1234aurel/MS
Updated
vasilcov
1234aurel
AI & ML interests
None yet
Recent Activity
updated
a model
about 1 hour ago
1234aurel/MS
updated
a model
about 1 hour ago
1234aurel/gfd
updated
a model
about 2 hours ago
1234aurel/hhhy
Organizations
None yet
1234aurel's activity
updated
2
models
about 1 hour ago
updated
a
model
about 2 hours ago
updated
a
model
about 11 hours ago
replied to
BrigitteTousi's
post
about 13 hours ago
reacted to
BrigitteTousi's
post with π€
about 13 hours ago
Post
1371
AI agents are transforming how we interact with technology, but how sustainable are they? π
Design choices β like model size and structure β can massively impact energy use and cost. β‘π° The key takeaway: smaller, task-specific models can be far more efficient than large, general-purpose ones.
π Open-source models offer greater transparency, allowing us to track energy consumption and make more informed decisions on deployment. π± Open-source = more efficient, eco-friendly, and accountable AI.
Read our latest, led by @sasha with assists from myself + @yjernite π€
https://huggingface.co/blog/sasha/ai-agent-sustainability
Design choices β like model size and structure β can massively impact energy use and cost. β‘π° The key takeaway: smaller, task-specific models can be far more efficient than large, general-purpose ones.
π Open-source models offer greater transparency, allowing us to track energy consumption and make more informed decisions on deployment. π± Open-source = more efficient, eco-friendly, and accountable AI.
Read our latest, led by @sasha with assists from myself + @yjernite π€
https://huggingface.co/blog/sasha/ai-agent-sustainability
-
- 1 reply
reacted to
sequelbox's
post with π
about 13 hours ago
Post
708
TITANIUM 2 Deepseek-R1 dataset is here! Open-source synthetic architecture and DevOps dataset:
sequelbox/Titanium2-DeepSeek-R1
Esper 3 will be coming out soon for multiple base models, trained on Titanium, Raiden, and more :)
with my love,
allegra
Esper 3 will be coming out soon for multiple base models, trained on Titanium, Raiden, and more :)
with my love,
allegra
reacted to
fdaudens's
post with β€οΈ
about 13 hours ago
Post
1480
I read the 456-page AI Index report so you don't have to (kidding). The wild part? While AI gets ridiculously more accessible, the power gap is actually widening:
1οΈβ£ The democratization of AI capabilities is accelerating rapidly:
- The gap between open and closed models is basically closed: difference in benchmarks like MMLU and HumanEval shrunk to just 1.7% in 2024
- The cost to run GPT-3.5-level performance dropped 280x in 2 years
- Model size is shrinking while maintaining performance - Phi-3-mini hitting 60%+ MMLU at fraction of parameters of early models like PaLM
2οΈβ£ But we're seeing concerning divides deepening:
- Geographic: US private investment ($109B) dwarfs everyone else - 12x China's $9.3B
- Research concentration: US and China dominate highly-cited papers (50 and 34 respectively in 2023), while next closest is only 7
- Gender: Major gaps in AI skill penetration rates - US shows 2.39 vs 1.71 male/female ratio
The tech is getting more accessible but the benefits aren't being distributed evenly. Worth thinking about as these tools become more central to the economy.
Give it a read - fascinating portrait of where AI is heading! https://hai-production.s3.amazonaws.com/files/hai_ai_index_report_2025.pdf
1οΈβ£ The democratization of AI capabilities is accelerating rapidly:
- The gap between open and closed models is basically closed: difference in benchmarks like MMLU and HumanEval shrunk to just 1.7% in 2024
- The cost to run GPT-3.5-level performance dropped 280x in 2 years
- Model size is shrinking while maintaining performance - Phi-3-mini hitting 60%+ MMLU at fraction of parameters of early models like PaLM
2οΈβ£ But we're seeing concerning divides deepening:
- Geographic: US private investment ($109B) dwarfs everyone else - 12x China's $9.3B
- Research concentration: US and China dominate highly-cited papers (50 and 34 respectively in 2023), while next closest is only 7
- Gender: Major gaps in AI skill penetration rates - US shows 2.39 vs 1.71 male/female ratio
The tech is getting more accessible but the benefits aren't being distributed evenly. Worth thinking about as these tools become more central to the economy.
Give it a read - fascinating portrait of where AI is heading! https://hai-production.s3.amazonaws.com/files/hai_ai_index_report_2025.pdf
- 1 reply
reacted to
luigi12345's
post with π
about 13 hours ago
Post
1391
π Metaβs Llama 4 Models Now on Hugging Face!
