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/requirements/cuda.txt

@@ -0,0 +1,19 @@
+# Common dependencies
+-r common.txt
+
+numba == 0.61.2 # Required for N-gram speculative decoding
+
+# Dependencies for NVIDIA GPUs
+torch==2.10.0
+torchaudio==2.10.0
+# These must be updated alongside torch
+torchvision==0.25.0 # Required for phi3v processor. See https://github.com/pytorch/vision?tab=readme-ov-file#installation for corresponding version
+# FlashInfer should be updated together with the Dockerfile
+flashinfer-python==0.6.6
+# Cap nvidia-cudnn-frontend (transitive dep of flashinfer) due to
+# breaking changes in 1.19.0
+nvidia-cudnn-frontend>=1.13.0,<1.19.0
+
+# QuACK and Cutlass DSL for FA4 (cute-DSL implementation)
+nvidia-cutlass-dsl>=4.4.0.dev1
+quack-kernels>=0.2.7