diff --git a/pyproject.toml b/pyproject.toml
index 51c3037235..49347c8ab0 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -36,7 +36,7 @@ huggingface = [
     "unsloth_zoo>=2024.11.8",
     "packaging",
     "tyro",
-    "transformers>=4.46.1",
+    "transformers>=4.46.1,<=4.46.3",
     "datasets>=2.16.0",
     "sentencepiece>=0.2.0",
     "tqdm",
@@ -247,7 +247,7 @@ colab-new = [
     "unsloth_zoo>=2024.11.8",
     "packaging",
     "tyro",
-    "transformers>=4.46.1",
+    "transformers>=4.46.1,<=4.46.3",
     "datasets>=2.16.0",
     "sentencepiece>=0.2.0",
     "tqdm",
diff --git a/unsloth/models/llama.py b/unsloth/models/llama.py
index 3e67b4e89a..f8fb7d945d 100644
--- a/unsloth/models/llama.py
+++ b/unsloth/models/llama.py
@@ -1054,7 +1054,6 @@ def CausalLM_fast_forward(fast_forward_inference):
                 self.extra_ignored_labels = torch.full((self.max_seq_length, 1), -100, device = "cuda:0")
             pass
             shift_labels = torch.hstack((labels[..., 1:], self.extra_ignored_labels[:labels.shape[0]]))
-            print(kwargs.get("num_items_in_batch", None) or kwargs.get("n_items", None))
             loss = fast_cross_entropy_loss(
                 logits = shift_logits,
                 labels = shift_labels,