[1/n][CI] Load models in CI from S3 instead of HF (#13205)

Signed-off-by: <> Co-authored-by: EC2 Default User <ec2-user@ip-172-31-20-117.us-west-2.compute.internal>
2025-02-18 23:34:59 -08:00
parent fd84857f64
commit d5d214ac7f
43 changed files with 225 additions and 76 deletions
--- a/tests/entrypoints/llm/test_collective_rpc.py
+++ b/tests/entrypoints/llm/test_collective_rpc.py
@@ -28,7 +28,7 @@ def test_collective_rpc(tp_size, backend):
        def echo_rank(self):
            return self.rank

-    llm = LLM(model="meta-llama/Llama-3.2-1B-Instruct",
+    llm = LLM(model="s3://vllm-ci-model-weights/Llama-3.2-1B-Instruct",
              enforce_eager=True,
              load_format="dummy",
              tensor_parallel_size=tp_size,