Skip to content

v0.2.0

Choose a tag to compare

@mamayer19 mamayer19 released this 09 Jun 09:27
· 148 commits to main since this release

Highlights

Added ModernBERT-based embedding models and LightOnOCR, along with startup model preloading and explicit concurrency control in the asynchronous Python client.

Features

  • Added GTE-ModernBERT-base, Snowflake Arctic Embed M v2.0, and IBM Granite English R2 embedding models, including the small variant.
  • Added LightOnOCR-2-1B for OCR in the transformers5 bundle.
  • Added max_concurrency to SIEAsyncClient and Haystack-convention import aliases under haystack_integrations.
  • Added anonymous usage telemetry, with opt-out through SIE_TELEMETRY_DISABLED=true or DO_NOT_TRACK=1.

Bug fixes

  • Model-affinity routing can spill requests to other workers instead of becoming stuck, and rejected requests now contribute to autoscaling demand.

Breaking changes

  • Worker startup no longer accepts --model to select models. Use --preload or SIE_PRELOAD_MODELS to load models at startup; otherwise models load on demand.