All pull requests

Pull requests

• fix(server): reap UNKNOWN worker records whose heartbeat expired

  fixsize/M
  #965·YiqiuLiu opened Nov 2, 2026Contributor·

  ·
  1

  6

• fix(client): partition kernel JIT cache artifacts by compile config digest

  fixsize/L
  #963·YiqiuLiu opened Nov 2, 2026Contributor·

  ·
  1

  5

• fix(vllm): finalize online quantization before weight post-processing

  fixsize/M
  #962·CDalezyb opened Nov 2, 2026Contributor·

  ·
  3

• perf(client): register CUDA allocations instead of tensors by default

  perfsize/M
  #958·aleksandr-inherent opened Oct 29, 2026·

  ·
  8

• feat(vllm): let reload_model restore weights another loader loaded

  featsize/L
  #957·aleksandr-inherent opened Oct 29, 2026·

  ·
  13

• Support vLLM P2P loading for separate speculative draft checkpoints

  size/L
  #956·sathunsuthakaran-cohere opened Oct 29, 2026·

  ·
  2

• test(ci): add PrimeRL MX refit E2E and align RL workflows

  size/XXLtest
  #955·zhengluo-nv opened Oct 29, 2026Contributor·

  ·
  2

  

• feat(sglang): enable independent draft P2P with NIXL and TransferEngine

  featsize/XXL
  #954·libang-linuxer opened Oct 29, 2026·

  ·
  3

• fix(sglang): keep distributed state local during tensor adoption

  fixsize/M
  #951·zcin opened Oct 28, 2026·

  ·
  5

• fix(tests): define vLLM loader mocks for no-vLLM CI

  fixsize/XS
  #950·zdtsw opened Oct 28, 2026Contributor·3 tasks·

  ·
  3

• feat: optimize draft loading logic

  featsize/L
  #949·vie-serendipity opened Oct 28, 2026·

  ·
  1

  6

• feat: coordinate concurrent vLLM cold starts with donor leases

  featsize/XXL
  #948·sathunsuthakaran-cohere opened Oct 28, 2026·

  ·
  35

• fix(client): reject pool registration of tensors that span allocations

  fixsize/M
  #945·undergg opened Oct 27, 2026Contributor·

  ·
  8

• fix(client): fail clearly when a per-device listener port is taken

  fixsize/L
  #944·undergg opened Oct 27, 2026Contributor·

  ·
  9

• fix(vllm): keep only DefaultModelLoader options in its extra config

  fixsize/M
  #942·undergg opened Oct 27, 2026Contributor·

  ·
  9

• feat(sglang): refill released weights over P2P on resume

  featsize/L
  #941·galletas1712 opened Oct 27, 2026·

  ·
  1

• feat(sglang): Support p2p metadata transfer

  featsize/XXL
  #940·hzxuzhonghu opened Oct 27, 2026·

  ·
  1

  11

• fix(crypto): hash with OpenSSL in the openssl build and keep sha2 out of it

  fixsize/XL
  #935·madhugoutham opened Oct 26, 2026Contributor·

  ·
  1

  9

• fix(vllm): preserve instance-local FlashInfer MLA counters during RDMA transfer

  fixsize/L
  #933·Ethan-ES opened Oct 25, 2026Contributor·

  ·
  3

• feat(refit): add Azure publication and startup support

  featsize/XXL
  #931·ipochi opened Oct 23, 2026Contributor·

  ·
  5

• fix(refit): apply the classic CUDA pool to the registered buffer's device

  fixsize/XL
  #930·yafshar opened Oct 23, 2026Contributor·

  ·
  3

• fix(client): reuse one gRPC channel across an artifact transfer

  fixsize/L
  #928·JiaweiGithub opened Oct 23, 2026·

  ·
  1

  5

• Exclude distributed communicator state from peer weight transfers

  size/M
  #921·jason136 opened Oct 21, 2026Contributor·

  ·
  5

• fix(client): treat unset source_load as unknown in topology_aware

  fixsize/M
  #913·v0ropaev opened Oct 20, 2026·

  ·
  3

• fix(s3): reject '.' and '..' as the bucket of an s3:// model name

  fixsize/S
  #912·v0ropaev opened Oct 20, 2026·

  ·
  4