Playing with Veo3 this morning. Share your prompt if you want me to create videos for you (bonus point if they funnily reference HF/open-source). These videos are "a cat on the moon rapping "I love Hugging Face""!
Despite the emergence of combining LLM and DiT architectures for T2I synthesis, its design remains severely understudied.
This was done long ago and got into CVPR25 -- super excited to finally share it now, along with the data and code ♥️
We explore several architectural choices that affect this design. We provide an open & reproducible training recipe that works at scale.
Works like Playground v3 have already explored a deep fusion between an LLM and a DiT, sharing their representations through layerwise attention. They exhibit excellent performance on T2I.
Despite its compelling results and other performance virtues, it remains unexplored, which is what we want to improve in our work. Specifically, we take a pre-trained LLM (Gemma-2B) and trainable DiT, and set out to explore what makes a "good deep fusion" between the two for T2I.
We explore several key questions in the work, such as:
Q1: How should we do attention? We considered several alternatives. PixArt-Alpha like attention (cross-attention) is very promising. Q2: Should we incorporate additional text modulation? Q3: Can we eliminate timestep conditioning? Q4: How do we do positional encodings? Q5: Do instruction-tuned LLMs help deep fusion? Q6: Would using a decoder LLM from a multimodal model be helpful? Q7: Does using a better variant of Gemma help?
Based on the above findings, we arrive at FuseDiT with the following components on top of the base architecture from the findings of our experiments.
* No AdaLN-Zero modules * 1D + 2D-RoPE * Gemma 2 2B, adjusting DiT configurations accordingly
We trained FuseDiT on a mixture from CC12M, JourneyDB, & SA (~26M image-text pairs) for 800 steps. While not the best model, it's encouraging to develop something in a guided manner using open datasets.
To know more (code, models, all are available), please check out the paper: https://lnkd.in/gg6qyqZX.
New in smolagents v1.16.0: 🔍 Bing support in WebSearchTool 🐍 Custom functions & executor_kwargs in LocalPythonExecutor 🔧 Streaming GradioUI fixes 🌐 Local web agents via api_base & api_key 📚 Better docs
SmolVLM-2 and SigLIP-2 are now part of transformers in dedicated releases!
They're added on top of the v4.49.0 release, and can be installed from the following tags: v4.49.0-SmolVLM-2 and v4.49.0-SigLIP-2.
This marks a new beginning for the release process of transformers. For the past five years, we've been doing monthly releases featuring many models (v4.49.0, the latest release, features 9 new architectures).
Starting with SmolVLM-2 & SigLIP2, we'll now additionally release tags supporting new models on a stable branch. These models are therefore directly available for use by installing from the tag itself. These tags will continue to be updated with fixes applied to these models.
Going forward, continue expecting software releases following semantic versioning: v4.50.0 will have ~10 new architectures compared to v4.49.0, as well as a myriad of new features, improvements and bug fixes. Accompanying these software releases, we'll release tags offering brand new models as fast as possible, to make them accessible to all immediately.
We are reproducing the full DeepSeek R1 data and training pipeline so everybody can use their recipe. Instead of doing it in secret we can do it together in the open!
🧪 Step 1: replicate the R1-Distill models by distilling a high-quality reasoning corpus from DeepSeek-R1.
🧠 Step 2: replicate the pure RL pipeline that DeepSeek used to create R1-Zero. This will involve curating new, large-scale datasets for math, reasoning, and code.
🔥 Step 3: show we can go from base model -> SFT -> RL via multi-stage training.
🌟 Digital Odyssey: AI Image & Video Generation Platform 🎨 Welcome to our all-in-one AI platform for image and video generation! 🚀 ✨ Key Features
🎨 High-quality image generation from text 🎥 Video creation from still images 🌐 Multi-language support with automatic translation 🛠️ Advanced customization options
💫 Unique Advantages
⚡ Fast and accurate results using FLUX.1-dev and Hyper-SD models 🔒 Robust content safety filtering system 🎯 Intuitive user interface 🛠️ Extended toolkit including image upscaling and logo generation
🎮 How to Use
Enter your image or video description Adjust settings as needed Click generate Save and share your results automatically
Introducing 📐𝐅𝐢𝐧𝐞𝐌𝐚𝐭𝐡: the best public math pre-training dataset with 50B+ tokens! HuggingFaceTB/finemath
Math remains challenging for LLMs and by training on FineMath we see considerable gains over other math datasets, especially on GSM8K and MATH.
We build the dataset by: 🛠️ carefully extracting math data from Common Crawl; 🔎 iteratively filtering and recalling high quality math pages using a classifier trained on synthetic annotations to identify math reasoning and deduction.
We conducted a series of ablations comparing the performance of Llama-3.2-3B-Base after continued pre-training on FineMath and observe notable gains compared to the baseline model and other public math datasets.
We hope this helps advance the performance of LLMs on math and reasoning! 🚀 We’re also releasing all the ablation models as well as the evaluation code.
After some heated discussion 🔥, we clarify our intent re. storage limits on the Hub
TL;DR: - public storage is free, and (unless blatant abuse) unlimited. We do ask that you consider upgrading to PRO and/or Enterprise Hub if possible - private storage is paid above a significant free tier (1TB if you have a paid account, 100GB otherwise)
