Open Weights Breakout
🧱
58
Play a Breakout game featuring 50 company logo bricks
🔄 In a Training Loop
John Johnson
jjokah
AI & ML interests
Natural Language Processing, AI Engineering, Small Language Models.
Recent Activity
liked a Space 2 days ago
burtenshaw/open-weights-breakout upvoted a paper 25 days ago
Vera: A Layered Diffusion Model for Content-Preserving Video EditingOrganizations
liked a
Space 2 days ago
upvoted an article 4 days ago
Article
Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident
- +2
hlarcher, XciD, raphael-gl, chris-rannou
• • 421 upvoted a paper 25 days ago
upvoted a paper 3 months ago
upvoted a changelog 4 months ago
Hugging Face Changelog
Storage Buckets for Spaces
• 146
upvoted an article 4 months ago
Article
Welcome Gemma 4: Frontier multimodal intelligence on device
- +5
merve, pcuenq, sergiopaniego, burtenshaw, Steveeeeeeen, alvarobartt, SaylorTwift
• • 918 upvoted a collection 4 months ago
upvoted a paper 5 months ago
commented on Small Language Models (SLM): A Comprehensive Overview 5 months ago
Here on Hugging Face (models section) use the Parameters filter and select the size range of the model you want.
commented on Small Language Models (SLM): A Comprehensive Overview 5 months ago
There has to be a trade-off here to really shrink the model size.
The model has to be scoped to specific tasks or domains rather than trying to be general-purpose.
And a good approach to achieve this is Knowledge Distillation: train a tiny student model on a specific task to mimic a larger teacher, but you lose generality. That's the trade-off.
upvoted 2
articles 6 months ago
Article
One-Shot Any Web App with Gradio's gr.HTML
- +1
ysharma, hysts, freddyaboulton
• • 35 Article
Introducing Daggr: Chain apps programmatically, inspect visually
- +3
merve, ysharma, abidlabs, hysts, pcuenq
• • 107 liked a
model 6 months ago
liked a
Space 6 months ago
liked a
model 7 months ago
upvoted 2
collections 7 months ago
reacted to their post with 🔥 7 months ago
Post
1114
TranslateGemma: Open Translation Models (Jan 2026)
Google introduces TranslateGemma, a new suite of open translation models based on Gemma 3, available in 4B, 12B, and 27B parameter sizes.
Key Highlights:
• Supports 55 languages with high-quality translation across high-, mid-, and low-resource languages
• Exceptional efficiency: 12B model outperforms 27B baseline on WMT24++ benchmark
• Built using two-stage fine-tuning process distilling knowledge from Gemini models
• Retains strong multimodal capabilities (can translate text within images)
• Trained on nearly 500 additional language pairs for research adaptation
• Designed for diverse deployment environments from mobile to cloud
The models achieve state-of-the-art performance while maintaining exceptional efficiency, making high-quality translation accessible across different devices and use cases.
https://huggingface.co/collections/google/translategemma
Google introduces TranslateGemma, a new suite of open translation models based on Gemma 3, available in 4B, 12B, and 27B parameter sizes.
Key Highlights:
• Supports 55 languages with high-quality translation across high-, mid-, and low-resource languages
• Exceptional efficiency: 12B model outperforms 27B baseline on WMT24++ benchmark
• Built using two-stage fine-tuning process distilling knowledge from Gemini models
• Retains strong multimodal capabilities (can translate text within images)
• Trained on nearly 500 additional language pairs for research adaptation
• Designed for diverse deployment environments from mobile to cloud
The models achieve state-of-the-art performance while maintaining exceptional efficiency, making high-quality translation accessible across different devices and use cases.
https://huggingface.co/collections/google/translategemma
posted an update 7 months ago
Post
1114
TranslateGemma: Open Translation Models (Jan 2026)
Google introduces TranslateGemma, a new suite of open translation models based on Gemma 3, available in 4B, 12B, and 27B parameter sizes.
Key Highlights:
• Supports 55 languages with high-quality translation across high-, mid-, and low-resource languages
• Exceptional efficiency: 12B model outperforms 27B baseline on WMT24++ benchmark
• Built using two-stage fine-tuning process distilling knowledge from Gemini models
• Retains strong multimodal capabilities (can translate text within images)
• Trained on nearly 500 additional language pairs for research adaptation
• Designed for diverse deployment environments from mobile to cloud
The models achieve state-of-the-art performance while maintaining exceptional efficiency, making high-quality translation accessible across different devices and use cases.
https://huggingface.co/collections/google/translategemma
Google introduces TranslateGemma, a new suite of open translation models based on Gemma 3, available in 4B, 12B, and 27B parameter sizes.
Key Highlights:
• Supports 55 languages with high-quality translation across high-, mid-, and low-resource languages
• Exceptional efficiency: 12B model outperforms 27B baseline on WMT24++ benchmark
• Built using two-stage fine-tuning process distilling knowledge from Gemini models
• Retains strong multimodal capabilities (can translate text within images)
• Trained on nearly 500 additional language pairs for research adaptation
• Designed for diverse deployment environments from mobile to cloud
The models achieve state-of-the-art performance while maintaining exceptional efficiency, making high-quality translation accessible across different devices and use cases.
https://huggingface.co/collections/google/translategemma
upvoted a collection 8 months ago