An API for hosted deep learning models
blog.algorithmia.com
An API for hosted deep learning models
1–10 of 35 posts
Re: An API for hosted deep learning models
#2How does this compare in price to AWS GPU instances?
Re: An API for hosted deep learning models
#3> "Using GPUs inside of containers is a challenge. There are driver issues, system dependencies, and configuration challenges. It’s a new space that’s not well-explored, yet. There’s not a lot of people out there trying to run multiple GPU jobs inside a Docker container.”
Er, Nvidia itself has an official Docker application which allows containers to interface with the host GPU, optimized for the deep learning use case: https://github.com/NVIDIA/nvidia-docker
Training models is one thing that can commoditized, like with this API, but building models and selecting features without breaking the rules of statistics is another story and is the true bottleneck for deep learning. That can't be automated as easily.
Re: An API for hosted deep learning models
#4The history of machine learning startups is littered with companies that thought a hosted web service was a good idea. The problem with this model is that big data, by definition, is costly to move. So if a managed service is not generating and storing the data you need to process with machine learning or deep learning (as you might conceivably with AWS), then you probably don't want to move your data to those algorithms or models. All you'll get are small-data users. The models and algos need to go to the data. That's the most efficient approach, and it means you have to go on prem... Fwiw, that's what we're trying to do with Skymind and Deeplearning4j.
Re: An API for hosted deep learning models
#5> "Using GPUs inside of containers is a challenge. There are driver issues, system dependencies, and configuration challenges. It’s a new space that’s not well-explored, yet. There’s not a lot of people out there trying to run multiple GPU jobs inside a Docker container.” Er, Nvidia itself has an official Docker application which allows containers to interface with the host GPU, optimized for the deep learning use ca…
Algorithmia founder here. nvidia-docker is helpful but does not address all the issues with running GPU computing inside of docker. There are driver issues on the host OS, and the real challenge is running multiple GPU jobs inside of separate docker containers and sharing the GPU.
I agree that building models is still definitely a big challenge, but the tooling and knowledge is getting better every day. Either way, our goal with Algorithmia is to create a channel for people to make their models available, and create incentive for people to put in the effort to train really solid, useful models.
Re: An API for hosted deep learning models
#6> "Using GPUs inside of containers is a challenge. There are driver issues, system dependencies, and configuration challenges. It’s a new space that’s not well-explored, yet. There’s not a lot of people out there trying to run multiple GPU jobs inside a Docker container.” Er, Nvidia itself has an official Docker application which allows containers to interface with the host GPU, optimized for the deep learning use ca…
What's the first step towards easily automating it? What would the long term roadmap look like?
Re: An API for hosted deep learning models
#7The history of machine learning startups is littered with companies that thought a hosted web service was a good idea. The problem with this model is that big data, by definition, is costly to move. So if a managed service is not generating and storing the data you need to process with machine learning or deep learning (as you might conceivably with AWS), then you probably don't want to move your data to those algori…
My understanding of the post is it would host learned models. You'd train them wherever, but host the learned model in algorithmia, which exposes it through an api, making it easy for others to use your model.
Re: An API for hosted deep learning models
#8How does this compare in price to AWS GPU instances?
The service operates at a higher level than EC2, and pricing is calculated on a per-second of compute basis. Comparing prices is going to depend a lot on the specifics of your workload and your affinity for managing infrastructure.
Re: An API for hosted deep learning models
#9The history of machine learning startups is littered with companies that thought a hosted web service was a good idea. The problem with this model is that big data, by definition, is costly to move. So if a managed service is not generating and storing the data you need to process with machine learning or deep learning (as you might conceivably with AWS), then you probably don't want to move your data to those algori…
There's a clear trend in the industry to increasingly rely on cloud services, so it seems reasonable that machine learning would follow the same trend. As long as the compute is in the same data center, data transfer is rarely the bottleneck for these kinds of deep learning algorithms, which is why we designed algorithmia to be able to operate anywhere -- on all the major cloud providers, as well as on premise.
Re: An API for hosted deep learning models
#10The history of machine learning startups is littered with companies that thought a hosted web service was a good idea. The problem with this model is that big data, by definition, is costly to move. So if a managed service is not generating and storing the data you need to process with machine learning or deep learning (as you might conceivably with AWS), then you probably don't want to move your data to those algori…
My understanding of the post is it would host learned models. You'd train them wherever, but host the learned model in algorithmia, which exposes it through an api, making it easy for others to use your model.
It takes big data to train but that trained model can work on "small data" or big data. One-off uses for apps really do lend themselves well to a hosted solution like this IMHO. If you need to classify lots of data – then you are probably at a point to either train your own model or buy it from the developer via this site, I'd think.