fix(vertex_ai): accept encoding_format param for Gemini/Vertex AI embeddings.

litellm/llms/vertex_ai/vertex_embeddings/transformation.py

@@ -72,7 +72,7 @@ def get_config(cls):
         }
 
     def get_supported_openai_params(self):
-        return ["dimensions"]
+        return ["dimensions", "encoding_format"]
 
     def map_openai_params(
         self, non_default_params: dict, optional_params: dict, kwargs: dict

tests/test_litellm/llms/vertex_ai/test_vertex_ai_embedding_encoding_format.py

@@ -0,0 +1,101 @@
+"""
+Test that encoding_format parameter is accepted for Vertex AI / Gemini embeddings.
+
+This test verifies the fix for the issue where the OpenAI client sends
+encoding_format: base64 automatically, which was causing UnsupportedParamsError
+for Gemini embeddings.
+
+Reference: https://github.com/BerriAI/litellm/issues/XXXX
+"""
+
+import pytest
+
+import litellm
+from litellm.utils import get_optional_params_embeddings, get_llm_provider
+
+
+class TestVertexAIEmbeddingEncodingFormat:
+    """Test encoding_format parameter handling for Vertex AI embeddings."""
+
+    def test_vertex_ai_embedding_accepts_encoding_format_without_drop_params(self):
+        """
+        Test that encoding_format is accepted without needing drop_params=True.
+
+        The OpenAI Python client automatically sends encoding_format: base64.
+        This should not raise an UnsupportedParamsError for Vertex AI embeddings.
+        """
+        litellm.drop_params = False
+
+        model, custom_llm_provider, _, _ = get_llm_provider(
+            model="vertex_ai/textembedding-gecko"
+        )
+
+        optional_params = get_optional_params_embeddings(
+            model=model,
+            encoding_format="base64",
+            custom_llm_provider=custom_llm_provider,
+        )
+
+        assert "encoding_format" not in optional_params
+
+    def test_gemini_embedding_accepts_encoding_format_without_drop_params(self):
+        """
+        Test that encoding_format is accepted for Gemini embeddings.
+
+        The OpenAI Python client automatically sends encoding_format: base64.
+        This should not raise an UnsupportedParamsError for Gemini embeddings.
+        """
+        litellm.drop_params = False
+
+        model, custom_llm_provider, _, _ = get_llm_provider(
+            model="gemini/text-embedding-004"
+        )
+
+        optional_params = get_optional_params_embeddings(
+            model=model,
+            encoding_format="base64",
+            custom_llm_provider=custom_llm_provider,
+        )
+
+        assert "encoding_format" not in optional_params
+
+    def test_vertex_ai_embedding_with_dimensions_and_encoding_format(self):
+        """
+        Test that both dimensions and encoding_format work together.
+
+        dimensions should be passed through, encoding_format should be ignored.
+        """
+        litellm.drop_params = False
+
+        model, custom_llm_provider, _, _ = get_llm_provider(
+            model="vertex_ai/text-embedding-005"
+        )
+
+        optional_params = get_optional_params_embeddings(
+            model=model,
+            dimensions=256,
+            encoding_format="base64",
+            custom_llm_provider=custom_llm_provider,
+        )
+
+        assert optional_params.get("outputDimensionality") == 256
+        assert "encoding_format" not in optional_params
+
+    def test_vertex_ai_embedding_encoding_format_float(self):
+        """
+        Test that encoding_format: float is also accepted and ignored.
+        """
+        litellm.drop_params = False
+
+        model, custom_llm_provider, _, _ = get_llm_provider(
+            model="vertex_ai/textembedding-gecko"
+        )
+
+        optional_params = get_optional_params_embeddings(
+            model=model,
+            encoding_format="float",
+            custom_llm_provider=custom_llm_provider,
+        )
+
+        assert "encoding_format" not in optional_params
+