Meta has released Llama 4 Scout and Llama 4 Maverick, now available on Hugging Face:
β’ Llama 4 Scout: 17B active parameters, 16-expert Mixture of Experts (MoE) architecture, 10M token context window, fits on a single H100 GPU. οΏΌ
β’ Llama 4 Maverick: 17B active parameters, 128-expert MoE architecture, 1M token context window, optimized for DGX H100 systems. οΏΌ
π₯ Key Features:
β’ Native Multimodality: Seamlessly processes text and images. οΏΌ
β’ Extended Context Window: Up to 10 million tokens for handling extensive inputs.
β’ Multilingual Support: Trained on 200 languages, with fine-tuning support for 12, including Arabic, Spanish, and German. οΏΌ
π οΈ Access and Integration:
β’ Model Checkpoints: Available under the meta-llama organization on the Hugging Face Hub.
β’ Transformers Compatibility: Fully supported in transformers v4.51.0 for easy loading and fine-tuning.
β’ Efficient Deployment: Supports tensor-parallelism and automatic device mapping.
These models offer developers enhanced capabilities for building sophisticated, multimodal AI applications. οΏΌ
Meta has released Llama 4 Scout and Llama 4 Maverick, now available on Hugging Face:
β’ Llama 4 Scout: 17B active parameters, 16-expert Mixture of Experts (MoE) architecture, 10M token context window, fits on a single H100 GPU. οΏΌ
β’ Llama 4 Maverick: 17B active parameters, 128-expert MoE architecture, 1M token context window, optimized for DGX H100 systems. οΏΌ
π₯ Key Features:
β’ Native Multimodality: Seamlessly processes text and images. οΏΌ
β’ Extended Context Window: Up to 10 million tokens for handling extensive inputs.
β’ Multilingual Support: Trained on 200 languages, with fine-tuning support for 12, including Arabic, Spanish, and German. οΏΌ
π οΈ Access and Integration:
β’ Model Checkpoints: Available under the meta-llama organization on the Hugging Face Hub.
β’ Transformers Compatibility: Fully supported in transformers v4.51.0 for easy loading and fine-tuning.
β’ Efficient Deployment: Supports tensor-parallelism and automatic device mapping.
These models offer developers enhanced capabilities for building sophisticated, multimodal AI applications. οΏΌ
updated
a
Space
1 day ago
replied to
seawolf2357's
post
5 days ago
reacted to
seawolf2357's
post with π€β€οΈ
5 days ago
Post
6862
π¨ Ghibli-Style Image Generation with Multilingual Text Integration: FLUX.1 Hugging Face Edition πβ¨
Hello creators! Today I'm introducing a special image generator that combines the beautiful aesthetics of Studio Ghibli with multilingual text integration! π
seawolf2357/Ghibli-Multilingual-Text-rendering
β¨ Key Features
Ghibli-Style Image Generation - High-quality animation-style images based on FLUX.1
Multilingual Text Rendering - Support for Korean, Japanese, English, and all languages! π°π·π―π΅π¬π§
Automatic Image Editing with Simple Prompts - Just input your desired text and you're done!
Two Stylistic Variations Provided - Get two different results from a single prompt
Full Hugging Face Spaces Support - Deploy and share instantly!
π How Does It Work?
Enter a prompt describing your desired image (e.g., "a cat sitting by the window")
Input the text you want to add (any language works!)
Select the text position, size, and color
Two different versions are automatically generated!
π― Advantages of This Model
No Tedious Post-Editing Needed - Text is perfectly integrated during generation
Natural Text Integration - Text automatically adjusts to match the image style
Perfect Multilingual Support - Any language renders beautifully!
User-Friendly Interface - Easily adjust text size, position, and color
One-Click Hugging Face Deployment - Use immediately without complex setup
π Use Cases
Creating multilingual greeting cards
Animation-style social media content
Ghibli-inspired posters or banners
Character images with dialogue in various languages
Sharing with the community through Hugging Face Spaces
This project leverages Hugging Face's FLUX.1 model to open new possibilities for seamlessly integrating high-quality Ghibli-style images with multilingual text using just prompts! π
Try it now and create your own artistic masterpieces! π¨β¨
#GhibliStyle #MultilingualSupport #AIImageGeneration #TextRendering #FLUX #HuggingFace
Hello creators! Today I'm introducing a special image generator that combines the beautiful aesthetics of Studio Ghibli with multilingual text integration! π
seawolf2357/Ghibli-Multilingual-Text-rendering
β¨ Key Features
Ghibli-Style Image Generation - High-quality animation-style images based on FLUX.1
Multilingual Text Rendering - Support for Korean, Japanese, English, and all languages! π°π·π―π΅π¬π§
Automatic Image Editing with Simple Prompts - Just input your desired text and you're done!
