All pull requests

Pull requests

• fix(server): restore Docling "<word>sign" glyph names that docling-parse leaves unmapped

  #961·krisztian-gajdar opened Nov 13, 2026Contributor·

  ·
  1

  4

• feat(server): add an RTX PRO 6000 32K thinking-bounded RLM profile for Qwen3.8-27B-FP8

  #949·svonava opened Nov 12, 2026Contributor·

  ·
  2

• fix(models): remove the bge-m3 banking profile, whose LoRA has no effect

  #948·krisztian-gajdar opened Nov 11, 2026Contributor·

  ·
  1

• fix(mcp): make sie-mcp work against SIE Serverless and redact every copy of a PII value

  #936·rparundekar opened Nov 9, 2026Contributor·4 tasks done·

  ·
  1

• chore(deps): bump the uv group across 29 directories with 3 updates

  dependenciespython:uv
  #920·dependabot[bot] opened Nov 7, 2026Bot·

  ·
  1

• chore(deps): bump the uv group across 3 directories with 4 updates

  dependenciespython:uv
  #917·dependabot[bot] opened Nov 7, 2026Bot·

  ·
  1

• fix(server): use image-only LightOnOCR-2 prompts by default

  #916·svonava opened Nov 6, 2026Contributor·

  ·
  1

• chore(deps): bump msgpack from 1.2.1 to 1.2.3 in /examples/wine-recommender in the pip group across 1 directory

  dependenciespython
  #893·dependabot[bot] opened Nov 5, 2026Bot·

  ·
  1

• fix(sdk): retry 503 LORA_LOADING on score and extract, like encode

  #883·paolo-perrone opened Nov 2, 2026·

  ·
  4

• fix(sdk): refuse non-uint8 image arrays instead of sending a black or white JPEG

  #876·paolo-perrone opened Nov 2, 2026·

  ·
  3

• feat(examples): compare Luna low reasoning on the router workload

  #861·svonava opened Oct 27, 2026Contributor·

  ·
  4

• chore(deps): bump the uv group across 49 directories with 5 updates

  dependenciespython:uv
  #860·dependabot[bot] opened Oct 26, 2026Bot·

  ·
  5

• chore(deps): bump the pip group across 5 directories with 3 updates

  dependenciespython
  #859·dependabot[bot] opened Oct 26, 2026Bot·

  ·
  1

• fix(server): tokenize the Qwen3 reranker row as the model card does

  #838·svonava opened Oct 20, 2026Contributor·

  ·
  1

• feat: run an SGLang LLM and encoders on one GPU in one pool (composite bundles)

  #836·svonava opened Oct 20, 2026Contributor·

  ·
  9

• fix: restart a worker with a wedged CUDA context in seconds

  #835·svonava opened Oct 20, 2026Contributor·

  ·
  4

• feat(server): keep evicted models in host RAM and wake them instead of reloading

  #834·svonava opened Oct 20, 2026Contributor·

  ·
  13

• feat(sglang): size more engines automatically and add shared-GPU LLM profiles

  #832·svonava opened Oct 20, 2026Contributor·

  ·
  5

• feat(gateway)!: let API pools live up to 30 days or never expire

  #831·svonava opened Oct 20, 2026Contributor·

  ·
  6

• perf(server): avoid masked hidden-state copy in CPU fp32 mean pooling

  #822·phiomegaruntime-rgb opened Oct 19, 2026Contributor·

  ·
  1

• fix(helm): schedule dcgm-exporter on GPU nodes outside GKE

  #810·caohhung opened Oct 19, 2026Contributor·

  ·
  43

• fix(examples): support single-arm rerank URL replay

  #707·svonava opened Oct 12, 2026Contributor·

  ·
  9

• fix(server): preserve Qwen reranker semantics without Flash Attention

  #693·svonava opened Oct 11, 2026Contributor·

  ·
  5

• feat(server): add opt-in exclude_labels and require_uppercase to GLiNER extraction

  #664·svonava opened Oct 9, 2026Contributor·

  ·
  1

• feat(server): add pinned OntoNotes entity extraction

  #661·svonava opened Oct 9, 2026Contributor·

  ·
  1