vllm
5f1af97f - [V1] [Hybrid] Enable Full CUDA graph by default for hybrid models in V1 (#22594)

Commit

104 days ago

[V1] [Hybrid] Enable Full CUDA graph by default for hybrid models in V1 (#22594) Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>

References

Author

tdoublep

Parents