netron
openpilot
netron | openpilot | |
---|---|---|
42 | 860 | |
31,292 | 55,933 | |
0.8% | 0.6% | |
9.8 | 9.9 | |
3 days ago | 5 days ago | |
JavaScript | Python | |
MIT License | MIT License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
netron
-
World Emulation via DNN
The model seems to be viewable here:
https://netron.app/?url=https://madebyoll.in/posts/world_emu...
- Show HN: Mycelium – A graph viewer library for neural networks
-
How to develop an AI app with a local model in Kotlin Multiplatform
Note the size of the scrollbars. The model is huge. You can explore it using Netron. In the real app, you would use some ready, pre-trained model.
-
ONNX: The Open Standard for Seamless Machine Learning Interoperability
You can open and visualize any ONNX model in this website: https://netron.app
-
Essential Deep Learning Checklist: Best Practices Unveiled
How to Accomplish: Use architecture visualization tools like NETRON or TensorBoard to inspect the model architecture. These tools provide a graphical representation of the model, making it easier to understand its structure and parameter count. Adjust the model complexity based on performance benchmarks and resource constraints, aiming for the simplest model that achieves the desired performance.
-
Giving Odin Intelligence
use handy visualizers, for example, https://netron.app/
- Google Edge AI Model Explorer
-
Your 14-Day Free Trial Ain't Gonna Cut It
They're data-dependence graphs for a neural-network scheduling problem. Like this but way bigger to start with and then lowered to more detailed representations several times: https://netron.app/?url=https://github.com/onnx/models/raw/m... My home-grown layout engine can handle the 12k nodes for llama2 in its highest-level form in 20s or so, but its not the most featureful, and they only get bigger from there. So I always have an eye out for potential tools.
-
What's the best PyTorch model visualization tool?
Netron seems to be the best that I've seen so far. https://github.com/lutzroeder/netron
- Visualizer for neural network, deep learning and machine learning models
openpilot
-
Ask HN: Do custom ROMs exist for electric cars, for example Teslas?
I don't know about electric cars, but for gas powered cars there are open source ECUs [0][1]. There are also tuners that directly modify the car's firmware to improve performance. Finally, you can connect a computer to the CAN bus [2], which allows you to capture and replay commands, as well as craft your own commands. This is how Comma's openpilot [3] works: it connects to the CAN bus and sends commands for all supported functionality.
What kind of features did you have in mind?
[0] https://en.wikipedia.org/wiki/Engine_control_unit
[1] https://rusefi.com/
[2] https://en.wikipedia.org/wiki/CAN_bus
[3] https://comma.ai/openpilot
-
I couldn't submit a PR, so I got hired and fixed it myself
Well, that's a bit of my time gone (re)looking into GeoHot, patent trolls, and now comma.ai.
Comma.AI by George Hotz sounds very interesting, it's basically a $999 "comma 3x" smartphone with an OBD-II connector and a $99 wiring harness that can add an equivalent of a Tesla Autopilot to many cars manufactured in the last 10 years (even Tesla's own cars, too), for a total cost of $1098, whilst being OSS and available on GitHub, and — get this — even having ssh access to your car! Optional cloud subscription plans are $10/mo for your own SIM, or $24/mo with bundled cellular data.
Sadly, it does NOT have an equivalent of Tesla Sentry Mode yet, https://github.com/commaai/openpilot/issues/29912, which is kind of unfortunate, because Tesla's own implementation of Sentry Mode is using 250W of power — depleting the entire 80kWh battery from 80% to 30% in like 7 days (".5*80kWh over 7 days" = 238W) — openpilot would have been a nice alternative at what'd presumably be around 5W or less ("40kWh / 5W" is 333 days).
-
Are We the Baddies?
The author should start with his own product, which stops working if it doesn't connect to the internet at least every 14 days.
https://github.com/commaai/openpilot/blob/a8ec08e5bbc2be0a32...
-
Autonomous drone defeats human champions in racing first
Yes. https://www.psu.edu/news/campus-life/story/hackers-who-remot...
Even without vulnerabilities like that, something like https://comma.ai/openpilot could very likely be used in the same way ArduPilot was used in the recent Ukrainian drone attacks.
-
Comma 3X: Initial Impressions
It's really not hard to just search "openpilot" and click the first link. There's also a link at the top of the comma.ai website that links to the same page.
https://comma.ai/openpilot
openpilot is an open source advanced driver assistance system that works on 300+ car models of Toyota, Hyundai, Honda, and many other brands.
Automated Lane Centering
-
Comma AI has sold 10K devices
(from https://github.com/commaai/openpilot/blob/master/README.md)
> I might or might not have fallen asleep one night on a long road trip and awoken to my car driving along happily
My nightmare is getting smashed to pieces by people like you and your insurance refusing to pay because you found it funny to fiddle with your car. You being bankrupt will not help me.
- Comma.ai – Make Driving Chill
- Every Hands-Free Driving System Available in 2024
-
Openpilot – Operating system for robotics. 275 supported cars
For those similarly curious, I went sniffing around the PRs and found that, yes, there are two .onnx (via git-lfs) in https://github.com/commaai/openpilot/tree/v0.9.7/selfdrive/m... - one for the forward camera and one for watching the driver. The driver one had a lot more features than I expected, but I'd guess they are related to the NTSB certification I saw mentioned elsewhere in the thread here. Regrettably, because I don't have a Panda-compatible car, I actually can't try out the self-hosting setup but it certainly seems plausible
Also, while poking around, I saw they use a "CTF" for on-boarding new contributors, which I thought was neat: https://github.com/commaai/openpilot/blob/v0.9.7/tools/CTF.m...
-
Comma.ai: Refactoring for Growth
> While changing lanes, openpilot is not capable of looking next to you or checking your blind spot. Only nudge the wheel to initiate a lane change after you have confirmed it's safe to do so. [1]
Seems they are up-front about it.
[1] https://github.com/commaai/openpilot/blob/master/docs/LIMITA...
What are some alternatives?
ncnn - ncnn is a high-performance neural network inference framework optimized for the mobile platform
sunnypilot - sunnypilot is an open source driver assistance system. sunnypilot offers the user a unique driving experience for over 300 supported car makes and models with modified behaviors of driving assist engagements. sunnypilot complies with the safety policy from comma.ai's openpilot as accurately as possible.
models - Models and examples built with TensorFlow
dragonpilot - dragonpilot - 基於 openpilot 的開源駕駛輔助系統
PlotNeuralNet - Latex code for making neural networks diagrams
opendbc - a Python API for your car