[KVCache] Make KVCacheSpec hashable (#21791)

Chen Zhang committed over 1 year ago

[Docs] Fix the outdated URL for installing from vLLM binaries (#21523)

Kay Yan committed over 1 year ago

[Model] Refactor JambaForCausalLM (#21394)

Jee Jee Li committed over 1 year ago

[V0 deprecation] Guided decoding (#21347)

Reza Barazesh committed over 1 year ago

[Model]: Fused MoE for nomic-embed-text-v2-moe (#18321)

Isotr0py committed over 1 year ago

[Doc] Link to RFC for pooling optimizations (#21806)

Cyrus Leung committed over 1 year ago

[Misc] Rework process titles (#21780)

Nick Hill committed over 1 year ago

Migrate InternVLImageInputs and InternVLVideoInputs to TensorSchema (#21684)

Benji Beck committed over 1 year ago

[AMD][CI/Build][Bugfix] Guarding CUDA specific functions by ifndef ROCM (#21766)

Gregory Shtrasberg committed over 1 year ago

skip fusedmoe layer for start_load_kv (#21378)

Calvin Chen committed over 1 year ago

[CI] Parallelize Kernels MoE Test (#21764)

Michael Goin committed over 1 year ago

[Refactor] Merge Compressed Tensor FP8 `CompressedTensorsW8A8Fp8MoEMethod` and `CompressedTensorsW8A8Fp8MoECutlassMethod` (#21775)

Wentao Ye committed over 1 year ago

[Docs] Minimize spacing for supported_hardware.md table (#21779)

Michael Goin committed over 1 year ago

[Kernel] SM90 CUTLASS FP8 GEMM: add support for swap AB + kernel tuning (#20396)

lyrisz committed over 1 year ago

[Perf] Disable chunked local attention by default with llama4 (#21761)

Lucas Wilkinson committed over 1 year ago

[Feat]: Add support for Dynamic Quant 4 bit CPU kleidiai kernels (#17112)

Nikhil Gupta committed over 1 year ago

[Bugfix] DeepGEMM is not enabled on B200 due to `_lazy_init()` (#21472)

Clayton Coleman committed over 1 year ago

Revert "[V1] Exception Handling when Loading KV Cache from Remote Store" (#21778)

Kuntai Du committed over 1 year ago

[AMD][CI/Build] Fix the AMD issue caused by inappropriate of symbol exposure (#21647)

Lu Fang committed over 1 year ago

[Bug] Enforce contiguous input for `dynamic_scaled_fp8_quant` and `static_scaled_fp8_quant` (#21773)

Wentao Ye committed over 1 year ago