EmbeddingGemma 2: Unified and Efficient Multimodal Representations
Paper • 2610.37364 • Published • 35
Dev Mode Explorers
community
AI & ML interests
None defined yet.
Recent Activity
View all activity
osanseviero
authored a
paper 33 minutes ago
Post
223
👋Hey there folks ,
I wanted to use a Jev model in a practical way , and I came up with this OCR pipeline using decision-models to evaluate OCR quality .
Check it out :
Tonic/OCR-with-JEV
I wanted to use a Jev model in a practical way , and I came up with this OCR pipeline using decision-models to evaluate OCR quality .
Check it out :
Tonic/OCR-with-JEV
- 1 reply
Post
4,188
𝐈𝐧𝐭𝐫𝐨𝐝𝐮𝐜𝐢𝐧𝐠 🧬 𝐅𝐢𝐧𝐞𝐃𝐍𝐀: 𝐭𝐡𝐞 𝐛𝐞𝐬𝐭 𝐝𝐚𝐭𝐚𝐬𝐞𝐭 𝐭𝐨 𝐭𝐞𝐚𝐜𝐡 𝐚𝐧𝐲 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐦𝐨𝐝𝐞𝐥 𝐃𝐍𝐀.
Training Qwen3-14B on it produces Carbon-X, the strongest open model on DNA benchmarks, on par with GPT-6.1 Sol.
It contains:
- Midtraining: 300B+ tokens of genomic sequences, papers and web pages mixing DNA with text
- SFT: 200K samples, from variant effect prediction to sequence question answering
- RL: 98K prompts with verifiable rewards
Data, models and the full recipe are open.
HuggingFaceBio/FineDNA
Enjoy this late Halloween treat! 🎃
Training Qwen3-14B on it produces Carbon-X, the strongest open model on DNA benchmarks, on par with GPT-6.1 Sol.
It contains:
- Midtraining: 300B+ tokens of genomic sequences, papers and web pages mixing DNA with text
- SFT: 200K samples, from variant effect prediction to sequence question answering
- RL: 98K prompts with verifiable rewards
Data, models and the full recipe are open.
HuggingFaceBio/FineDNA
Enjoy this late Halloween treat! 🎃
- 5 replies
nielsr
submitted a
paper to Daily Papers 15 days ago
Post
3,785
Introducing Live Subtitles, a Windows tray app that captions anything your PC is playing and isn't butt ugly. live and fully local. It runs NVIDIA's Parakeet-unified 0.6B on your GPU, puts words on screen about 0.36 s after they're spoken, and uses less than one CPU core while it does it.
I built a benchmark for it out of real TV, podcasts, earnings calls and a short film. Parakeet-unified scores 12.1% WER there, close to Nemotron streaming's 11.4%, but with much better punctuation and lower worst-case delay. Once a caption is on screen it never gets rewritten or jumps lines, and the overlay disappears when nobody's talking.
No NVIDIA GPU? It falls back to Nemotron on the CPU, which also covers 30 other languages. It can also write SRT/ASS/VTT files for videos on disk.
Windows only. The GPU engine needs an NVIDIA card with ~3.2 GB of free VRAM. MIT.
Edit: the first version of this post said 13.8% vs 17.0%. The replies caught a scoring bug (numbers weren't normalized), so the benchmark now uses Whisper's English normalizer and the numbers above are re-scored.
Get it here:
https://github.com/Nymbo/Live-Subtitles
Models:
nvidia/parakeet-unified-en-0.6b
nvidia/nemotron-speech-streaming-en-0.6b
I built a benchmark for it out of real TV, podcasts, earnings calls and a short film. Parakeet-unified scores 12.1% WER there, close to Nemotron streaming's 11.4%, but with much better punctuation and lower worst-case delay. Once a caption is on screen it never gets rewritten or jumps lines, and the overlay disappears when nobody's talking.
No NVIDIA GPU? It falls back to Nemotron on the CPU, which also covers 30 other languages. It can also write SRT/ASS/VTT files for videos on disk.
Windows only. The GPU engine needs an NVIDIA card with ~3.2 GB of free VRAM. MIT.
Edit: the first version of this post said 13.8% vs 17.0%. The replies caught a scoring bug (numbers weren't normalized), so the benchmark now uses Whisper's English normalizer and the numbers above are re-scored.
