Model Pulse

Model Pulse

Shows daily download history, likes and derivative families for models on the Hugging Face Hub since July 2024, with a model comparison chart and weekly rankings. It suits model builders and researchers.

Made by tardellirs on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: Model Pulse tracks how models on the Hugging Face Hub are downloaded and liked. It keeps daily history going back to July 2024. It also separates direct use from use through models derived from it.

Who it suits: Model builders, researchers and analysts who want to see which models are growing and which families dominate. It is an independent project and has no tie to Hugging Face.

Title: Who watches model trends
Industry | Who uses it | How they use it | Example
AI research | Researchers | Track how a model's downloads and likes change over time | Look up a model and read its daily history
Model publishing | Model authors | See how much of their reach comes from quantizations and fine-tunes | Check a base model's derivative family
Analysis | Analysts and writers | Put as many as five models side by side on a single chart | Chart several competing models together
Community | Hub users | Read the weekly rankings for new breakouts | See which models grew fastest this week

If you need download numbers for something other than Hub models, you can skip it.

Where to use it:

• A web browser: the Space runs on Hugging Face with nothing to install
• Windows, macOS or Linux, by opening the downloaded page in a browser
• The README of a model, through the badge it offers

Use case | Who | What they get back
Look up one model | Researchers | Daily downloads, likes and derivative families
Compare models | Analysts | Up to five models drawn on one chart
Watch the weekly rankings | Community members | Lists of the most downloaded, fastest growing and new breakouts
Add a badge to a README | Model authors | Monthly downloads with a sparkline, updated daily

How to set it up: To use it online you need nothing. To keep a copy of the page you need Python and pip for the download.

• A web browser
• Python with pip, only to download a copy

1. Install the Hugging Face hub package and download the Space into a folder.
2. Move into that folder.
3. Load index.html into your browser.

pip install -U huggingface_hub
hf download tardellirs/model-pulse --repo-type space --local-dir model-pulse
cd model-pulse

How to use it: Open the page and browse the rankings, or ask for one model by adding a model parameter to the page address in the form below. A second parameter picks a special view, either galaxy or wrapped.

?model=org/name
?view=galaxy
?view=wrapped

Tool | What it does
Model lookup | Shows one model's daily download history, likes and derivative families
Compare | Draws up to five models on one chart
Derivative share | Shows how much reach comes from quantizations, fine-tunes, adapters and merges
Galaxy | Draws every model built on a base model as a galaxy
Wrapped | Summarizes one author's past year on the Hub as six cards
Weekly rankings | Ranks models by downloads, speed of growth, new breakouts and family size, and ranks organizations

Example: You look up Qwen/Qwen3-8B. You see its daily downloads and likes since July 2024, and how much of its reach comes from quantizations, fine-tunes, adapters and merges of it.

For newcomers and experts: If you are new to this, open the Space and read the weekly rankings first. Then look up one model you already use and see how its downloads have moved.

If you know the field, the numbers come from daily snapshots of cfahlgren1/hub-stats, compiled into the dataset modelpulse/model-pulse-data. Splitting a model's reach by quantizations, fine-tunes, adapters and merges separates direct use from use through derivatives. Download counts are the Hub's own, so they carry whatever limits those counts have.

Daily snapshots of Hub stats
Compiled into one dataset
The page reads that dataset
You see history, rankings and families

Good to know:

• Independence: made by an outside author with no link to Hugging Face
• Range: daily history goes back to July 2024
• License: apache-2
• Cost to run: none named
• Data it reaches: the two Hugging Face datasets it names, cfahlgren1/hub-stats and modelpulse/model-pulse-data

On Hugging Face, 13 November 2026

• Likes: 214
• Built with: static
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/tardellirs/model-pulse
• Files: https://huggingface.co/spaces/tardellirs/model-pulse/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.