RunsOnMine

RunsOnMine

A browser registry of LLM serving benchmarks measured on consumer GPUs, dual Tesla T4 and an RTX 3050 laptop, with concurrency scaling curves and a way to submit your own measurements.

Made by prajwalpujari16 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: RunsOnMine is a browser registry of LLM serving benchmarks measured on consumer GPUs. It holds 24 physical measurements from a pair of Tesla T4 cards and an RTX 3050 laptop, along with curves that show how performance scales as concurrency rises from 1 to 8. Visitors can also add their own measurements through a one-click pull request portal.

Who it suits: It suits people who want real numbers for serving language models on modest hardware, not vendor estimates.

Title: Who might use this registry
Industry | Who uses it | How they use it | Example
Hobbyist computing | Enthusiasts | Check what a home GPU can serve | Look up the RTX 3050 laptop results
Startups | Small teams | Judge whether cheap GPUs are enough | Read the dual T4 scaling curves
AI research | Researchers | Collect empirical serving data | Compare measurements across concurrency levels

If you want data on data-center GPUs, you can skip it, since the registry covers consumer hardware.

Where to use it: You can use it online or open its files on your own computer.

• Web browser: the Space's page on Hugging Face, with nothing to install
• Windows, macOS or Linux: the downloaded files, opened in a browser

Title: What the registry is good for
Use case | Who | What they get back
See real serving numbers | Enthusiasts | Measurements taken on physical hardware
Study concurrency scaling | Engineers | Curves for 1 through 8 concurrent requests
Share your own results | Contributors | A one-click path to submit a measurement

How to set it up: The hosted page needs nothing. To get the files you need Python and pip.

• Python, with pip
• Any web browser

1. Install the huggingface_hub package.
2. Download the Space's files into a folder.
3. Move into that folder.
4. Open index.html in a web browser.

pip install -U huggingface_hub
hf download prajwalpujari16/runsonmine --repo-type space --local-dir runsonmine
cd runsonmine

How to use it: Open the Space online, or open index.html from your downloaded folder. Then browse the measurements, look at the scaling curves and use the submission portal if you have results to add.

Title: What the page shows
Tool | What it does
Physical measurements | Lists 24 measurements on dual Tesla T4 and an RTX 3050 laptop
Concurrency scaling curves | Show results as concurrency goes from 1 to 8
PR submission portal | Offers a one-click way to propose a new measurement

Example: You open the registry and look at the dual T4 results. You follow a scaling curve from 1 concurrent request up to 8 to see how serving behaves as load grows.

For newcomers and experts: If you are new to this, start with the hardware closest to yours and read one curve. Concurrency here means how many requests the GPU handles at once.

If you know the field, the data is empirical and comes from physical machines, not from a model of them. The Space names Qwen2.5 0.5B Instruct and Llama 3.1 8B, and a consumer silicon benchmarks dataset.

Measure serving on a consumer GPU
Add the numbers to the registry
The page plots concurrency curves
You compare hardware and load

Good to know: The registry is small and covers two hardware setups.

• Coverage: 24 measurements on dual Tesla T4 and an RTX 3050 laptop
• Static page: plain HTML and CSS, with no server code
• Contributions: the portal opens a pull request for you
• License: apache-2
• Cost to run: none named
• Data it reaches: nothing beyond its own files is named in the notes, apart from the pull request you choose to send
• Not for: data-center GPU benchmarks

On Hugging Face, 13 November 2026

• Likes: 1
• Built with: static
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/prajwalpujari16/runsonmine
• Files: https://huggingface.co/spaces/prajwalpujari16/runsonmine/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.