Загрузка видео...

Не удалось загрузить видео

Возникла проблема при загрузке этого видео. Это может быть связано с временными проблемами сети или видео может быть недоступно.

На главную

Here it is! I rebuilt a visual iniertial multicamera pipeline in rust + python, alll with support for Rerun OSS data catalog + visualization. code here - < The core is rust to make sure it runs fast, but you can also easily call it from python. I've been... show more

Pablo Vela

3,146 subscribers

17,561 просмотров • 10 дней назад •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

Комментарии: 12

Фото профиля Pablo Vela

Pablo Vela10 дней назад

1. Multi Camera - I'm convinced 4 or more cameras are optimal here, sure you can do purely monocular or stereo, but IMO most robots and actual deployed systems out there that need to run at realtime on low powered hardware should AND need to be robust against occlusion/motion blur/ect work best with a multicam setup. Theres a reason aria gen2 switched to 4 global shutter cameras. The cuvslam paper argues the same point for general robotics.

Фото профиля Pablo Vela

Pablo Vela10 дней назад

2. IMU support - bascially the same point as above, realtime robotics needs to be robust, and imu + cameras are perfect for each other. They each cover the others weakness

Фото профиля Pablo Vela

Pablo Vela10 дней назад

3. Diverse hardware - I didn't want this to just run on pytorch/nvidia devices. I want this to work on my linux machine, or my mac mini, or my Raspberry Pi! 4. Hardware acceleration - In the same way I want diversity of hardware, I also want to make sure we can take advantage of hardware acceleration. If I have a 5090, I should get the most out of it =] and if I have a rpi5 I should also get the most out of it! It has a GPU on it after all

Фото профиля Pablo Vela

Pablo Vela10 дней назад

5. Opensource - I'm sure theres plenty of great vio/slam pipelines out there that do the above, but none that are open that I know of =/

Фото профиля Pablo Vela

Pablo Vela10 дней назад

So thats exactly what I did, taking inspiration from basalt/cuvslam/glide this repo has imu+multicamera support, works on diverse hardware (tested on my mac mini + 5090 linux machine + rpi5 + rockchip 3588) AND support GPU thanks to the awesome CubeCL library that lets you write kernels on rust and support wgpu/metal/vulkan I also leaned heavily on @kornia_foss and the great work done by the folks there. I used many of the components in kornia-rs and got lots of inspiration from kornia-slam I plan to rip out parts that make sense and contribute them back to kornia. The biggest problem now is that this is VIO, so it has lots of drift. I tried walking for a mile or two and returning to the same spot and the drift is pretty bad. That'll a problem for later me, but shouldn't be too hard to add loop closure. I'm looking at cuSFM + colmap for this.

Фото профиля clankr

clankr10 дней назад

@rerundotio Looks great. Could models like World Labs’ Atlas be used to test SLAM algorithms? What do you think?

Фото профиля Pablo Vela

Pablo Vela10 дней назад

@rerundotio I’m not sure, if I get access I’d be up to try

Фото профиля Mateo de Mayo

Mateo de Mayo10 дней назад

@rerundotio This has many similarities to what we submitted to ICRA two days ago 😅

Фото профиля Pablo Vela

Pablo Vela10 дней назад

@rerundotio Awesome, ya'll do great work! I'd be curious to see how we differ in implementation. I still think theres a bunch that could be done on the GPU, I really only got a good speedup on my 5090 machine.

Фото профиля Mateo de Mayo

Mateo de Mayo10 дней назад

@rerundotio I'll do a post when we release the code :)

Фото профиля Carlos Pinheiro

Carlos Pinheiro9 дней назад

@rerundotio Nice! And btw thanks for the reference, GLidE-SLAM will be presented at iros26.

Фото профиля Damir Wallener

Damir Wallener10 дней назад

@rerundotio Have you ever watched a pigeon walk?

