Tom Aarsen's picture

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

liked a model about 4 hours ago

HuggingFaceTB/SmolVLM-256M-Instruct

new activity about 4 hours ago

sentence-transformers/all-MiniLM-L6-v2:Should I be concerned about "UserWarning: TypedStorage is deprecated. " when using the model with python 3.11?

posted an update about 5 hours ago

I just released Sentence Transformers v3.4.0, featuring a memory leak fix, compatibility between the powerful Cached... losses and the Matryoshka loss modifier, and a bunch of fixes & small features. 🪆 Matryoshka & Cached loss compatibility It is now possible to combine the powerful Cached... losses (which use in-batch negatives & a caching mechanism to allow for endless batch size & negatives) with the Matryoshka loss modifier which modifies a base loss such that it is trained not only on the maximum dimensionality (e.g. 1024 dimensions), but also on many lower dimensions (e.g. 768, 512, 256, 128, 64, 32). After training, these models' embeddings can be truncated for faster retrieval, etc. 🎞️ Resolve memory leak when Model and Trainer are reinitialized Due to a circular dependency between Trainer -> Model -> ModelCardData -> Trainer, deleting both the trainer & model still didn't free up the memory. This led to a memory leak in scripts where you repeatedly do so. ➕ New Features Many new small features, e.g. multi-GPU support for 'mine_hard_negatives', a 'margin' parameter to TripletEvaluator, and Matthews Correlation Coefficient in the BinaryClassificationEvaluator. 🐛 Bug Fixes Also a bunch of fixes, for example that subsequent batches were not sorted when using the "no_duplicates" batch sampler. See the release notes for more details. Full release notes: https://github.com/UKPLab/sentence-transformers/releases/tag/v3.4.0 Big thanks to all community members who assisted in this release. 10 folks with their first contribution this time around!

View all activity

Articles

Train 400x faster Static Embedding Models with Sentence Transformers

Finally, a Replacement for BERT: Introducing ModernBERT

Welcome Gemma 2 - Google's new open LLM

Training and Finetuning Embedding Models with Sentence Transformers v3

Blazing Fast SetFit Inference with 🤗 Optimum Intel on Xeon

Binary and Scalar Embedding Quantization for Significantly Faster & Cheaper Retrieval

🪆 Introduction to Matryoshka Embedding Models

SetFitABSA: Few-Shot Aspect Based Sentiment Analysis using SetFit

🕳️ Attention Sinks in LLMs for endless fluency

Organizations

tomaarsen's activity

upvoted an article about 9 hours ago

Article

SmolVLM Grows Smaller – Introducing the 250M & 500M Models!

about 21 hours ago

• 35

upvoted an article about 12 hours ago

Article

Mastering Long Contexts in LLMs with KVPress

By

•

about 13 hours ago

• 42

upvoted an article 1 day ago

Article

Hugging Face and FriendliAI partner to supercharge model deployment on the Hub

2 days ago

• 26

upvoted an article 3 days ago

Article

Fine-tune ModernBERT for RAG with Synthetic Data

By

•

3 days ago

• 28

upvoted a paper 3 days ago

CodeXEmbed: A Generalist Embedding Model Family for Multiligual and Multi-task Code Retrieval

Paper • 2411.12644 • Published Nov 19, 2024 • 3

upvoted an article 7 days ago

Article

Timm ❤️ Transformers: Use any timm model with transformers

8 days ago

• 33

upvoted an article 8 days ago

Article

Train 400x faster Static Embedding Models with Sentence Transformers

9 days ago

• 119

upvoted an article 10 days ago

Article

Python Is All You Need? Introducing Dria-Agent-α

By

•

13 days ago

• 22

upvoted a paper 11 days ago

Agentless: Demystifying LLM-based Software Engineering Agents

Paper • 2407.01489 • Published Jul 1, 2024 • 59

upvoted 2 papers 16 days ago

Trans-Tokenization and Cross-lingual Vocabulary Transfers: Language Adaptation of LLMs for Low-Resource NLP

Paper • 2408.04303 • Published Aug 8, 2024 • 16

Fietje: An open, efficient LLM for Dutch

Paper • 2412.15450 • Published Dec 19, 2024 • 4

upvoted an article 17 days ago

Article

Announcing NVIDIA Cosmos World Foundation Models

By

•

17 days ago

• 23

upvoted a paper 17 days ago

KaLM-Embedding: Superior Training Data Brings A Stronger Embedding Model

Paper • 2501.01028 • Published 22 days ago • 13

upvoted a collection 17 days ago

KaLM-embedding

7 items • Updated about 14 hours ago • 22

upvoted 2 collections 19 days ago

OLMo 2

Artifacts for the second set of OLMo models. • 22 items • Updated 17 days ago • 74

PubMedBERT Embeddings M2V

Models distilled with Model2Vec - 100K / 500K / 1M / 2M / 8M parameter variants. • 5 items • Updated 16 days ago • 3

upvoted a collection 20 days ago

ModernGLiNER

GLiNER models based on modern encoder architectures • 2 items • Updated about 1 month ago • 6

upvoted an article 24 days ago

Article

Fine-tune ModernBERT for text classification using synthetic data

By

•

24 days ago

• 26

upvoted a collection 26 days ago

Granite 3.1 Language Models

A series of language models with 128K context length trained by IBM licensed under Apache 2.0 license. • 8 items • Updated Dec 18, 2024 • 48

upvoted a paper about 1 month ago

Spectrum: Targeted Training on Signal to Noise Ratio

Paper • 2406.06623 • Published Jun 7, 2024 • 13