Two Stylistic Variations Provided - Get two different results from a single prompt
Full Hugging Face Spaces Support - Deploy and share instantly!
π How Does It Work?
Enter a prompt describing your desired image (e.g., "a cat sitting by the window")
Input the text you want to add (any language works!)
Select the text position, size, and color
Two different versions are automatically generated!
π― Advantages of This Model
No Tedious Post-Editing Needed - Text is perfectly integrated during generation
Natural Text Integration - Text automatically adjusts to match the image style
Perfect Multilingual Support - Any language renders beautifully!
User-Friendly Interface - Easily adjust text size, position, and color
One-Click Hugging Face Deployment - Use immediately without complex setup
π Use Cases
Creating multilingual greeting cards
Animation-style social media content
Ghibli-inspired posters or banners
Character images with dialogue in various languages
Sharing with the community through Hugging Face Spaces
This project leverages Hugging Face's FLUX.1 model to open new possibilities for seamlessly integrating high-quality Ghibli-style images with multilingual text using just prompts! π
Try it now and create your own artistic masterpieces! π¨β¨
#GhibliStyle #MultilingualSupport #AIImageGeneration #TextRendering #FLUX #HuggingFace
-
- 3 replies
reacted to
jasoncorkill's
post with π§ π§
5 days ago
Post
2334
π Rapidata: Setting the Standard for Model Evaluation
Rapidata is proud to announce our first independent appearance in academic research, featured in the Lumina-Image 2.0 paper. This marks the beginning of our journey to become the standard for testing text-to-image and generative models. Our expertise in large-scale human annotations allows researchers to refine their models with accurate, real-world feedback.
As we continue to establish ourselves as a key player in model evaluation, weβre here to support researchers with high-quality annotations at scale. Reach out to info@rapidata.ai to see how we can help.
Lumina-Image 2.0: A Unified and Efficient Image Generative Framework (2503.21758)
Rapidata is proud to announce our first independent appearance in academic research, featured in the Lumina-Image 2.0 paper. This marks the beginning of our journey to become the standard for testing text-to-image and generative models. Our expertise in large-scale human annotations allows researchers to refine their models with accurate, real-world feedback.
As we continue to establish ourselves as a key player in model evaluation, weβre here to support researchers with high-quality annotations at scale. Reach out to info@rapidata.ai to see how we can help.
Lumina-Image 2.0: A Unified and Efficient Image Generative Framework (2503.21758)
replied to
jasoncorkill's
post
5 days ago
reacted to
jasoncorkill's
post with β€οΈ
5 days ago
Post
2554
This dataset was collected in roughly 4 hours using the Rapidata Python API, showcasing how quickly large-scale annotations can be performed with the right tooling!
All that at less than the cost of a single hour of a typical ML engineer in Zurich!
The new dataset of ~22,000 human annotations evaluating AI-generated videos based on different dimensions, such as Prompt-Video Alignment, Word for Word Prompt Alignment, Style, Speed of Time flow and Quality of Physics.
Rapidata/text-2-video-Rich-Human-Feedback
All that at less than the cost of a single hour of a typical ML engineer in Zurich!
The new dataset of ~22,000 human annotations evaluating AI-generated videos based on different dimensions, such as Prompt-Video Alignment, Word for Word Prompt Alignment, Style, Speed of Time flow and Quality of Physics.
Rapidata/text-2-video-Rich-Human-Feedback
-
- 1 reply
replied to
jasoncorkill's
post
5 days ago
reacted to
jasoncorkill's
post with π
5 days ago
Post
2726
We benchmarked
@xai-org
's Aurora model, as far as we know the first public evaluation of the model at scale.
We collected 401k human annotations in over the past ~2 days for this, we have uploaded all of the annotation data here on huggingface with a fully permissive license
Rapidata/xAI_Aurora_t2i_human_preferences
We collected 401k human annotations in over the past ~2 days for this, we have uploaded all of the annotation data here on huggingface with a fully permissive license
Rapidata/xAI_Aurora_t2i_human_preferences
-
- 1 reply
published
a
model
6 days ago