Похожие видео

As part of my exo ego pipeline, I haven't yet found a VIO/Slam solution that hits all of the features I want. Now that agents make it super easy to write low level systems code, I've almost finished building a portable rust/python slam pipeline that works on both a 5090 AND a Rockchip-class computer. To really make this work, I needed a good ground truth, and the wonderful work from Collabora and Mateo de Mayo on the monado slam dataset has been indispensable. I registered it as a Rerun dataset in the OSS catalog to make it very easy to verify that the GT actually looks right and use it as an eval, and it has worked REALLY well. I'll have more on the VIO pipeline here in the coming days

As part of my exo ego pipeline, I haven't yet found a VIO/Slam solution that hits all of the features I want. Now that agents make it super easy to write low level systems code, I've almost finished building a portable rust/python slam pipeline that works on both a 5090 AND a Rockchip-class computer. To really make this work, I needed a good ground truth, and the wonderful work from Collabora and Mateo de Mayo on the monado slam dataset has been indispensable. I registered it as a Rerun dataset in the OSS catalog to make it very easy to verify that the GT actually looks right and use it as an eval, and it has worked REALLY well. I'll have more on the VIO pipeline here in the coming days

Pablo Vela

18,939 просмотров • 13 дней назад

Hey people, so in this video i am explaining the scaling architecture of the two programming languages. Rust vs Python: Scaling Vertical Scaling: Rust uses every CPU core efficiently with no GIL. Python's GIL limits true parallelism—multiprocessing adds overhead. Horizontal Scaling: Rust needs fewer nodes for the same throughput, cutting infrastructure costs. Python scales out easily with mature tooling but requires more instances. Trade-off: Rust optimizes for performance and co zenks. ;)

Hey people, so in this video i am explaining the scaling architecture of the two programming languages. Rust vs Python: Scaling Vertical Scaling: Rust uses every CPU core efficiently with no GIL. Python's GIL limits true parallelism—multiprocessing adds overhead. Horizontal Scaling: Rust needs fewer nodes for the same throughput, cutting infrastructure costs. Python scales out easily with mature tooling but requires more instances. Trade-off: Rust optimizes for performance and co zenks. ;)

Dikshika Sharma

69,623 просмотров • 8 месяцев назад

I've been showing the output of my flight reconstruction tool for a while and I wanted to share something like it which wasn't a mess of hand tuned stuff that required X-Plane. So I created a completely new version in javascript that runs in an HTML page so that anyone can use it. The output is a KML file which is a tour that works in Google Earth pro with all the attitude information included. This is a long way from perfect, I messed around with multiple smoothing algorithms and none of them compares well with the python code, but they're readable and understandable for anyone.

I've been showing the output of my flight reconstruction tool for a while and I wanted to share something like it which wasn't a mess of hand tuned stuff that required X-Plane. So I created a completely new version in javascript that runs in an HTML page so that anyone can use it. The output is a KML file which is a tour that works in Google Earth pro with all the attitude information included. This is a long way from perfect, I messed around with multiple smoothing algorithms and none of them compares well with the python code, but they're readable and understandable for anyone.

Scott Manley

19,808 просмотров • 7 месяцев назад

Here is a cheap solo boat I'm using, I call it the cruiser, sail powered not particularly fast but gets the job done. Anyone got any other simple designs? #Rust

Here is a cheap solo boat I'm using, I call it the cruiser, sail powered not particularly fast but gets the job done. Anyone got any other simple designs? #Rust

Protox

24,316 просмотров • 7 месяцев назад

Well that was an adventure. Toad's fuzzy file search boosts first characters, smallest number of groups, and it highlights the best scoring match. So you can locate a file with the bare minimum of keypresses. It's an expensive algorithm as it scores every possible combination of the query characters. For any given path there could be many indices which match, and each of them is scored separately. Still, it was fast enough to search as you type for my largest repository. Alas, it was slow for very large repos. Like Microsoft's Typescript, which has more than 84,000 files. I figured I'd reach the limit of what I could do with Python. So my options were to use something like ripgrep or re-implement the fuzzy searching in a faster language. I didn't want to add the ripgrep dependency, so I built a Rust version (with the help of AI) which eventually yielded a ~14X speedup. Nice. Better. But it could still become painful to type under certain pathological conditions. An inefficient algorithm can always be slow if you add more items. So I scrapped the Rust solution and built an index in Python—so the fuzzy search could throw away most paths that wouldn't match or have low scores. And its faster than the Rust solution was. The only downside is that there is a little additional work upfront (done in the background). I think this will do for a while. Until I get an issue that it is slow with 10million files. Rust is still an option, as is doing the work in parallel. This will be in the next version of Toad. Here I am searching the Typescript repo... #Python

