1883 270 857

Tom Aarsen

tomaarsen

https://linkedin.com/in/tomaarsen

AI & ML interests

NLP: text embeddings, information retrieval, named entity recognition, few-shot text classification

Recent Activity

new activity about 2 hours ago

opensearch-project/opensearch-neural-sparse-encoding-v2-distill:Use the DistilBertTokenizer for this DistilBERT-based model

new activity about 2 hours ago

opensearch-project/opensearch-neural-sparse-encoding-doc-v2-distill:Use the DistilBertTokenizer for this DistilBERT-based model

upvoted a paper 2 days ago

MIEB: Massive Image Embedding Benchmark

View all activity

Organizations

tomaarsen's activity

New activity in opensearch-project/opensearch-neural-sparse-encoding-v2-distill about 2 hours ago

Use the DistilBertTokenizer for this DistilBERT-based model

#6 opened about 2 hours ago by

tomaarsen

New activity in opensearch-project/opensearch-neural-sparse-encoding-doc-v2-distill about 2 hours ago

Use the DistilBertTokenizer for this DistilBERT-based model

#4 opened about 2 hours ago by

tomaarsen

upvoted a paper 2 days ago

MIEB: Massive Image Embedding Benchmark

Paper • 2504.10471 • Published 3 days ago • 14

posted an update 2 days ago

Post

2228

I just released Sentence Transformers v4.1; featuring ONNX and OpenVINO backends for rerankers offering 2-3x speedups and improved hard negatives mining which helps prepare stronger training datasets. Details:

🏎️ ONNX, OpenVINO, Optimization, Quantization
- I've added ONNX and OpenVINO support with just one extra argument: "backend" when loading the CrossEncoder reranker, e.g.: CrossEncoder("cross-encoder/ms-marco-MiniLM-L6-v2", backend="onnx")
- The export_optimized_onnx_model, export_dynamic_quantized_onnx_model, and export_static_quantized_openvino_model functions now work with CrossEncoder rerankers, allowing you to optimize (e.g. fusions, gelu approximations, etc.) or quantize (int8 weights) rerankers.
- I've uploaded ~340 ONNX & OpenVINO models for all existing models under the cross-encoder Hugging Face organization. You can use these without having to export when loading.

⛏ Improved Hard Negatives Mining
- Added 'absolute_margin' and 'relative_margin' arguments to mine_hard_negatives.
- absolute_margin ensures that sim(query, negative) < sim(query, positive) - absolute_margin, i.e. an absolute margin between the negative & positive similarities.
- relative_margin ensures that sim(query, negative) < sim(query, positive) * (1 - relative_margin), i.e. a relative margin between the negative & positive similarities.
- Inspired by the excellent NV-Retriever paper from NVIDIA.

And several other small improvements. Check out the full release notes here: https://github.com/UKPLab/sentence-transformers/releases/tag/v4.1.0

With this release, I introduce near-feature parity between the SentenceTransformer embedding & CrossEncoder reranker models, which I've wanted to do for quite some time! With rerankers very strongly supported now, it's time to look forward to other useful architectures!

liked 2 models 2 days ago

BAAI/bge-reranker-v2-m3

Text Classification • Updated Jun 24, 2024 • 1.53M • 610

BAAI/bge-reranker-base

Text Classification • Updated Jun 24, 2024 • 1.17M • • 183

New activity in cross-encoder/ms-marco-MiniLM-L12-v2 2 days ago

Push tokenizer again

#7 opened 9 days ago by

tomaarsen

updated 11 models 2 days ago