24 109 105

Tolga Cangöz

tolgacangoz

standard_ai

AI & ML interests

AIGC

Recent Activity

liked a model about 11 hours ago

diffusers/Wan2.1-VAE

reacted to tomaarsen's post with ❤️ about 11 hours ago

An assembly of 18 European companies, labs, and universities have banded together to launch 🇪🇺 EuroBERT! It's a state-of-the-art multilingual encoder for 15 European languages, designed to be finetuned for retrieval, classification, etc. 🇪🇺 15 Languages: English, French, German, Spanish, Chinese, Italian, Russian, Polish, Portuguese, Japanese, Vietnamese, Dutch, Arabic, Turkish, Hindi 3️⃣ 3 model sizes: 210M, 610M, and 2.1B parameters - very very useful sizes in my opinion ➡️ Sequence length of 8192 tokens! Nice to see these higher sequence lengths for encoders becoming more common. ⚙️ Architecture based on Llama, but with bi-directional (non-causal) attention to turn it into an encoder. Flash Attention 2 is supported. 🔥 A new Pareto frontier (stronger *and* smaller) for multilingual encoder models 📊 Evaluated against mDeBERTa, mGTE, XLM-RoBERTa for Retrieval, Classification, and Regression (after finetuning for each task separately): EuroBERT punches way above its weight. 📝 Detailed paper with all details, incl. data: FineWeb for English and CulturaX for multilingual data, The Stack v2 and Proof-Pile-2 for code. Check out the release blogpost here: https://huggingface.co/blog/EuroBERT/release * https://huggingface.co/EuroBERT/EuroBERT-210m * https://huggingface.co/EuroBERT/EuroBERT-610m * https://huggingface.co/EuroBERT/EuroBERT-2.1B The next step is for researchers to build upon the 3 EuroBERT base models and publish strong retrieval, zero-shot classification, etc. models for all to use. I'm very much looking forward to it!

upvoted an article 1 day ago

What changed in the Transformer architecture

View all activity

Organizations

tolgacangoz's activity

liked a model about 11 hours ago

diffusers/Wan2.1-VAE

Updated about 12 hours ago • 2 • 1

reacted to tomaarsen's post with ❤️ about 11 hours ago

Post

4361

An assembly of 18 European companies, labs, and universities have banded together to launch 🇪🇺 EuroBERT! It's a state-of-the-art multilingual encoder for 15 European languages, designed to be finetuned for retrieval, classification, etc.

🇪🇺 15 Languages: English, French, German, Spanish, Chinese, Italian, Russian, Polish, Portuguese, Japanese, Vietnamese, Dutch, Arabic, Turkish, Hindi
3️⃣ 3 model sizes: 210M, 610M, and 2.1B parameters - very very useful sizes in my opinion
➡️ Sequence length of 8192 tokens! Nice to see these higher sequence lengths for encoders becoming more common.
⚙️ Architecture based on Llama, but with bi-directional (non-causal) attention to turn it into an encoder. Flash Attention 2 is supported.
🔥 A new Pareto frontier (stronger *and* smaller) for multilingual encoder models
📊 Evaluated against mDeBERTa, mGTE, XLM-RoBERTa for Retrieval, Classification, and Regression (after finetuning for each task separately): EuroBERT punches way above its weight.
📝 Detailed paper with all details, incl. data: FineWeb for English and CulturaX for multilingual data, The Stack v2 and Proof-Pile-2 for code.

Check out the release blogpost here: https://huggingface.co/blog/EuroBERT/release
* EuroBERT/EuroBERT-210m
* EuroBERT/EuroBERT-610m
* EuroBERT/EuroBERT-2.1B

The next step is for researchers to build upon the 3 EuroBERT base models and publish strong retrieval, zero-shot classification, etc. models for all to use. I'm very much looking forward to it!