Get it here:
https://github.com/Nymbo/Live-Subtitles
Models:
nvidia/parakeet-unified-en-0.6b
nvidia/nemotron-speech-streaming-en-0.6b
- 7 replies
Post
5,304
@Banaxi-Tech stop hiding my comments. AND STOP STEALING PAPERS AND SPREADING MISINFORMATION.
your BGA blog is a copy of NSA (deepseek, 2025) branded under your name. literally the same top16 selected blocks, 512 local window, router over block summaries, all you did was change block size from 64 to 128.
you didnt cite NSA once but you put a “please cite BGA” bibtex at the bottom.
i commented under your post and said that there is no way that you can support claims like: “The Accuracy Should BE WAy better than DSA but untested yet.” you didnt run a single experiment. and the 256x isnt from BGA, its just n/2k with k=2048 so the exact same k DSA uses. if opus wrote this for you, at least read it before posting.
i commented again after you hid my comment despite it having constructive and correct feedback and you hid that too. and again.
you can hide the truth and just try to get hf post likes..... but is it really the thing that needs to be done? do you really want to take papers and make them yours while barely even changing the params?
admitting your mistakes and doing something about them needs humbleness, intelligence, humanness.
i encourage you to admit your mistakes and try to do better next time (at least read what blog your ai wrote or do proper experiments to back your stuff up).
your BGA blog is a copy of NSA (deepseek, 2025) branded under your name. literally the same top16 selected blocks, 512 local window, router over block summaries, all you did was change block size from 64 to 128.
you didnt cite NSA once but you put a “please cite BGA” bibtex at the bottom.
i commented under your post and said that there is no way that you can support claims like: “The Accuracy Should BE WAy better than DSA but untested yet.” you didnt run a single experiment. and the 256x isnt from BGA, its just n/2k with k=2048 so the exact same k DSA uses. if opus wrote this for you, at least read it before posting.
i commented again after you hid my comment despite it having constructive and correct feedback and you hid that too. and again.
you can hide the truth and just try to get hf post likes..... but is it really the thing that needs to be done? do you really want to take papers and make them yours while barely even changing the params?
admitting your mistakes and doing something about them needs humbleness, intelligence, humanness.
i encourage you to admit your mistakes and try to do better next time (at least read what blog your ai wrote or do proper experiments to back your stuff up).
- 49 replies
Post
162
SlopSoup TV. A live, never-ending 24/7 pixel-art TV network inspired by old-school late night Adult Swim. No human makes any of it.
Tune in anytime: https://huggingface.co/proxy/severian-slopsoup.hf.space or https://www.youtube.com/live/Bikb9JGy6vg?si=YPpIqLjW6y9yML6m
20+ shows in the spirit of late-night Adult Swim:
- an hourly news desk with exhausted overnight anchors
- a game show with a reigning champion
- small-claims court and a cursed cooking show
- a talk show hosted by a houseplant or a microwave
- a 1 a.m. call-in hotline
- an ’80s sitcom on a decaying VHS tape
- 4 a.m. infomercials that slide into existential dread
- wordless generative music videos
- many more shows and content added daily...
Fully AI-run: AI writes the scripts, invents new characters every episode, designs each one a unique voice, directs the camera, cuts the edit, schedules the day and runs the broadcast.
Nothing ever airs twice: a memory of everything that’s aired blocks repeated premises, jokes and characters, so every minute is new.
Real facts, delivered weird: the news, trivia and explainers are sourced and cited; the comedy is the chaos around them.
Its own engine: a custom pixel renderer with a virtual camera, lip-synced close-ups and 50+ animated sets, all running in a browser on one cloud box.
It’s weird, it’s for adults, and it never stops. Happy to answer questions about how it works.
Tune in anytime: https://huggingface.co/proxy/severian-slopsoup.hf.space or https://www.youtube.com/live/Bikb9JGy6vg?si=YPpIqLjW6y9yML6m
20+ shows in the spirit of late-night Adult Swim:
- an hourly news desk with exhausted overnight anchors
- a game show with a reigning champion
- small-claims court and a cursed cooking show
- a talk show hosted by a houseplant or a microwave
- a 1 a.m. call-in hotline
- an ’80s sitcom on a decaying VHS tape
- 4 a.m. infomercials that slide into existential dread
- wordless generative music videos
- many more shows and content added daily...
Fully AI-run: AI writes the scripts, invents new characters every episode, designs each one a unique voice, directs the camera, cuts the edit, schedules the day and runs the broadcast.
Nothing ever airs twice: a memory of everything that’s aired blocks repeated premises, jokes and characters, so every minute is new.
