A Qwen3-4B text embedder trained with bagging-based model merging for OOD-robust retrieval.
A workable 4B-parameter dense embedding model from ICT-CAS TIME / Querit. Treat the modality benchmarks above as the leading indicator of fit — composite scoring across modalities is still maturing.
Generated from this model’s benchmarks and ranking signals. Editor reviews refine it over time.
Access model weights, configuration files, and documentation.
See which devices can run this model and at what quality level.
A Qwen3-4B-based general text embedding model from the TIME group at ICT-CAS in collaboration with Querit. Applies BOOM (Bagging-based rObust mOdel Merging), training multiple embedding models on bootstrap-sampled subsets of the multi-task corpus and linearly merging them, to deliver ensemble-like robustness and OOD generalization at single-model inference cost; supports lightweight incremental updates via merge-in.
View on Hugging Face