diff --git a/inst/scripts/PredictScTourPseudotime.py b/inst/scripts/PredictScTourPseudotime.py
index a7803f5d..147089b5 100644
--- a/inst/scripts/PredictScTourPseudotime.py
+++ b/inst/scripts/PredictScTourPseudotime.py
@@ -6,6 +6,8 @@
 import anndata
 from anndata import AnnData
 
+torch.serialization.add_safe_globals([AnnData, anndata._core.file_backing.AnnDataFileManager]):
+
 def PredictPseudotime(GEXfile, model_file, ptime_out_file, embedding_out_file):
     #read count data and variable genes
     adataObj = sc.read_10x_h5(GEXfile)
@@ -16,8 +18,7 @@ def PredictPseudotime(GEXfile, model_file, ptime_out_file, embedding_out_file):
         print('AnnData object is a csr matrix, converting to dense because scipy depreciated the .A shorthand')
         adataObj.X = adataObj.X.toarray()
 
-    with torch.serialization.add_safe_globals([AnnData, anndata._core.file_backing.AnnDataFileManager]):
-        checkpoint = torch.load(model_file, map_location=torch.device('cpu'))
+    checkpoint = torch.load(model_file, map_location=torch.device('cpu'), weights_only = True)
     model_adata = checkpoint['adata']
 
     genes_in_model = model_adata.var.index.values.tolist()
@@ -30,8 +31,8 @@ def PredictPseudotime(GEXfile, model_file, ptime_out_file, embedding_out_file):
     #subset to genes found in the pretrained model.
     adataObj = adataObj[:, genes_in_model]
     #initalize a trainer and pull a previously saved model from model_file
-    with torch.serialization.safe_globals([AnnData, anndata._core.file_backing.AnnDataFileManager]):
-        tnode = sct.predict.load_model(model_file)
+
+    tnode = sct.predict.load_model(model_file)
     pred_t = sct.predict.predict_time(adata = adataObj, model = tnode)
     adataObj.obs['ptime'] = pred_t
     mix_zs, zs, pred_zs  = sct.predict.predict_latentsp(adata = adataObj, model = tnode)