No 70b

worker/tests/test_multimodel/test_inference_llama70B.py

@@ -1,7 +1,7 @@
 import asyncio
+import os
 from logging import Logger
 from typing import Callable
-import os
 
 import pytest
 
@@ -151,7 +151,13 @@ async def test_2_runner_inference(
 
 
 
-
+@pytest.mark.skipif(
+    not (
+        os.path.exists(os.path.expanduser("~/.exo/models/mlx-community--Llama-3.3-70B-Instruct-4bit/")) 
+        and _get_model_size_gb(os.path.expanduser("~/.exo/models/mlx-community--Llama-3.3-70B-Instruct-4bit/")) > 30
+    ),
+    reason="This test only runs when model mlx-community/Llama-3.3-70B-Instruct-4bit is downloaded"
+)
 async def test_parallel_inference(
     logger: Logger,
     pipeline_shard_meta: Callable[[int, int], PipelineShardMetadata],