diff --git a/.github/workflows/python-package.yml b/.github/workflows/python-package.yml
index 622cb32..01857dc 100644
--- a/.github/workflows/python-package.yml
+++ b/.github/workflows/python-package.yml
@@ -18,7 +18,7 @@ jobs:
         python-version: ["3.10", "3.12", "3.13"]
 
     steps:
-      - uses: actions/checkout@v6
+      - uses: actions/checkout@v7
       - name: Set up Python ${{ matrix.python-version }}
         uses: actions/setup-python@v6
         with:
@@ -29,7 +29,7 @@ jobs:
           pip install uv
           uv pip install flake8 pytest sentencepiece rdkit --system
           uv pip install git+https://github.com/novonordisk-research/pepfunn.git --system
-          uv pip install .  "transformers<5" --system
+          uv pip install . "transformers<5.13" --system
           if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
       - name: Lint with flake8
         run: |
diff --git a/autopeptideml/db/negative_sampling.py b/autopeptideml/db/negative_sampling.py
index 5c5e69f..61a861c 100644
--- a/autopeptideml/db/negative_sampling.py
+++ b/autopeptideml/db/negative_sampling.py
@@ -149,7 +149,8 @@ def add_negatives_from_db(
     verbose: bool = True,
     sample_by: str = 'mw',
     n_jobs: int = cpu_count(),
-    random_state: int = 1
+    random_state: int = 1,
+    _to_seq: bool = False
 ) -> pd.DataFrame:
     """
     Augments a dataset with negative samples from a target database to achieve a desired negative/positive ratio.
@@ -206,7 +207,7 @@ def add_negatives_from_db(
 
     if isinstance(target_db, pd.DataFrame):
         db = target_db
-
+        path = None
     else:
         db, path = get_neg_db(target_db, verbose=verbose, return_path=True)
 
@@ -231,11 +232,14 @@ def add_negatives_from_db(
             n_bins = 50
         else:
             n_bins = 10
-    if 'sequence' not in db:
+    if 'sequence' not in db and _to_seq:
         from autopeptideml.pipeline import get_pipeline
         pipe = get_pipeline('to-sequences')
         db['sequence'] = pipe(db['smiles'], n_jobs=n_jobs, verbose=verbose)
-        db.to_csv(path, index=False)
+        if path is not None:
+            db.to_csv(path, index=False)
+    if not _to_seq:
+        db['sequence'] = [None for s in range(len(db))]
     db[sample_by], disc = discretizer(db[sample_by].to_numpy(),
                                       n_bins=n_bins,
                                       return_discretizer=True)