[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>

tests/engine/test_skip_tokenizer_init.py

@@ -2,16 +2,21 @@
 
 import pytest
 
+from vllm.config import LoadFormat
 from vllm.entrypoints.llm import LLM
 from vllm.sampling_params import SamplingParams
 
+from ..conftest import MODEL_WEIGHTS_S3_BUCKET
 
-@pytest.mark.parametrize("model", ["facebook/opt-125m"])
+
+@pytest.mark.parametrize("model", [f"{MODEL_WEIGHTS_S3_BUCKET}/distilgpt2"])
 def test_skip_tokenizer_initialization(model: str):
     # This test checks if the flag skip_tokenizer_init skips the initialization
     # of tokenizer and detokenizer. The generated output is expected to contain
     # token ids.
-    llm = LLM(model=model, skip_tokenizer_init=True)
+    llm = LLM(model=model,
+              skip_tokenizer_init=True,
+              load_format=LoadFormat.RUNAI_STREAMER)
     sampling_params = SamplingParams(prompt_logprobs=True, detokenize=True)
 
     with pytest.raises(ValueError, match="cannot pass text prompts when"):