fix(invoke_handler.py): remove hard coded bedrock final usage chunk

litellm/llms/bedrock/chat/invoke_handler.py

@@ -1274,13 +1274,6 @@ class AWSEventStreamDecoder:
     def converse_chunk_parser(self, chunk_data: dict) -> ModelResponseStream:
         try:
             verbose_logger.debug("\n\nRaw Chunk: {}\n\n".format(chunk_data))
-            chunk_data["usage"] = {
-                "inputTokens": 3,
-                "outputTokens": 392,
-                "totalTokens": 2191,
-                "cacheReadInputTokens": 1796,
-                "cacheWriteInputTokens": 0,
-            }
             text = ""
             tool_use: Optional[ChatCompletionToolCallChunk] = None
             finish_reason = ""

litellm/proxy/_new_secret_config.yaml

@@ -5,7 +5,14 @@ model_list:
     api_key: os.environ/AZURE_API_KEY
     api_base: http://0.0.0.0:8090
     rpm: 3
-    
+ - model_name: "gpt-4o-mini-openai"
+   litellm_params:
+    model: gpt-4o-mini
+    api_key: os.environ/OPENAI_API_KEY
+ - model_name: "bedrock-nova"
+   litellm_params:
+    model: us.amazon.nova-pro-v1:0
+
 litellm_settings:
   num_retries: 0