vllm-project/vllm

Pull Requests Commits

khluu committed 1 day ago

2ad06142

[CI/Build][AMD] Skip tests in test_fusions_e2e and test_dbo_dp_ep_gsm8k that require non-existing imports for ROCm (#30417)

rasmith committed 1 day ago

Verified 48661d27

[FIX]Patch run-cluster.sh (fix for #28328) (#30002)

evberrypi committed 1 day ago

Verified d527cf0b

[ROCM][CI] Fix AMD Examples Test Group (#30276)

Concurrensee committed 1 day ago

Verified 2cc5affc

[EPLB] Support EPLB w/ NVFP4 (#29804)

andrewbriand committed 1 day ago

Verified a00d8897

[Refactor] Remove useless syncwarp (#30510)

yewentao256 committed 1 day ago

Verified 61249b17

[Perf] Optimize deepgemm experts initialization, 3.9% TTFT improvement (#30494)

yewentao256 committed 1 day ago

Verified c817b141

[Docs][CPU backend] Add pre-built Arm CPU Docker images (#30491)

ioghiban committed 1 day ago

Verified 3efdc3fe

[Core] Whisper Enable Encoder Batching (#29421)

NickLucche committed 1 day ago

Verified 0efd9f86

[BugFix][MM]support VLLM_RANDOMIZE_DP_DUMMY_INPUTS (#30472)

charlotte12l committed 1 day ago

Verified 90d6cf92

Standardise `get_rope` to use `rope_parameters["partial_rotary_factor"]`, not `rotary_dim` (#30389)

hmellor committed 1 day ago

Verified cf3eacfe

[compile] Stop one-off setting enable_aot_compile and use context manager instead. (#30503)

zhxchen17 committed 1 day ago

Verified 92fea56f

[Misc] Add mcp to requirements (#30474)

yeqcharlotte committed 1 day ago

Verified e458270a

[ROCm][Bugfix] Add MLACommonMetadata to allowed attention types for speculative decoding (#30430)

AndreasKaratzas committed 1 day ago

Verified 72aaac5b

[Feature] AWQ marlin quantization support for fused moe with lora (#30442)

princepride committed 1 day ago

Verified 0e71eaa6

Add Eagle and Eagle3 support to Transformers modeling backend (#30340)

hmellor committed 1 day ago

Verified 8781cd6b

[IMPROVEMENT] Change MistralReasoningParser behavior (#30391)

juliendenize committed 1 day ago

Verified aa3c250c

[CI] refine more logic when generating and using nightly wheels & indices, add cuda130 build for aarch64, specify correct manylinux version (#30341)

Harry-Chen committed 1 day ago

Verified 305b168a

Give pooling examples better names (#30488)

hmellor committed 1 day ago

Verified 93db3256

[Docs][CPU Backend] Add nightly and per revision pre-built Arm CPU wheels (#30402)

ioghiban committed 1 day ago

Verified 17cb5402

Make the `httpx` logger less annoying when Transformers v5 is installed (#30480)

hmellor committed 1 day ago

Verified 97a042f3

[Misc] Improve error message for `is_multimodal` (#30483)

DarkLight1337 committed 2 days ago

Verified 3a3b06ee

[KVConnector] Add KV events to KV Connectors (#28309)

hickeyma committed 2 days ago

Verified f4417f84

[Misc][PCP&DCP] relocate PCP feature check (#30050)

pisceskkk committed 2 days ago

Verified a11f4a81

Fix typo of endpoint name in CLI args docs (#30473)

kmaehashi committed 2 days ago

Verified 853611bb

[Bugfix] Fix `task` still being passed in tests/benchmarks (#30476)

DarkLight1337 committed 2 days ago

Verified d917747c

[Deprecation] Deprecation `--convert reward`, use `--convert embed` instead. (#30463)

noooop committed 2 days ago

Verified a5f9fb59

[Fix] Update lazing loading of video loader backend (#30444)

jeremyteboul committed 2 days ago

Verified 4515eb1a

[Deprecation] Remove missed fallback for `embed_input_ids` (#30469)

DarkLight1337 committed 2 days ago

Verified 13d63b65

[Fix]fix import error from lmcache (#30376)

wz1qqx committed 2 days ago

Verified b4e8b912

Older