Fix passing top_k parameter for Bedrock Anthropic models

litellm/llms/bedrock/chat/converse_transformation.py

@@ -332,9 +332,27 @@ def _transform_inference_params(self, inference_params: dict) -> InferenceConfig
         if "top_k" in inference_params:
             inference_params["topK"] = inference_params.pop("top_k")
         return InferenceConfig(**inference_params)
+
+    def _handle_top_k_value(self, model: str, inference_params: dict) -> dict:
+        base_model = self._get_base_model(model)
+
+        val_top_k = None
+        if "topK" in inference_params:
+            val_top_k = inference_params.pop("topK")
+        elif "top_k" in inference_params:
+            val_top_k = inference_params.pop("top_k")
+
+        if val_top_k:
+            if (base_model.startswith("anthropic")):
+                return {"top_k": val_top_k}
+            if base_model.startswith("amazon.nova"):
+                return {'inferenceConfig': {"topK": val_top_k}}                
+
+        return {}
 
     def _transform_request_helper(
         self,
+        model: str,
         system_content_blocks: List[SystemContentBlock],
         optional_params: dict,
         messages: Optional[List[AllMessageValues]] = None,
@@ -361,35 +379,20 @@ def _transform_request_helper(
                 )
 
         inference_params = copy.deepcopy(optional_params)
-        additional_request_keys = []
-        additional_request_params = {}
         supported_converse_params = list(
             AmazonConverseConfig.__annotations__.keys()
         ) + ["top_k"]
         supported_tool_call_params = ["tools", "tool_choice"]
         supported_guardrail_params = ["guardrailConfig"]
+        total_supported_params = supported_converse_params + supported_tool_call_params + supported_guardrail_params
         inference_params.pop("json_mode", None)  # used for handling json_schema
 
-        # send all model-specific params in 'additional_request_params'
-        for k, v in inference_params.items():
-            if (
-                k not in supported_converse_params
-                and k not in supported_tool_call_params
-                and k not in supported_guardrail_params
-            ):
-                additional_request_params[k] = v
-                additional_request_keys.append(k)
-        for key in additional_request_keys:
-            inference_params.pop(key, None)
+        # keep supported params in 'inference_params', and set all model-specific params in 'additional_request_params'
+        additional_request_params = {k: v for k, v in inference_params.items() if k not in total_supported_params}
+        inference_params = {k: v for k, v in inference_params.items() if k in total_supported_params}
 
-        if "topK" in inference_params:
-            additional_request_params["inferenceConfig"] = {
-                "topK": inference_params.pop("topK")
-            }
-        elif "top_k" in inference_params:
-            additional_request_params["inferenceConfig"] = {
-                "topK": inference_params.pop("top_k")
-            }
+        # Only set the topK value in for models that support it
+        additional_request_params.update(self._handle_top_k_value(model, inference_params))
 
         bedrock_tools: List[ToolBlock] = _bedrock_tools_pt(
             inference_params.pop("tools", [])
@@ -437,6 +440,7 @@ async def _async_transform_request(
         ## TRANSFORMATION ##
 
         _data: CommonRequestObject = self._transform_request_helper(
+            model=model,
             system_content_blocks=system_content_blocks,
             optional_params=optional_params,
             messages=messages,
@@ -483,6 +487,7 @@ def _transform_request(
         messages, system_content_blocks = self._transform_system_message(messages)
 
         _data: CommonRequestObject = self._transform_request_helper(
+            model=model,
             system_content_blocks=system_content_blocks,
             optional_params=optional_params,
             messages=messages,

tests/llm_translation/test_bedrock_completion.py

@@ -2580,3 +2580,72 @@ def test_bedrock_custom_deepseek():
         except Exception as e:
             print(f"Error: {str(e)}")
             raise e
+
+@pytest.mark.parametrize(
+    "model, expected_output", 
+    [
+        ("bedrock/anthropic.claude-3-sonnet-20240229-v1:0", {"top_k": 3}),
+        ("bedrock/converse/us.amazon.nova-pro-v1:0", {'inferenceConfig': {"topK": 3}}),
+        ("bedrock/meta.llama3-70b-instruct-v1:0", {}),
+    ]
+)
+def test_handle_top_k_value_helper(model, expected_output):
+    assert litellm.AmazonConverseConfig()._handle_top_k_value(model, {"topK": 3}) == expected_output
+    assert litellm.AmazonConverseConfig()._handle_top_k_value(model, {"top_k": 3}) == expected_output
+
+@pytest.mark.parametrize(
+    "model, expected_params", 
+    [
+        ("bedrock/anthropic.claude-3-sonnet-20240229-v1:0", {"top_k": 2}),
+        ("bedrock/converse/us.amazon.nova-pro-v1:0", {'inferenceConfig': {"topK": 2}}),
+        ("bedrock/meta.llama3-70b-instruct-v1:0", {}),
+        ("bedrock/mistral.mistral-7b-instruct-v0:2", {}),
+
+    ]
+)
+def test_bedrock_top_k_param(model, expected_params):
+    import json
+
+    client = HTTPHandler()
+
+    with patch.object(client, "post") as mock_post:
+        mock_response = Mock()
+
+        if ("mistral" in model):
+            mock_response.text = json.dumps({"outputs": [{"text": "Here's a joke...", "stop_reason": "stop"}]}) 
+        else:
+            mock_response.text = json.dumps(
+                {
+                    "output": {
+                        "message": {
+                            "role": "assistant",
+                            "content": [
+                                {
+                                    "text": "Here's a joke..."
+                                }
+                            ]
+                        }
+                    },
+                    "usage": {"inputTokens": 12, "outputTokens": 6, "totalTokens": 18},
+                    "stopReason": "stop"
+                }
+            ) 
+
+        mock_response.status_code = 200
+        # Add required response attributes
+        mock_response.headers = {"Content-Type": "application/json"}
+        mock_response.json = lambda: json.loads(mock_response.text)
+        mock_post.return_value = mock_response
+
+
+        litellm.completion(
+            model=model,
+            messages=[{"role": "user", "content": "Hello, world!"}],
+            top_k=2,
+            client=client
+        )  
+        data = json.loads(mock_post.call_args.kwargs["data"])
+        if ("mistral" in model):
+            assert (data["top_k"] == 2)
+        else:
+            assert (data["additionalModelRequestFields"] == expected_params)