diff --git a/tests/unit/encoders/test_vit.py b/tests/unit/encoders/test_vit.py
index 72d01b349bc83a0b824e2f2f746d29c0553c94ce..9970e31075a44396c5fbfeaff0a2ce0a846487ca 100644
--- a/tests/unit/encoders/test_vit.py
+++ b/tests/unit/encoders/test_vit.py
@@ -4,8 +4,14 @@ import torch
 from PIL import Image
 
 from semantic_router.encoders import VitEncoder
+from transformers import AutoTokenizer, AutoModel
 
 test_model_name = "hf-internal-testing/tiny-random-vit"
+
+# force the model download
+tokenizer = AutoTokenizer.from_pretrained(test_model_name, force_download=True)
+model = AutoModel.from_pretrained(test_model_name, force_download=True)
+
 vit_encoder = VitEncoder(name=test_model_name)
 embed_dim = 32