The top 10 Hugging Face models and datasets by download growth over the 7-day window ending 9 November 2026, measured by RiseFinder.
| Rank | Name | Before → after | Change |
|---|---|---|---|
| 1 | LiquidAI/LFM3-Omni-2B | 8,376 → 16,547 downloads | +97.6% |
| 2 | Diffuspace/Clover-35B-A3B-Preview | 2,352 → 4,061 downloads | +72.7% |
| 3 | RecursiveSI/flint | 26,703 → 44,352 downloads | +66.1% |
| 4 | Vyber07/cyber-security | 8,571 → 13,494 downloads | +57.4% |
| 5 | moondream/refcoco-m | 1,046 → 1,455 downloads | +39.1% |
| 6 | bones-studio/seed | 8,485 → 11,419 downloads | +34.6% |
| 7 | ShapeNet/ShapeNetCore | 11,459 → 15,409 downloads | +34.5% |
| 8 | hongzhuyi/VDEBench | 6,305 → 7,680 downloads | +21.8% |
| 9 | openbmb/JevOCRBench | 920 → 1,100 downloads | +19.6% |
| 10 | venvoo/china-a-share-l2-level2-limit-order-book-tick-data | 183,337 → 218,724 downloads | +19.3% |
LiquidAI/LFM3-Omni-2B is an AI model that handles multiple kinds of input and output in a single system. It understands text, images, video and audio, and responds in text or speech. It is built for on-device use, replacing what would otherwise require a separate ASR model, vision-language model, LLM and TTS model.
Hugging Face model
Open source ↗Clover-35B-A3B-Preview is a diffusion language model and the first model released by DiffuSpace. It combines a 35B-parameter mixture-of-experts backbone with approximately 3B active parameters. At release, it delivered state-of-the-art performance among open-source diffusion language models on the benchmarks reported in the launch blog.
Hugging Face model
Open source ↗Flint is a mixture-of-experts language model made by Recursive for use with coding agents. It has 39.8 billion total parameters with approximately 2.5 billion active per token. It is designed for software engineering work such as investigating codebases, fixing bugs, implementing features, and writing tests.
Hugging Face model
Open source ↗Vyber07/cyber-security is a cybersecurity knowledge base dataset intended for AI training. It covers cybersecurity domains including offensive red teaming, bug bounty exploitation, defensive SOC operations, digital forensics, and AI/LLM security. The dataset is 16 GB in size, contains 507 files, and spans more than 30 domains drawn from more than 15 platforms.
Hugging Face dataset
Open source ↗This dataset serves as a benchmark for referring expression segmentation. While it has historically been a standard benchmark, the evidence notes that it contains issues such as poor mask quality and harmful referring expressions. Furthermore, modern models are capable of producing masks that are more accurate than the ground-truth annotations provided in the dataset, which suggests the dataset may be an imprecise measure.
Hugging Face dataset
Open source ↗bones-studio/seed, also called BONES-SEED (Skeletal Everyday Embodiment Dataset), is an open dataset of 142,220 annotated human motion animations intended for humanoid robotics. It contains motion capture data in SOMA and Unitree G1 formats, along with natural language descriptions, temporal segmentation, and detailed skeletal metadata.
Hugging Face dataset
Open source ↗ShapeNetCore is a dataset of annotated 3D models. This repository contains ShapeNetCore (v2), a subset of ShapeNet covering 55 common object categories with about 51,300 unique 3D models. Each model is linked to an appropriate synset in WordNet 3.0, and use of the data is subject to the ShapeNet terms of use.
Hugging Face dataset
Open source ↗VDEBench is a bilingual benchmark dataset for visual document editing, containing 3,506 examples split into 1,578 text edits and 1,928 table edits. Each example provides the input document image, an editing instruction, a target image, editing-region annotations, and recorded source provenance. Available masks indicate which regions of a document are allowed for editing.
Hugging Face dataset
Open source ↗JevOCRBench is a benchmark for OCR correctness decisions. It tests whether a decision model can tell whether an OCR result is correct. The dataset contains 2,267 Chinese and English records covering text, formulas, and tables, with each record mainly containing the source image and the OCR candidate text.
Hugging Face dataset
Open source ↗venvoo/china-a-share-l2-level2-limit-order-book-tick-data is a Hugging Face dataset containing a historical archive of Chinese exchange Level-2 market data from 2017 to 2026, supplied through a vendor export and stored in Parquet format. It includes ten-level quote snapshots, individual order messages, and trade-stream records, covering A-share stocks as well as non-stock instruments such as ETFs and bonds.
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.