Well that was an adventure. Toad's fuzzy file search boosts first characters, smallest number of groups, and it highlights the best scoring match. So you can locate a file with the bare minimum of keypresses. It's an expensive algorithm as it scores every possible combination of the query characters. For any given path there could be many indices which match, and each of them is scored separately. Still, it was fast enough to search as you type for my largest repository. Alas, it was slow for very large repos. Like Microsoft's Typescript, which has more than 84,000 files. I figured I'd reach the limit of what I could do with Python. So my options were to use something like ripgrep or re-implement the fuzzy searching in a faster language. I didn't want to add the ripgrep dependency, so I built a Rust version (with the help of AI) which eventually yielded a ~14X speedup. Nice. Better. But it could still become painful to type under certain pathological conditions. An inefficient algorithm can always be slow if you add more items. So I scrapped the Rust solution and built an index in Python—so the fuzzy search could throw away most paths that wouldn't match or have low scores. And its faster than the Rust solution was. The only downside is that there is a little additional work upfront (done in the background). I think this will do for a while. Until I get an issue that it is slow with 10million files. Rust is still an option, as is doing the work in parallel. This will be in the next version of Toad. Here I am searching the Typescript repo... #Python

Will McGugan

12,261 просмотров • 7 месяцев назад

Vibe-coded a proof of concept for a visual scripting solution in Godot. I know it’s rough, but I think there are some interesting ideas here. It supports accelerator keys, so you can just press V → I to declare an integer, then F → R to call the _ready function. The idea is to make it faster to type than normal code while keeping it 1:1 with GDScript.

Vibe-coded a proof of concept for a visual scripting solution in Godot. I know it’s rough, but I think there are some interesting ideas here. It supports accelerator keys, so you can just press V → I to declare an integer, then F → R to call the _ready function. The idea is to make it faster to type than normal code while keeping it 1:1 with GDScript.

FR3NKD Dev

85,853 просмотров • 1 месяц назад

0.32 has shipped, and it's a massive release from Rerun. There's a ton of cool new features, and I wanted to highlight 2 in particular 1. OSS Server streaming from disk 2. Dataset review I walk you through them in the video, so take a look. I'll have a much longer blog post next week about the entire pipeline. With 0.32, much of the foundation is set for a unified data layer for physical data, and I'll be getting into the details of it with all that I've built over the past year. This will cover 1. Raw Data Collection 2. Data Ingestion 3. Catalog Registration 4. Query and Review 5. Post Process 6. Training so lots to share

0.32 has shipped, and it's a massive release from Rerun. There's a ton of cool new features, and I wanted to highlight 2 in particular 1. OSS Server streaming from disk 2. Dataset review I walk you through them in the video, so take a look. I'll have a much longer blog post next week about the entire pipeline. With 0.32, much of the foundation is set for a unified data layer for physical data, and I'll be getting into the details of it with all that I've built over the past year. This will cover 1. Raw Data Collection 2. Data Ingestion 3. Catalog Registration 4. Query and Review 5. Post Process 6. Training so lots to share

Pablo Vela

11,264 просмотров • 4 месяцев назад

I asked Codex 5.4 to reverse engineer a DOS game with no source code. It’s been running for 6 hours, I can’t look away. It unpacked assets, disassembled the EXE, rebuilt the renderer, and built my childhood favorite SkyRoads in Rust! Now think of all the games we can revive.

I asked Codex 5.4 to reverse engineer a DOS game with no source code. It’s been running for 6 hours, I can’t look away. It unpacked assets, disassembled the EXE, rebuilt the renderer, and built my childhood favorite SkyRoads in Rust! Now think of all the games we can revive.

Ammaar Reshi

511,458 просмотров • 6 месяцев назад

Why Rust is overused right now Martin Odersky (Creator of Scala): "Right now, rust is actually overused because a lot of people push Rust for things higher up in the stack Where a garbage collector is fine, but essentially you still want to code without. And that is, for me, a bit an exercise in doing it intellectually just because you can I don't really think there's a big sense in it. If you have the memory for a garbage collector, you should absolutely use one, because it makes a lot of things simpler. So, yes, of course, if given enough brains, I can write code around and I can do that, but why should you? It can be much simpler."