Real facts, delivered weird: the news, trivia and explainers are sourced and cited; the comedy is the chaos around them.
Its own engine: a custom pixel renderer with a virtual camera, lip-synced close-ups and 50+ animated sets, all running in a browser on one cloud box.
It’s weird, it’s for adults, and it never stops. Happy to answer questions about how it works.
- 2 replies
Post
353
tokenizers 0.23.3 is out!https://github.com/huggingface/tokenizers/releases/tag/v0.23.3
The release is python-only and fixes a few very annoying issues
alielfilali01
posted an update about 1 month ago
Post
597
multi-agents orchestration is old news by now. Try multi-teams of agents ! that is a whole different nightmare ... This is the real signal about the sparks of AGI.
Post
322
I put together a Space to showcase tokenization-in-the-browser libs we maintain at HF
Let me know what you think!
sbrandeis/tokenizers-wasm-demo
Let me know what you think!
sbrandeis/tokenizers-wasm-demo
nielsr
submitted a
paper to Daily Papers about 2 months ago
mrfakename
posted an update about 2 months ago
Post
2,694
We’ve been working with LAION on a voice acting arena. You listen to two models doing the same scene and compare how well they pull it off.
It’s ready to try now - would love to hear what you think 🙂
TTS-AGI/voice-acting-arena
It’s ready to try now - would love to hear what you think 🙂
TTS-AGI/voice-acting-arena
Post
2,962
SPECK 2 IS ALREADY OUT: specklabs/Speck2-140M
Pretrained on 4x more tokens than the previous releases (20b vs 5b).
Instruct tuned versions are coming soon.
Very interesting models are coming soon too (hint: super long context).
Thanks for everyone supporting!
Pretrained on 4x more tokens than the previous releases (20b vs 5b).
Instruct tuned versions are coming soon.
Very interesting models are coming soon too (hint: super long context).
Thanks for everyone supporting!
- 3 replies
Post
166
NEW SPECK UPDATES:
Just hit #14 and #15 with out FIRST models on Open SLM Leaderboard. The models were trained on 5B tokens, while competing with similarly sized models trained on more than 6-20x the data.
A new base model Speck1.5-140M being trained right now on a higher quality corpus and will be released soon.
SpeckChat3 is coming very soon with 1 million samples, specifically designed to post train small base models.
Also, just to clarify stuff, we will NOT release anything that is NOT MIT licensed EVER. Openness is needed in small language research.
Thanks to everyone supporting the project, and stay tuned for new releases!
Just hit #14 and #15 with out FIRST models on Open SLM Leaderboard. The models were trained on 5B tokens, while competing with similarly sized models trained on more than 6-20x the data.
A new base model Speck1.5-140M being trained right now on a higher quality corpus and will be released soon.
SpeckChat3 is coming very soon with 1 million samples, specifically designed to post train small base models.
Also, just to clarify stuff, we will NOT release anything that is NOT MIT licensed EVER. Openness is needed in small language research.
Thanks to everyone supporting the project, and stay tuned for new releases!
Post
1,900
SPECK UPDATES:
1 New instruct model tuned on top of Speck1-140M: specklabs/Speck1-140M-Instruct
2 Instruction tuning datasets
2 GGUFs
Much more coming soon:
Speck1.1-140M-Instruct that is post trained on SpeckChat2 will be coming very soon
New base model Speck1.5-140M is coming with a much higher quality corpus
Thanks to everyone who is already supporting the project, and stay tuned for new releases!
1 New instruct model tuned on top of Speck1-140M: specklabs/Speck1-140M-Instruct
2 Instruction tuning datasets
2 GGUFs
Much more coming soon:
Speck1.1-140M-Instruct that is post trained on SpeckChat2 will be coming very soon
New base model Speck1.5-140M is coming with a much higher quality corpus
Thanks to everyone who is already supporting the project, and stay tuned for new releases!
- 3 replies
nielsr
submitted a
paper to Daily Papers 2 months ago
Post
2,135
FIRST SPECK MODEL RELEASED:
specklabs/Speck1-140M
new models coming very soon (both instruct and much better models), with much much higher training scale as i am getting marenostrum5 access soon!
we will be looking at 100b-2t token budgets :)
specklabs/Speck1-140M
new models coming very soon (both instruct and much better models), with much much higher training scale as i am getting marenostrum5 access soon!
we will be looking at 100b-2t token budgets :)
- 4 replies
nielsr
submitted 2
papers to Daily Papers 3 months ago