[V1][Kernel] Flashinfer HND KV cache layout (#19280)

Nicolò Lucchesi committed over 1 year ago

[doc] split "Other AI Accelerators" tabs (#19708)

David Xia committed over 1 year ago

[doc][mkdocs] Add edit button to documentation (#19637)

Reid committed over 1 year ago

[Kernel] Add Split-KV Support to Unified Triton Attention Kernel (#19152)

jvlunteren committed over 1 year ago

Add a doc on how to update PyTorch version (#19705)

Huy Do committed over 1 year ago

[Doc] Add missing llava family multi-image examples (#19698)

Isotr0py committed over 1 year ago

[Core] add remove_seq_from_computed_blocks_tracker to BlockSpaceManager (#19686)

quanliu committed over 1 year ago

Fixes IMA for TP w/ flex-attention (#19712)

Driss Guessous committed over 1 year ago

[DOC] fix doc typos (#19600)

Di Liu committed over 1 year ago

[Frontend] add chunking audio for > 30s audio (#19597)

nguyenhoangthuan99 committed over 1 year ago

[Wheel Size] Only build FA2 8.0+PTX (#19336)

Lucas Wilkinson committed over 1 year ago

[doc] add project flag to gcloud TPU command (#19664)

David Xia committed over 1 year ago

[Fix] Fall back to Gloo when NCCL backend is unavailable (#19641)

Conroy Cheers committed over 1 year ago

[Quantization] Remove FP4 emulation; Fall-back to marlin for device < 100 (#19563)

Dipika Sikka committed over 1 year ago

[V1] Change return type on get_multimodal_embeddings() (#19446)

Russell Bryant committed over 1 year ago

[Model] Add support for MiniMaxM1ForCausalLM (shares architecture with MiniMaxText01ForCausalLM) (#19677)

qscqesze committed over 1 year ago

[Kernels] Use empty for modular MoE workspaces (#19667)

bnellnm committed over 1 year ago

[Bugfix] fix missing 'finish_reason': null in streaming chat (#19662)

Chauncey committed over 1 year ago

[MISC] bump huggingface_hub pkg to 0.33.0 (#19547)

Ning Xie committed over 1 year ago

[Bugfix] Fix TP inference for Flex attention backend (#19657)

Isotr0py committed over 1 year ago