Fall Back CPU. Not enough space", shape, patch_size, splits, s)
+                            print("Fall Back CPU. Not enough space; s[j] == 1", shape, patch_size, splits, s)
                             break
-                        splits[j] += 1
                         shape_split = [ceil(s / sp) for s, sp in zip(shape, splits)]
                         # print(shape, patch_size, splits, s, np.prod(shape) / 1000000)
                         if check_mem(shape_split):
@@ -543,6 +542,7 @@ def check_mem(shape):
                                 print(e)
                                 break
 
+                        splits[j] += 1
                 predicted_logits, n_predictions = self._run_sub(data, network, device, slicers, pbar)
                 pbar.desc = "finish"
                 pbar.update(0)