Hugging Face's logo

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up

All HF Hub posts
burtenshaw 
posted an update 2 days ago
view post
Post
4539
new smol course

If you’re building with or learning about post training AI models right now, we have a new FREE and CERTIFIED course.

🔗 Follow the org to join in smol-course

The course builds on smol course v1 which was the fastest way to learn to train your custom AI models. It now has:

- A leaderboard for students to submit models to
- Certification based on exams and leaderboards
- Prizes based on Leaderboards
- Up to date content on TRL and SmolLM3
- Deep integration with the Hub’s compute for model training and evaluation

We will release chapters every few weeks, so you can follow the org to stay updated.
  • 2 replies
·
tomaarsen 
posted an update 1 day ago
view post
Post
1954
ModernBERT goes MULTILINGUAL! One of the most requested models I've seen, The Johns Hopkins University's CLSP has trained state-of-the-art massively multilingual encoders using the ModernBERT architecture: mmBERT.

Model details:
- 2 model sizes:
- jhu-clsp/mmBERT-small
- jhu-clsp/mmBERT-base
- Uses the ModernBERT architecture, but with the Gemma2 multilingual tokenizer (so: flash attention, alternating global/local attention, unpadding/sequence packing, etc.)
- Maximum sequence length of 8192 tokens, on the high end for encoders
- Trained on 1833 languages using DCLM, FineWeb2, and many more sources
- 3 training phases: 2.3T tokens pretraining on 60 languages, 600B tokens mid-training on 110 languages, and 100B tokens decay training on all 1833 languages.
- Both models are MIT Licensed, and the full datasets and intermediary checkpoints are also publicly released

Evaluation details:
- Very competitive with ModernBERT at equivalent sizes on English (GLUE, MTEB v2 English after finetuning)
- Consistently outperforms equivalently sized models on all Multilingual tasks (XTREME, classification, MTEB v2 Multilingual after finetuning)
- In short: beats commonly used multilingual base models like mDistilBERT, XLM-R (multilingual RoBERTa), multilingual MiniLM, etc.
- Additionally: the ModernBERT-based mmBERT is much faster than the alternatives due to its architectural benefits. Easily up to 2x throughput in common scenarios.

Check out the full blogpost with more details. It's super dense & gets straight to the point: https://huggingface.co/blog/mmbert

Based on these results, mmBERT should be the new go-to multilingual encoder base models at 300M and below. Do note that the mmBERT models are "base" models, i.e. they're currently only trained to perform Mask Filling. They'll need to be finetuned for downstream tasks like semantic search, classification, clustering, etc.
Reubencf 
posted an update 2 days ago
view post
Post
3848
Introducing the Nano Banana Node Editor! 🍌

Now you can control and manipulate Nano Banana images with a powerful, intuitive node-based system. Explore the creative possibilities at: Reubencf/Nano_Banana_Editor

This version is clearer, more inviting, and emphasizes the creative potential of your tool.
  • 2 replies
·
andywu-kby 
posted an update 3 days ago
view post
Post
2823
Hello everyone
Good day!

We have launched the product - Virtual Try On 🚀
Say goodbye to the uncertainty of online shopping with Miragic’s Virtual Try-On solution! Our cutting-edge AI technology lets you try on clothes virtually, offering a seamless and interactive shopping experience. Whether you're exploring new outfits or simply trying before you buy, Miragic gives you a realistic view of how items will look on you—without ever stepping into a store.

Miragic-AI/Miragic-Virtual-Try-On

🌟 Key Features:
- Realistic 3D Try-On: See how clothes fit and look on your virtual self in real-time.
- Personalized Fit: Using advanced body-scanning tech, Miragic adjusts the fit based on your unique measurements.
- Wide Fashion Selection: Browse through various brands and styles, all available for a virtual try-on.
- Sustainable Shopping: Reduce the need for returns and make more eco-friendly choices with a virtual experience that helps you shop smarter.

👚 Why Virtual Try-On?
- Save time and money while shopping smarter.
- Discover new styles, fit options, and combinations in a way that’s fast and fun.
- Enjoy a unique, tech-driven shopping experience from the comfort of your home!

