dkobran

15 ms·

91.
by dkobran 8y ago
Also on HN today: “Amazon was just granted a patent for tracking Bitcoin transactions and selling the data to the government” https://www.techspot.com/news/74246-amazon-granted-patent-tr...
92.
How to implement a YOLO (v3) object detector from scratch in PyTorch(blog.paperspace.com)
2 points by dkobran 8y ago|0 comments
93.
by dkobran 8y ago
Just that you’re limited to the number of saved and running experiments (Jobs) on the free plan. For most research/non-production environments, this is sufficient. Hope that helps.
94.
by dkobran 8y ago
There are referral codes floating all over the internet ;)
95.
by dkobran 8y ago
Paperspace Cofounder here. We just launched Gradient, a really simple and affordable DL platform: https://paperspace.com/gradient Feel free to ping me with any questions.
96.
by dkobran 8y ago
I agree, how we are tracked needs to be more transparent/upfront. Clearly these companies will make every possible effort to hide what’s going on behind the scenes so this stuff needs to be regulated. Companies like Google/Facebo
97.
by dkobran 8y ago
Next year the AI will enter the mayoral race itself :)
98.
by dkobran 9y ago
How are you going to release a Broadwell based server in 2018? Sourcing Broadwell processors is already difficult today.
99.
Gradient° – Train deep learning models directly from your command line(medium.com)
3 points by dkobran 9y ago|0 comments
100.
by dkobran 9y ago
NVIDIA makes a chip called the P6000 that offers 24GB memory. It isn't high-throughput (HBM2) memory but regardless, the chip will benchmark relatively close to the Volta V100 and costs about half. Here's a deep learning benchmar
101.
by dkobran 9y ago
HBM2 to be specific: https://en.wikipedia.org/wiki/High_Bandwidth_Memory#HBM_2
102.
China Moves to Let Xi Stay in Power by Abolishing Term Limit(nytimes.com)
84 points by dkobran 9y ago|2 comments
103.
by dkobran 9y ago
It's my understanding that fp16 (available on the previous generation P100) and mixed-precision (major innovation of V100) are different things and the speedup of TensorCores is entirely missing from this benchmark. Unlike the general
104.
by dkobran 9y ago
Just to clarify, is this benchmark leveraging mixed-precision mode on the Volta V100? The major innovation of the Volta generation is mixed-precision which NVIDIA claims is a huge performance increase over the Pascal generation (P100 in th
105.
by dkobran 9y ago
Dan from Paperspace here. Our fast.ai template is Linux based but you can spinup a GPU backed Windows instance on our cloud if you’re interested in a remote option. Our streaming tech is GPU accelerated so it feels snappy.
106.
by dkobran 9y ago
This makes me want to die
107.
Dimension Reduction with Multi-Dimension Scaling (MDS)(blog.paperspace.com)
2 points by dkobran 9y ago|0 comments
108.
Ask HN: It's 2018 – why aren't all text messages delivered over the internet?
4 points by dkobran 9y ago|3 comments
109.
Understanding Dimension Reduction with Principal Component Analysis(blog.paperspace.com)
1 points by dkobran 9y ago|0 comments
110.
by dkobran 9y ago
People are always talking about spot pricing when GPU compute comes up which is valid but not for the ML use-case (which this article is discussing). Spot is 100% incompatible with ML training. You need to be guaranteed that your many hour&
111.
by dkobran 9y ago
Dan here (also Paperspace team). Totally agree that transfer costs are a significant pain point which is why we do not charge for it. We can peer with other providers (eg with AWS we can leverage Direct Connect directly from our datacenter
112.
by dkobran 9y ago
Totally agree though the concept of the uncanny valley is based on computer generated or robotic representations of humans specifically -- not other forms of life or inanimate objects. Meaning, not just realism in general.
113.
by dkobran 9y ago
Thinking that tech money can't destroy a respected hundred year old publication is naive: https://www.theatlantic.com/magazine/archive/2017/09/when-si...
114.
by dkobran 9y ago
While the job doesn't sound like it involves "cool things that matter" and paints a pretty believable picture of working at a large corporation, I do find it interesting that this was published by thr Washington Post. Being o
115.
by dkobran 9y ago
Taking a stand against companies fighting to undermine privacy is a good reason to choose a product like Firefox. I don't think it's as much about being a for-profit [3] as it is about the intentions of the company, its influence,
116.
by dkobran 9y ago
Really interesting glimpse into this world. It's unfortunate that this article is getting negative feedback on HN. I think we all need to be careful not to belittle/condemn issues that don't personally affect us.
117.
Unsupervised Politics: Clustering in machine learning(blog.paperspace.com)
2 points by dkobran 9y ago|0 comments
118.
by dkobran 9y ago
If you want some current generation GPUs ;) check out https://paperspace.com
119.
by dkobran 9y ago
One of the interesting variables in calculating ML training costs is developer time. The cost of a Data Scientist (or similar role) on an hourly basis will far outweigh the most expensive compute resource by several orders of magnitude. Whe
120.
by dkobran 9y ago
The citation is we're building a GPU cloud and have tested almost every GPU in existence :) Just kidding, here are a few examples: http://vfio.blogspot.com.au/2014/08/vfiovga-faq.html https://www.

More ›