vllm-project/vllm

Pull Requests Commits

[V1][Metrics] Add request_success_total counter, labelled with finish reason (#12579)

markmc committed 133 days ago

Verified 233df6f5

[Bugfix] Fix CI failures for InternVL and Mantis models (#12728)

DarkLight1337 committed 133 days ago

Verified 18016a5e

[Build] update requirements of no-device for plugin usage (#12630)

sducouedic committed 133 days ago

Verified 649550f2

Avoid unnecessary multi-modal input data copy when len(batch) == 1 (#12722)

imkero committed 133 days ago

Verified 62467a83

[Bugfix] Fix loading of fine-tuned models based on Phi-3-Small (#12689)

mgtk77 committed 133 days ago

Verified 6469038b

[VLM] merged multimodal processor and V1 support for idefics3 (#12660)

Isotr0py committed 133 days ago

Verified 815079de

[V1] Remove scheduling constraint on partial requests (#12674)

WoosukKwon committed 133 days ago

Verified 18a88fcc

[VLM] Merged multi-modal processor for InternVL-based models (#12553)

DarkLight1337 committed 133 days ago

Verified d1ca7df8

[Misc] Add BNB quantization for Whisper (#12381)

jeejeelee committed 133 days ago

Verified 96b23621

[AMD][ROCm] Enable DeepSeek model on ROCm (#12662)

hongxiayang committed 133 days ago

Verified c36ac98d

[Quant] Fix use_mla TypeError and support loading pure-sparsity Compressed Tensors configs (#12711)

kylesayrs committed 133 days ago

Verified 4896d0c2

[Doc] Replace ibm-fms with ibm-ai-platform (#12709)

tdoublep committed 133 days ago

Verified bb392af4

Support Pixtral-Large HF by using llava multimodal_projector_bias config (#12710)

mgoin committed 133 days ago

Verified 5d98d560

[Core] Improve hash collision avoidance in prefix caching (#12621)

russellb committed 134 days ago

Verified 73b35cca

[V1] Revert `uncache_blocks` and support recaching full blocks (#12415)

comaniac committed 134 days ago

Verified 5095e966

[MISC] Remove model input dumping when exception (#12582)

comaniac committed 134 days ago

Verified cf58b9c4

[Model] Add Deepseek V3 fp8_w8a8 configs for B200 (#12707)

kushanam committed 134 days ago

Verified 4797dad3

Squelch MLA warning for Compressed-Tensors Models (#12704)

kylesayrs committed 134 days ago

Verified 6dd5e528

[Bugfix][Kernel] Fix per-token/per-channel quantization for Hopper scaled mm (#12696)

tlrmchlsmth committed 134 days ago

Verified c11de33d

[Misc] Fix improper placement of SPDX header in scripts (#12694)

russellb committed 134 days ago

Verified 33e0602e

[Model]: Add `transformers` backend support (#11330)

ArthurZucker committed 134 days ago

Verified a1a2aaad

[ci/build] fix gh200 test (#12681)

youkaichao committed 134 days ago

Verified 1298a400

[cuda] manually import the correct pynvml module (#12679)

youkaichao committed 134 days ago

Verified ad4a9dc8

Fix for attention layers to remain unquantized during moe_wn16 quant (#12570)

srikanthsrnvs committed 134 days ago

Verified b9986454

Properly check if all fused layers are in the list of targets (#12666)

eldarkurtic committed 134 days ago

Verified c5932e5d

make sure mistral_common not imported for non-mistral models (#12669)

youkaichao committed 134 days ago

Verified 20579c0f

[Kernel] port sgl moe_align_block_size kernels (#12574)

chenyang78 committed 134 days ago

Verified 95460fc5

[Doc] Deprecate Discord (#12668)

zhuohan123 committed 134 days ago

Verified 326fcc8b

[doc][misc] clarify VLLM_HOST_IP for multi-node inference (#12667)

youkaichao committed 135 days ago

Verified e6433091

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

russellb committed 135 days ago

Verified e489ad7a

Newer Older