4 ms·
Show HN: Panini AI – A platform to serve ML/DL models at low latency
- nl 8y agoThis claim (3x faster than TF serving) and the metrics on the site (~500 predictions per second vs ~200 for TF serving) seem more a function of scaling than any technology. Given that you can horizontally scale model prediction infinitely the only sensible way to compare is to include price. I agree that this looks compelling while it is free! But will it be price competitive later? And if price competitiveness is claimed, then how is it possible? Yes, you can do the whole spot instance thing, but that is difficult to make reliable enough at scale.
- avin_regmi 8y agoHey, both prediction for TF serving and panini serving was done in a single thread in the same specification machine. We used a simple model for image classification of CIFAR dataset. Roughly, 500 predictions were made for panini and 200 predictions for TF serving. You can always download the entire panini in your own private server and not pay anything. Ie. used Helm to install in your own kubernetes or DockerHub. For now, We're making it free for models under 2GB. Our main goal is to make it usable and we don't want cost to be a factor.
- nl 8y agoSo you claim that TF Serving (written in C++ I believe) has over double the overhead compared to Panini? This seems surprising. What makes it so much faster? Edit: Unless of course you are hitting the cache for a lot of the predictions?
- avin_regmi 8y agoOptimized TF serving would perform similarly to Panini however, it's really hard to find good documentation on optimizing TF serving compilation parameters. Panini automatically finds the right batch size to maximize the throughput and it adaptively changes. We also have a technique to reduce bound tail latency. I would love it for you to try it and provide me some feedback. Thanks
- tedivm 8y agoIs it actually 3x faster, or does it just scale more? In other words, if I had a model that previously took three seconds to get a response from would this platform respond in one second?
- dang 8y agoWe changed the title from "Show HN: 3x Faster Than Tensorflow Serving" to what the page says, which is less baity. https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html
- tedivm 8y agoThat's definitely an improvement, but I'm hoping someone from the Panini team will step in and clarify regardless.
- avin_regmi 8y agoSorry, I should've been more clear. Both predictions for TF serving and panini serving was done in a single thread in the same specification machine. We used a simple model for image classification of CIFAR dataset. Roughly, 500 predictions were made for panini and 200 predictions for TF serving. The graph on the website is for throughput. I'm planning to write a medium post soon regarding the benchmark test.
- etaioinshrdlu 8y agoThis is so fishy looking on many levels! It is very hard to believe that deploying your models on GKE is going to be cost saving for anyone involved.
- avin_regmi 8y agoHey, you don't have to deploy in GKE and it's not GKE that makes it faster. We also give you option to deploy in your own private Kubernetes via Helm or private server via DockerHub. GKE may not be the right option for you depending on your application. Your feedback would be very valuable to us. Please tell me why you think its fishy? We're always tryiing to make it better.
- tedivm 8y agoIt sounds super fishy, especially since you won't answer the question above confirming whether you're talking about latency or throughput. Google has also dumped a lot into tensorflow serve, so if you are outperforming it by that much it would be great to know how.
- avin_regmi 8y agoSorry, I should've been more clear. Both predictions for TF serving and panini serving was done in a single thread in the same specification machine. We used a simple model for image classification of CIFAR dataset. Roughly, 500 predictions were made for panini and 200 predictions for TF serving. The graph on the website is for throughput. I'm planning to write a medium post soon regarding the benchmark test. There are many other projects getting higher throughput compare to TF serving. I've heard TF Serving could be optimized to make it more efficient but making it more optimized is not documented properly. We're planning to make it open source if there is enough interest from the community!
- tedivm 8y agoI'd definitely be willing to try this if it was open source.