diff --git a/tests/test_pipeline_mixin.py b/tests/test_pipeline_mixin.py
index eebf4eaa256..3447577ac57 100644
--- a/tests/test_pipeline_mixin.py
+++ b/tests/test_pipeline_mixin.py
@@ -502,8 +502,12 @@ def validate_test_components(test_case, task, model, tokenizer, processor):
     if tokenizer is not None:
         config_vocab_size = getattr(model.config, "vocab_size", None)
         # For CLIP-like models
-        if config_vocab_size is None and hasattr(model.config, "text_config"):
-            config_vocab_size = getattr(model.config.text_config, "vocab_size", None)
+        if config_vocab_size is None:
+            if hasattr(model.config, "text_config"):
+                config_vocab_size = getattr(model.config.text_config, "vocab_size", None)
+            elif hasattr(model.config, "text_encoder"):
+                config_vocab_size = getattr(model.config.text_encoder, "vocab_size", None)
+
         if config_vocab_size is None and model.config.__class__.__name__ not in CONFIG_WITHOUT_VOCAB_SIZE:
             raise ValueError(
                 "Could not determine `vocab_size` from model configuration while `tokenizer` is not `None`."
diff --git a/tests/utils/tiny_model_summary.json b/tests/utils/tiny_model_summary.json
index e084780268f..2a1efa7d88d 100644
--- a/tests/utils/tiny_model_summary.json
+++ b/tests/utils/tiny_model_summary.json
@@ -4151,7 +4151,7 @@
         ],
         "processor_classes": [],
         "model_classes": [],
-        "sha": "37e9ae5dafb601daa8364e9ac17da31cd82b274b"
+        "sha": "f67d387eaaa7c71ddf88af95eda4bf14ace08d49"
     },
     "MusicgenForConditionalGeneration": {
         "tokenizer_classes": [
@@ -4161,7 +4161,7 @@
         "model_classes": [
             "MusicgenForConditionalGeneration"
         ],
-        "sha": "b71611b88832e53370e676da53b65042f7fc78ee"
+        "sha": "16102cdf580e70cf0b4e0e2cda5bc75b934da92c"
     },
     "MvpForCausalLM": {
         "tokenizer_classes": [