Lead author of DSSTNE here... 1. DSSTNE was designed two years ago specifically for product recommendations from Amazon's catalog. At that time, there was no TensorFlow, only Theano and Torch. DSSTNE differentiated from these two frameworks by optimizing for sparse data and multi-GPU spanning neural networks. What it's not currently is another framework for running AlexNet/VGG/GoogleNet etc, but about 500 lines of co…
Where do you wok now? It's interesting to hear what offer you couldn't refuse after being in so many places
Amazon DSSTNE: Deep Scalable Sparse Tensor Network Engine
51–57 of 57 posts
Re: Amazon DSSTNE: Deep Scalable Sparse Tensor Network Engine
#52First TensorFlow and now this. Tensor is quickly becoming a mathematical-term-that-sounds-familiar-to-developers-but-most-don't-know-what-it-is-actually. Another example is topology =)
But can we say we know what vectors are though? As far as I know tensors are derived from vectors and I would imagine programmers don't know what vectors are in a mathematical sense.
Tensors can work on other things apart from vector spaces (e.g. modules), but programmers don't know those either.
Re: Amazon DSSTNE: Deep Scalable Sparse Tensor Network Engine
#53Soo... what is the application for this (other than buzzwords)
Re: Amazon DSSTNE: Deep Scalable Sparse Tensor Network Engine
#54Earlier quoted context omitted.
It's more than that, and it's in use in production at Amazon. 8 TitanX GPUs can contain networks with up to 6 billion weights. As Geoffrey Hinton once said: "My belief is that we’re not going to get human-level abilities until we have systems that have the same number of parameters in them as the brain." And you're right that it's a specialized framework/engine. But IMO making it more general purpose is a matter of c…
> "My belief is that we’re not going to get human-level abilities until we have systems that have the same number of parameters in them as the brain." An interesting quote. Replicating functioning of the brain, or some major subsystem of it, is no doubt going to require far more than just billions of parameters. The cortex contains >15 billion neurons, but there are also the neurons contained in all the other brain s…
> "My belief is that we’re not going to get human-level abilities until we have systems that have the same number of parameters in them as the brain."
While that may be true, I find this compelling:
"The fundamental unit of biological information processing is the molecule, rather than any higher level structure like a neuron or a synapse; molecular level information processing evolved very early in the history of life."
http://www.softmachines.org/wordpress/?p=1558#more-1558
Edit: formatting
Re: Amazon DSSTNE: Deep Scalable Sparse Tensor Network Engine
#55Deep Learning systems are becoming C++11's halo projects. Here are some deep learning libraries from the Internet Big 4.
Amazon DSSTNE - https://github.com/amznlabs/amazon-dsstne
Google TensorFlow - https://github.com/tensorflow/tensorflow/
Microsoft CNTK - https://github.com/Microsoft/CNTK/
Facebook fbcunn - https://github.com/facebook/fbcunn/
They all utilize C++11 or later. Just as Hadoop pushed Java in the big data, map-reduce realm, I think these libraries will push C++11 in the Deep Learning realm.
Re: Amazon DSSTNE: Deep Scalable Sparse Tensor Network Engine
#56Soo... what is the application for this (other than buzzwords)
srsly? all the discussion above 10hrs+ before your comment and that's your question?
RTFM: https://github.com/amznlabs/amazon-dsstne/blob/master/FAQ.md