The top 5 Hugging Face models and datasets by download growth over the 1-day window ending 7 November 2026, measured by RiseFinder.
| Rank | Name | Before → after | Change |
|---|---|---|---|
| 1 | TacVerse/TacVerse | 1,637 → 5,596 downloads | +241.8% |
| 2 | 294-288-pretrain-interaction/160M-exp | 4,477 → 8,563 downloads | +91.3% |
| 3 | UniParser/RxnMiner-Benchmark | 309 → 519 downloads | +68% |
| 4 | Axion1-0/Axion-1.0-35B | 596 → 815 downloads | +36.7% |
| 5 | bitmanagerai-speech-lab/darbuka_arabic_tts_dataset | 1,758 → 2,158 downloads | +22.8% |
TacVerse is a large-scale dataset for bimanual manipulation research in which every frame carries four streams of high-resolution visuotactile sensing, two per gripper, alongside synchronized camera views. It was collected with a handheld UMI gripper pair carrying Xense Neutron visuotactile sensors across three real-world scene families: office, home, and workbench.
Hugging Face dataset
Open source ↗This model represents five 160M GPT-NeoX training runs. These runs utilized separate M/C/G source pools within Dolma3-16B. The run directories, such as Clean, M500, C500, G500, and MCG500, contain all existing step-.pt checkpoints. One checkpoint is the final step-122071.pt, which was reached after 16,000,090,112 tokens.
Hugging Face model
Open source ↗UniParser/RxnMiner-Benchmark is a benchmark dataset for evaluating systems that extract complete, structured chemical reactions from literature figures. It covers reaction schemes, substrate-scope tables with R-group templates, condition-optimization tables, multistep routes and total syntheses. Each figure has a manually annotated, expert-reviewed reference answer listing every reaction drawn or implied, with exact molecular identities such as SMILES and stereochemistry.
Hugging Face dataset
Open source ↗Axion-1.0-35B is a large language model checkpoint that reproduces Axion 1.0-35B following the recipe described in Axion 1.0: Open Agent Foundation Models, Simplified. The model is designed for scientific reasoning, information seeking, and tool-augmented problem solving, meaning it can combine its own reasoning with external tools when working through problems.
Hugging Face model
Open source ↗This dataset is intended for research in the fields of speech recognition and text-to-speech applications. It compiles Arabic speech samples along with their corresponding automatic transcripts and audio-quality measurements. The data collection has a specific focus on capturing speech patterns from Egyptian and Saudi speakers.
Hugging Face dataset
Open source ↗Get early signals on the websites, apps, repositories and startups gaining momentum. Join the list for RiseFinder email briefings.
By subscribing, you ask us to store your email and send Risefinder updates. You can unsubscribe at any time. Privacy.
Want a sample first? Read past daily emails.