Join us today and transform the way you shop online with Virtual Try-On.
·
Kseniase 
posted an update 3 days ago
view post
Post
6729
10 Latest Preference Optimization Techniques

Models need feedback on what makes outputs “good” or “bad.” Policy optimization (PO) turns preferences and rewards into actual training signals. This field is evolving quickly, moving far beyond classics like PPO and GRPO. So here is our overview of 10 newest PO methods:

1. Pref-GRPO → Pref-GRPO: Pairwise Preference Reward-based GRPO for Stable Text-to-Image Reinforcement Learning (2508.20751)
Stabilizes text-to-image reinforcement learning (RL) with pairwise preference rewards and a unified UNIGENBENCH benchmark

2. PVPO (Policy with Value Preference Optimization) → PVPO: Pre-Estimated Value-Based Policy Optimization for Agentic Reasoning (2508.21104)
This critic-free RL method uses a pre-trained model as a reference anchor to reduce bias and guide learning, selecting high-value examples through data pre-sampling

3. DCPO (Dynamic Clipping Policy Optimization) → DCPO: Dynamic Clipping Policy Optimization (2509.02333)
Uses dynamic clipping, which adjusts probability limits per token for better token exploration, and smooth reward standardization to balance rewards over training steps and prevent wasted updates

4. ARPO (Agentic Reinforced Policy Optimization) → Agentic Reinforced Policy Optimization (2507.19849)
Optimizes multi-turn LLM agents that use external tools. It uses an entropy-based adaptive rollout to explore post-tool use and an advantage attribution method to better assign credit across steps, leading to more efficient tool use with fewer resources

5. GRPO-RoC (Group Relative Policy Optimization with Resampling-on-Correct) → rStar2-Agent: Agentic Reasoning Technical Report (2508.20722)
Oversamples rollouts, then resamples them to keep diverse mistakes and only the highest-quality correct answers. It reduces noises and ends up with stronger reasoning in a code environment

Read further below ⬇️
If you like this, also subscribe to the Turing post: https://www.turingpost.com/subscribe
  • 1 reply
·
prithivMLmods 
posted an update about 20 hours ago
view post
Post
1529
Build something cool with Nano Banana aka Gemini 2.5 Flash Image AIO [All-in-One]. Draw and transform on canvas, edit images, and generate images—all in one place!🍌

✦︎ Constructed with the Gemini API (GCP). Try it here: https://nano-banana-aio-op72ohwdda-uw.a.run.app/

⚠️ Note: The server’s health status is currently stable, but this may change at any time. If you experience network issues, please refresh the current app tab or trigger the discussion below.
merve 
posted an update 2 days ago
salma-remyx 
posted an update about 20 hours ago
hesamation 
posted an update 6 days ago
view post
Post
4802
a senior engineer at google just dropped a 400-page free book on docs for review: agentic design patterns.

the table of contents looks like everything you need to know about agents + code:
> advanced prompt techniques
> multi-agent patterns
> tool use and MCP
> you name it

read it here: https://docs.google.com/document/d/1rsaK53T3Lg5KoGwvf8ukOUvbELRtH-V0LnOIFDxBryE/edit?tab=t.0#heading=h.pxcur8v2qagu

you can also pre-order on Amazon (published by Springer) and the royalties goes to Save the Children: https://www.amazon.com/Agentic-Design-Patterns-Hands-Intelligent/dp/3032014018/
Abhaykoul 
posted an update 1 day ago
view post
Post
422
🚀 Ever dreamed of training your own Large Language Model from scratch? What if I told you it doesn't require a supercomputer or PhD in ML? 🤯

Introducing LLM Trainer - the educational framework that makes LLM training accessible to EVERYONE! Whether you're on a CPU-only laptop or scaling to distributed GPUs, we've got you covered. 💻➡️🖥️

Why LLM Trainer? Because existing tools are either too simplistic (hiding the magic) or too complex (requiring expert knowledge). We bridge the gap with:

🎓 Educational transparency - every component built from scratch with clear code
💻 CPU-first approach - start training immediately, no GPU needed
🔧 Full customization - modify anything you want
📈 Seamless scaling - from laptop to cluster without code changes
🤝 HuggingFace integration - works with existing models & tokenizers

Key highlights:
✅ Built-in tokenizers (BPE, WordPiece, HF wrappers)
✅ Complete Transformer implementation from scratch
✅ Optimized for CPU training
✅ Advanced features: mixed precision, gradient checkpointing, multiple generation strategies
✅ Comprehensive monitoring & metrics

Perfect for:
- Students learning transformers
- Researchers prototyping new ideas
- Developers building domain-specific models

Ready to train your first LLM? It's easier than you think!

🔗 Check it out: https://github.com/HelpingAI/llm-trainer
📚 Docs: Getting Started Guide
💬 Join the community: GitHub Discussions

#AI #MachineLearning #LLM #DeepLearning #OpenSource #Python #HuggingFace #NLP

Special thanks to HuggingFace and PyTorch teams for the amazing ecosystem! 🙏
Recently active users