Issues

• Late interaction normalization produces nonfinite float16 embeddings

  qdrant/fastembed#825 · Filter by author IlyaKhalafi opened on Oct 15, 2026

• Worker crashes go unnoticed while waiting for a result

  qdrant/fastembed#823 · Filter by author IlyaKhalafi opened on Oct 15, 2026

• Non-positive worker counts wait for results without starting workers

  qdrant/fastembed#817 · Filter by author IlyaKhalafi opened on Oct 14, 2026

• Failed worker pool construction raises an extra destructor exception

  qdrant/fastembed#814 · Filter by author IlyaKhalafi opened on Oct 14, 2026

• [Bug]: rerank_pairs raises TypeError when pairs are passed as a tuple of pairs

  qdrant/fastembed#796 · Filter by author michaelkillgta opened on Oct 9, 2026

• [Feature]: Support huggingface-hub 2.0

  qdrant/fastembed#742 · Filter by author joein opened on Sep 28, 2026
  

• Silent all-NaN embeddings when a model file is incompatible with the runtime (0.8.0 still ships the fp16 jina-de model fixed by #623)

  qdrant/fastembed#688 · Filter by author MaximilianBuenning opened on Aug 24, 2026
  1

• Resolving License Issues

  qdrant/fastembed#679 · Filter by author scheung-lbg opened on Aug 18, 2026
  1

• [Bug]: Segmentation Fault or AssertionError during initialization on Python 3.14.2

  qdrant/fastembed#618 · Filter by author ElissonRodrigues opened on Mar 26, 2026
  4

• [Model]: perplexity-ai/pplx-embed-v1-0.6b

  qdrant/fastembed#613 · Filter by author fvisticot opened on Mar 15, 2026
  1

• [Feature]: Support Chinese Tokenization for BM25 using Jieba

  qdrant/fastembed#610 · Filter by author sandsc opened on Mar 6, 2026
  2

• GPU users: onnxruntime (CPU) overwrites onnxruntime-gpu binaries when both are installed by pip/uv

  qdrant/fastembed#608 · Filter by author michelkluger opened on Feb 25, 2026
  1

• [Model]: jina v5

  qdrant/fastembed#607 · Filter by author mohamad-tohidi opened on Feb 25, 2026
  1

• [Feature]: RFC: Biological embeddings support

  qdrant/fastembed#597 · Filter by author nleroy917 opened on Jan 12, 2026

• [Feature]: support qwen3-vl

  qdrant/fastembed#596 · Filter by author wey-gu opened on Jan 10, 2026

• [Feature]: Add custom token corpus for other languages for bm25

  qdrant/fastembed#594 · Filter by author Nevermetyou65 opened on Dec 19, 2025
  1

• [Feature] Support nomic-ai/CodeRankEmbed model

  qdrant/fastembed#587 · Filter by author ku-nal opened on Dec 5, 2025
  1

• [Feature]: add docs for exposed onnx session options, cache_dir and local_files_only

  qdrant/fastembed#579 · Filter by author joein opened on Nov 25, 2025
  1

• [Feature]: Add CLI to launch FastEmbed as an Embedding API service (OpenAI-compatible)

  qdrant/fastembed#571 · Filter by author magicbrighter opened on Nov 5, 2025
  2

• [Bug]: downloaded embedding models should not go to a temporary folder by default

  qdrant/fastembed#569 · Filter by author vemonet opened on Oct 31, 2025
  3

• Sparse Embeddings (BM25) and Persian Language Support

  qdrant/fastembed#558 · Filter by author alibaghernejad opened on Aug 28, 2025
  2

• Does converting model to onnx format cause performance degradation of model?

  qdrant/fastembed#554 · Filter by author hoangthvn2201 opened on Aug 20, 2025
  5

• how could fastembed load the local model?

  qdrant/fastembed#549 · Filter by author helxsz opened on Aug 7, 2025
  4

• [Bug]: SparseTextEmbedding needs to implement aembed for non-blocking use (async/await)

  qdrant/fastembed#546 · Filter by author sglebs opened on Jul 24, 2025
  2

• [Feature]: miniCOIL: no way to control input sequence length (and VRAM usage) natively

  qdrant/fastembed#544 · Filter by author elia-morrison opened on Jul 16, 2025
  1