A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 10, 09:00 UTC
main
[Bugfix] Fix integer overflow in Gemma3n audio processing (#31657)
Jeremy Teboul committed 10 months ago
[Bugfix] fix offline chat output prompt (#32076)
Ning Xie committed 10 months ago
[Benchmark][1/2] Generalize SLA criterion validation from binary flags to margins (#32075)
Cyrus Leung committed 10 months ago
Frelam committed 10 months ago
[Misc] Delay deprecation of CommonAttentionMetadata properties (#32074)
Lucas Wilkinson committed 10 months ago
[ROCm][CI] Fix flaky `test_function_calling_with_stream` and reduce schema test examples (#32063)
Andreas Karatzas committed 10 months ago
Update modelopt KV cache quantization resolution to new scheme (#31895)
roikoren755 committed 10 months ago
[Refactor] Separate sequence and token pooling types (#32026)
Cyrus Leung committed 10 months ago
[Core] Refactor ColumnParallelLinear: remove unused parameter and optimize forward (#31939)
maang committed 10 months ago
[Bugfix][Hardware][AMD] Use dynamic WARP_SIZE in sampler vectorized_process (#31295)
Kevin McKay committed 10 months ago
Fuse RoPE and MLA KV-cache write (#25774)
PatrykSaffer committed 10 months ago
Akshat Shrivastava committed 10 months ago
[Misc][LLaMa4] Compile LLaMa Vision Encoder (#30709)
Lucas Kabela committed 10 months ago
resolve pydantic error in startup benchmark (#31348)
Ning Xie committed 10 months ago
[Bugfix] Narrow broad exceptions in compilation backends (#31616)
Kevin McKay committed 10 months ago
[CI] Allow Deprecated Quantization For LM Eval Tests (#32065)
Micah Williamson committed 10 months ago
[Perf] Optimize async scheduling placeholder using empty (#32056)
Wentao Ye committed 10 months ago
[Core] Use weights_only=True with torch.load (#32045)
Russell Bryant committed 10 months ago
[2/N][Attention] Fix pre-commit errors (#32052)
Matthew Bonanni committed 10 months ago
[Misc][BE] Type coverage for vllm/compilation [2/3] (#31744)
Lucas Kabela committed 10 months ago