burn
antfarm | burn | |
---|---|---|
3 | 9 | |
3 | 7,223 | |
- | 7.0% | |
2.2 | 9.8 | |
almost 1 year ago | 5 days ago | |
TypeScript | Rust | |
- | Apache License 2.0 |
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.
antfarm
-
3 years of fulltime Rust game development, and why we're leaving Rust behind
I've experienced a lot of these concerns while building https://github.com/MeoMix/symbiants
I have a simple question that maybe someone smarter than me can answer confidently:
If I want to build something akin to Dwarf Fortress (in terms of simulation complexity) as a browser-first experience - what stack should I choose?
Originally, I prototyped something out using React, PixiJS, and ReactPixi (https://github.com/MeoMix/antfarm). The two main issues I ran into were the performance of React's reconciler processing tens of thousands of entities when most weren't changing (despite heavy memoization) and GC lurching due to excess object allocations. My takeaway was that if I wanted to continue writing in JS/TS that I would need to write non-idiomatic code to avoid excess allocations and abandon React. This approach would result in me effectively creating my own engine to manage state.
I decided to not go that direction. I chose Rust because no GC is a language feature (especially good since GCs in WASM are heavy) and I chose Bevy because it seemed like a fairly structured way to mutate a large amount of code.
Progress has been slow for a lot of the reasons listed in this article. I've written a lot of this off to WASM being a new frontier for game dev and rationalized it by noting there's not a lot of complex simulation games running in browser (that I'm aware of). It's not clear to me if that's actually true, though.
-
Ask HN: Good resources for architecting simulation games?
I recently made an attempt at building a simulation game. I found it challenging even though I'm a seasoned software developer. I was finding that my experience in building out web apps, dashboards, and crud stacks didn't immediately lend itself to architecting a game properly. It feels like everything is business logic and it has been tough to see ways to tease out and modularize concepts. I don't know if I am making good trade-offs w.r.t performance and immutability. I find myself copying an entire "world" structure with every loop in an attempt to represent the world immutably, but I also seem to need to process changes serially as entities interact with one another inside the world. It doesn't matter that the world is immutable, I can't easily parallelize the simulation. Perhaps immutability is only making things harder, then?
I'm not too concerned about graphics or 3D math. I'm working in a 2D space with the most amateur of graphics. I'm interested in having some looming trade-offs pointed out to me so that I can make decisions with my architecture to best receive those trade-offs. I'm interested in design patterns that allow me to modularly enrich the complexity of my simulation.
If you're curious, https://github.com/MeoMix/antfarm/blob/main/src/util.ts Here is a file that became quite the dumping ground. I am not proud of this code. I had intended to tease out some utility methods which take a world and return a cloned and updated world. It's easy enough to separate out view concepts, and it's easy enough to separate out constructors, but all of the "interesting" simulation logic found its way to this dumping ground.
-
Hi! I'm building a virtual ant farm. This version is so pre-alpha it shouldn't even be live, but, in the interest of shipping first and asking questions later... it is! So, check it out.
open source! https://github.com/MeoMix/antfarm
burn
-
3 years of fulltime Rust game development, and why we're leaving Rust behind
You can use libtorch directly via `tch-rs`, and at present I'm porting over to Burn (see https://burn.dev) which appears incredibly promising. My impression is it's in a good place, if of course not close to the ecosystem of Python/C++. At very least I've gotten my nn models training and running without too much difficulty. (I'm moving to Burn for the thread safety - their `Tensor` impl is `Sync` - libtorch doesn't have such a guarantee.)
Burn has Candle as one of its backends, which I understand is also quite popular.
- Burn: Deep Learning Framework built using Rust
-
Transitioning From PyTorch to Burn
[package] name = "resnet_burn" version = "0.1.0" edition = "2021" [dependencies] burn = { git = "https://github.com/tracel-ai/burn.git", rev = "75cb5b6d5633c1c6092cf5046419da75e7f74b11", features = ["ndarray"] } burn-import = { git = "https://github.com/tracel-ai/burn.git", rev = "75cb5b6d5633c1c6092cf5046419da75e7f74b11" } image = { version = "0.24.7", features = ["png", "jpeg"] }
- Burn Deep Learning Framework Release 0.12.0 Improved API and PyTorch Integration
-
Supercharge Web AI Model Testing: WebGPU, WebGL, and Headless Chrome
Great!
For Burn project, we have WebGPU example and I was looking into how we could add automated tests in the browser. Now it seems possible.
Here is the image classification example if you'd like to check out:
https://github.com/tracel-ai/burn/tree/main/examples/image-c...
-
Burn Deep Learning Framework 0.11.0 Released: Just-in-Time Automatic Kernel Fusion & Founding Announcement
Full Release Note: https://github.com/tracel-ai/burn/releases/tag/v0.11.0
- Burn Deep Learning Framework v0.11.0 Released: Just-in-Time Kernel Fusion
- Burn – comprehensive dynamic Deep Learning Framework built using Rust
- Burn: Deep Learning Framework in Rust
What are some alternatives?
dfdx - Deep learning in Rust, with shape checked tensors and neural networks
candle - Minimalist ML framework for Rust
wonnx - A WebGPU-accelerated ONNX inference run-time written 100% in Rust, ready for native and the web
tch-rs - Rust bindings for the C++ api of PyTorch.
rust-mlops-template - A work in progress to build out solutions in Rust for MLOPs
llama2.rs - A fast llama2 decoder in pure Rust.
corgi - A neural network, and tensor dynamic automatic differentiation implementation for Rust.
albumin - Simple Hierarchical Album Generator
Enzyme - High-performance automatic differentiation of LLVM and MLIR.
rust-ndarray - ndarray: an N-dimensional array with array views, multidimensional slicing, and efficient operations
are-we-learning-yet - How ready is Rust for Machine Learning?
cob-webber - Shelob of Cirith Ungol