Why Rust is overused right now Martin Odersky (Creator of Scala): "Right now, rust is actually overused because a lot of people push Rust for things higher up in the stack Where a garbage collector is fine, but essentially you still want to code without. And that is, for me, a bit an exercise in doing it intellectually just because you can I don't really think there's a big sense in it. If you have the memory for a garbage collector, you should absolutely use one, because it makes a lot of things simpler. So, yes, of course, if given enough brains, I can write code around and I can do that, but why should you? It can be much simpler."

Ryan Peterman

215,580 просмотров • 27 дней назад

19 days after Figma FIGMA MOTION dropped and I was able to squeeze this out of it. For a software in Beta, not bad.. 🚀🔥💜 I did my best to pay as much attention to the original as I could😄. I've been seeing alot of fancy & cool stuffs done with it, but I actually wanted to see it do the hard stuff that brands could pay $$ for. Thank you all once again for your support, following the process from the day I started. It gets better from here🚀💜. UI design & Motion - Figma Sound & Video edit - CapCut

19 days after Figma FIGMA MOTION dropped and I was able to squeeze this out of it. For a software in Beta, not bad.. 🚀🔥💜 I did my best to pay as much attention to the original as I could😄. I've been seeing alot of fancy & cool stuffs done with it, but I actually wanted to see it do the hard stuff that brands could pay $$ for. Thank you all once again for your support, following the process from the day I started. It gets better from here🚀💜. UI design & Motion - Figma Sound & Video edit - CapCut

Khlaseek designer

16,243 просмотров • 2 месяцев назад

Movie is out! Go watch if you can! Aside from reviewing, giving notes, redlining, and giving as much support to the animators as I much as I can behind the scenes, I got to do some roughs here and there when time allowed it. Here are some of my most faves!

Movie is out! Go watch if you can! Aside from reviewing, giving notes, redlining, and giving as much support to the animators as I much as I can behind the scenes, I got to do some roughs here and there when time allowed it. Here are some of my most faves!

Liem Nguyen (阮清廉)

54,141 просмотров • 2 месяцев назад

"It means everything. I've been here since I can remember. I've probably been here more than I've been at home sometimes, but I wouldn't change it for the world." 💙 From Everton Academy to the @PremierLeague: 'Harrison Armstrong 24/7' is out now. ▶️

"It means everything. I've been here since I can remember. I've probably been here more than I've been at home sometimes, but I wouldn't change it for the world." 💙 From Everton Academy to the @PremierLeague: 'Harrison Armstrong 24/7' is out now. ▶️

Everton

26,467 просмотров • 8 месяцев назад

🐰: Thank you so much for all the support, i think calling it "support" is right! I saw it all, and there was so much of it. I was so amazed watching it too. I really wanted to thank you once again, and to all the armys wishing me a happy birthday, thank you so much.

🐰: Thank you so much for all the support, i think calling it "support" is right! I saw it all, and there was so much of it. I was so amazed watching it too. I really wanted to thank you once again, and to all the armys wishing me a happy birthday, thank you so much.

vjk ☆

18,191 просмотров • 26 дней назад

Take any Python code and build a dynamic GUI on top of it by dragging and dropping components. Python is already the most popular language in the world. This makes it even better! The attached video is only 45 seconds. It will show you how it works. Here is the important part: You can now build an interface for your Python code without learning HTML, JavaScript, or CSS. Go to this link to try this: Remember to work smarter, not harder! Thanks to the team behind Taipy for partnering with me on this post.

Take any Python code and build a dynamic GUI on top of it by dragging and dropping components. Python is already the most popular language in the world. This makes it even better! The attached video is only 45 seconds. It will show you how it works. Here is the important part: You can now build an interface for your Python code without learning HTML, JavaScript, or CSS. Go to this link to try this: Remember to work smarter, not harder! Thanks to the team behind Taipy for partnering with me on this post.

Santiago

155,254 просмотров • 2 лет назад

Josh Hokit reveals he's never heard from Dana White or Hunter Campbell regarding his controversial comments: "They would also message me if they really felt a certain way. At the end of the day, all the other people weren't there when I was down. They weren't there when I was broken. I'm here to make a splash, and I'm here to back up my words as well. If anybody feels some type of way, they know where to find me. That's why I respect them so much... They can always feel some type of way. I respect their opinions, but we do have freedom of speech, so it is what it is. I don't like what other people say, but I've got to deal with it."

Josh Hokit reveals he's never heard from Dana White or Hunter Campbell regarding his controversial comments: "They would also message me if they really felt a certain way. At the end of the day, all the other people weren't there when I was down. They weren't there when I was broken. I'm here to make a splash, and I'm here to back up my words as well. If anybody feels some type of way, they know where to find me. That's why I respect them so much... They can always feel some type of way. I respect their opinions, but we do have freedom of speech, so it is what it is. I don't like what other people say, but I've got to deal with it."

Ariel Helwani

113,604 просмотров • 3 месяцев назад

There's so much focus on "how can AI do my work for me?" I think the more important question is "what work can I now do with AI that I would have never attempted before?" Earlier this year I wrote freestiler, a vector tiling engine for R and Python, with the help of Claude and Codex. I knew what the ideal engine looked like and how it would work at a high level. I didn't know how to put it together, and I don't know Rust, the language I wanted under the hood. Previously I would never have attempted this project as the ROI wasn't there. It would have taken me a year or more to learn the internals of a vector tiling engine and enough Rust to implement one. With Opus-level models, I could take it on. freestiler now powers all my vector tiling pipelines, including the map below rendering 143 million jobs from LODES, and it has 114 GitHub stars. Building this way has required a different set of skills. I don't review the code line by line. I set up adversarial agents to do that and write the test suites. What I review is the architecture, the behavior, and the results. Agent teams surface findings and explain their reasoning; I evaluate and critique. My job isn't to stress over code formatting, but instead to focus on questions like whether the engine is designed right, whether the output is correct, and if the UX makes sense. This means that I haven't "replaced my work." I've taken on entirely new work, with the help of agents, that I would have never done otherwise. It has taken some getting used to shipping code I haven't personally typed. In the old way of working, I built understanding through writing that code. Now I build understanding through managing the project - writing a spec, reviewing structure, evaluating UX. And that's helped me think a whole lot bigger in terms of what I can now do.

There's so much focus on "how can AI do my work for me?" I think the more important question is "what work can I now do with AI that I would have never attempted before?" Earlier this year I wrote freestiler, a vector tiling engine for R and Python, with the help of Claude and Codex. I knew what the ideal engine looked like and how it would work at a high level. I didn't know how to put it together, and I don't know Rust, the language I wanted under the hood. Previously I would never have attempted this project as the ROI wasn't there. It would have taken me a year or more to learn the internals of a vector tiling engine and enough Rust to implement one. With Opus-level models, I could take it on. freestiler now powers all my vector tiling pipelines, including the map below rendering 143 million jobs from LODES, and it has 114 GitHub stars. Building this way has required a different set of skills. I don't review the code line by line. I set up adversarial agents to do that and write the test suites. What I review is the architecture, the behavior, and the results. Agent teams surface findings and explain their reasoning; I evaluate and critique. My job isn't to stress over code formatting, but instead to focus on questions like whether the engine is designed right, whether the output is correct, and if the UX makes sense. This means that I haven't "replaced my work." I've taken on entirely new work, with the help of agents, that I would have never done otherwise. It has taken some getting used to shipping code I haven't personally typed. In the old way of working, I built understanding through writing that code. Now I build understanding through managing the project - writing a spec, reviewing structure, evaluating UX. And that's helped me think a whole lot bigger in terms of what I can now do.

Kyle Walker

13,867 просмотров • 2 месяцев назад

Ye thanks his fans for their years of support ❤️ “You are cordially invited to the greatest show on earth. You could have been anywhere else in the world, but it wouldn't have been as good as here. Y’all now rocking with the greatest that ever did it, that ever lived it. You know it, they know it, THEY KNOW, THEY KNOW, THEY KNOW IT!” “I want to thank y’all for sticking by me all these years. Through the hard times, through the low times. Y’all know what it is. I love you for that. LA, I love you!!!”

Ye thanks his fans for their years of support ❤️ “You are cordially invited to the greatest show on earth. You could have been anywhere else in the world, but it wouldn't have been as good as here. Y’all now rocking with the greatest that ever did it, that ever lived it. You know it, they know it, THEY KNOW, THEY KNOW, THEY KNOW IT!” “I want to thank y’all for sticking by me all these years. Through the hard times, through the low times. Y’all know what it is. I love you for that. LA, I love you!!!”

yzyupdates

128,502 просмотров • 5 месяцев назад