Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Leaf: Machine learning framework in Rust

github.com

Post reply on HN
Print view View original

31–40 of 53 posts

Re: Leaf: Machine learning framework in Rust

#31

Post by kvb »

YeGoblynQueenne wrote: post #16

> super-human image recognition That's a bold claim. As far as I know there was one paper that reported a model beating human scores in a specific test (imagenet, I believe). Whether that translates to "superhuman" results in general is followed by a very big question mark. In general I really struggle to see how any algorithm that learns from examples, especially one that minimises a measure of error against further…

Ensembles of humans can outperform the average human, and in the same way an algorithm trained on data labeled by an ensemble of humans can outperform the average human.

Re: Leaf: Machine learning framework in Rust

#32

Post by deepnet »

tomp wrote: post #23
benbou09 wrote: post #19

Earlier quoted context omitted.

It is much faster than humans

Computers have been faster than humans for the last 40 years. That doesn't make them more intelligent.

Then by implication this task does not require intelligence ;)

Computers are faster serial processors but brains do more in parallel.

Parallel pipelines only really hit Neural Nets with GPU's and the Imagenet convnet solvers like Alexnet were among the 1st parallel implementations - this gave 30 - 300 speedup but still relatively tiny compared with squishy wetware.

Re: Leaf: Machine learning framework in Rust

#33

Post by somerandomness »

kingnothing wrote: post #29

I'm completely new to ML and what real world applications it's suitable for. Are we at the point yet where you can train a computer to look at arbitrary images and count the number of people in it? What if it was the largely on the same background and only the number of people were changing -- for example, a camera shooting a queue of people to determine queue depth at a bus station.

I actually think this is quite do-able and has been for a while. Although deep learning has revolutionized object recognition, face detection has been working reasonably well for some time, e.g. your cell phone camera or Google street view face blurring.

Re: Leaf: Machine learning framework in Rust

#34

Post by hobofan »

eggy wrote: post #26

I will take a look at it, but are the benchmarks comparable, since to quote the site, "For now we can use C Rust wrappers for performant libraries."? Torch is LuaJit over C, and Tensorflow has Python and C++. Is Rust making it fast, or the interface code to the C libraries?

The interface code to the C libraries (which is written in Rust). We are however optimistic that there will be Rust libraries popping up in the future that outperform the current C implementations. (Optimistic as a Rust user, not as developer as Leaf)

Re: Leaf: Machine learning framework in Rust

#36

Post by hobofan »

rck wrote: post #28

The benchmarks would be a lot more useful if the context around them were more obvious. In particular, it would be nice to know if the benchmarks are for a single input, or for a batch of inputs. If for a batch, then the batch size is important too. Maybe this stuff is somewhere on their site, but it shouldn't require digging. Without this information it's hard to make a useful comparison at all.

You are right, batchsize is important and we should make that more clear.

The numbers in the benchmark are taken from our deep-learning-benchmarks[1] which we are still in the process of building up. It might actually make sense to test the same model with different batch sizes. The current benchmarks are based on the convnet-benchmarks[2] where the Alexnet model has a batch size of 128. (Alexnet was chosen because out of the benchmarks that's the one I am most familiar with, since it small enough that I can work with it on my Laptop)

In some informal tests Leaf was generally faster than other frameworks in smaller batch sizes, but no benchmarks that we could publish with confidence yet.

[1]: https://github.com/autumnai/deep-learning-benchmarks [2]: https://github.com/soumith/convnet-benchmarks

Re: Leaf: Machine learning framework in Rust

#38

Post by dang »

mastax wrote: post #35

I'm glad that rust has crossed the point where posts to HN that would be "_ in Rust" are now just "_". I hope this means that Rust is starting to be used for its own merits rather than just novelty.

We changed the title to say "in Rust" because someone else complained about "for Hackers". I suppose we could take both of them out, but the project highlights its Rustiness so this seems more representative.

Re: Leaf: Machine learning framework in Rust

#39

Post by emcq »

wall_words wrote: post #20

The performance graph is deceptive for two reasons: (1) Leaf with CuDNN v3 is a little slower than Torch with CuDNN v3, yet the bar for leaf is positioned to the left of the one for Torch, and (2) there's a bar for Leaf with CuDNN v4, but not for Torch. It's good to see alternatives to Torch, Theano, and TensorFlow, but it's important to be honest with the benchmarks so that people can make informed decisions about w…

Yes, on their site they post Torch CuDNN v4 as faster than Leaf [0]. Seems exciting for an early release.

Can it get much faster than something like Torch? I would think if CuDNN is doing most of the computation time it would be hard to see big improvements. Perhaps go the route of Neon and tune your GPGPU code like crazy [1, 2], or MXNet and think about distributed computing performance [3].

[0] http://autumnai.com/deep-learning-benchmarks

[1] https://github.com/soumith/convnet-benchmarks

[2] https://github.com/NervanaSystems/neon

[3] http://alex.smola.org/talks/NIPS15.pdf

Re: Leaf: Machine learning framework in Rust

#40

Post by danielvf »

kingnothing wrote: post #29

I'm completely new to ML and what real world applications it's suitable for. Are we at the point yet where you can train a computer to look at arbitrary images and count the number of people in it? What if it was the largely on the same background and only the number of people were changing -- for example, a camera shooting a queue of people to determine queue depth at a bus station.

In the scale of computer vision problems, the stationary camera case is relatively easy. It's not too hard to isolate moving objects from a background, it's not too hard to decide if an object is a person or not, and it's not too hard to keep track of an object once you've identified it. You would still have to handle overlapping people, scene illumination changes, etc, but these can be solved and have been done before.

If you would like to play with some of this stuff, take a look at OpenCV. http://opencv.org

Post reply on HN