feat: async run

tasks.py

@@ -162,6 +162,11 @@ class Task:
 
         return self.result
 
+    async def arun(self, pipeline):
+        self.outputs = self.outputs or await pipeline(self.samples)
+
+        return self.result
+
 
 def multichoice(responses: Any, references: list[str]):
     if isinstance(responses[0], str):

tlem.py

@@ -103,6 +103,16 @@ class Suite(EvaluationSuite):
             _ = task.run(model_or_pipeline)
         return self.aggregate(deepcopy(self.suite))
 
+    def arun(self, model_or_pipeline):
+        async def sync_function():
+            return await tqdm.gather(
+                *[task.run(model_or_pipeline) for task in self.tasks], leave=False
+            )
+
+        asyncio.run(sync_function())
+
+        return self.aggregate(deepcopy(self.suite))
+
     def get_suite(self, name) -> dict[str, Task]:
         chat = False
         match name: