Add vLLM v0.18.1 source tree with KV transfer abort fix

third_party/vllm/ now tracked in git for direct patch management.
Based on vLLM v0.18.1 release with one patch applied:

  vllm/v1/core/sched/scheduler.py:
    Replace fatal assert with graceful skip when KV transfer callback
    arrives for an already-aborted request during PD disaggregated serving.

Future vLLM modifications should be made directly in third_party/vllm/
and committed normally. The patches/ directory is kept as documentation
of what changed from upstream.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

third_party/vllm/tests/transformers_utils/test_config.py

@@ -0,0 +1,32 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+"""
+This test file includes some cases where it is inappropriate to
+only get the `eos_token_id` from the tokenizer as defined by
+`BaseRenderer.get_eos_token_id`.
+"""
+
+from vllm.tokenizers import get_tokenizer
+from vllm.transformers_utils.config import try_get_generation_config
+
+
+def test_get_llama3_eos_token():
+    model_name = "meta-llama/Llama-3.2-1B-Instruct"
+
+    tokenizer = get_tokenizer(model_name)
+    assert tokenizer.eos_token_id == 128009
+
+    generation_config = try_get_generation_config(model_name, trust_remote_code=False)
+    assert generation_config is not None
+    assert generation_config.eos_token_id == [128001, 128008, 128009]
+
+
+def test_get_blip2_eos_token():
+    model_name = "Salesforce/blip2-opt-2.7b"
+
+    tokenizer = get_tokenizer(model_name)
+    assert tokenizer.eos_token_id == 2
+
+    generation_config = try_get_generation_config(model_name, trust_remote_code=False)
+    assert generation_config is not None
+    assert generation_config.eos_token_id == 50118