Very cool, well done! I love the training workflow. Is it possible to download trained models to use them offline (e.g. in Enterprise package maybe)?
Launch HN: Nyckel (YC W22) – Train and deploy ML classifiers in minutes
41–50 of 50 posts
Re: Launch HN: Nyckel (YC W22) – Train and deploy ML classifiers in minutes
#42This looks cool, I signed up. I may have to employ this in my algotrading pipeline. I've collected historical data from 100s of different indicators and I'm currently using a largely heuristic approach with some light ML on subsets, but so far I haven't had the ML breakthrough I've been looking for, https://grizzlybulls.com/models/vix-ta-macro-advanced
I think you're going to have a hard time using historical data to deploy a real time algo trader, unless your strategy involves only trading on a macro scale or against some time horizon. One hour and 30 minutes is pretty common as you typically see a flurry of volume at these intervals. 30 minutes also has some legacy pit trader strategies that I'm sure automated systems use. Arbitrary time slices of price or volume data doesn't really dictate how the market works in real time however. Hmmm, now that I think about it this might be viable, but execution is an entirely different beast in production depending on which instrument you're trading. Have to consider current volatility, slippage not to mention all of the crazy stuff that happens at the liquidity level... And even then things get spoofed all the time. Then you start dealing with the HFT stuff.
Re: Launch HN: Nyckel (YC W22) – Train and deploy ML classifiers in minutes
#43Very cool, well done! I love the training workflow. Is it possible to download trained models to use them offline (e.g. in Enterprise package maybe)?
Thank you! We do think that "model export" is important, but we're still working out how to do it in the most seamless and non-ML-expert friendly way. Do you have a use-case and target hardware in mind?
Re: Launch HN: Nyckel (YC W22) – Train and deploy ML classifiers in minutes
#44In what ways is this different from some of the cloud based AutoML tools such as Google Cloud Vision AutoML? Or tools like Roboflow?
There are a continuum of offerings in this space. Some have lots of custom control of the training pipeline and deployment, and on the other side, things like RoboFlow that try to make it easy / hide the complexity. We consider ourselves even further to the “hide complexity” side, since we try several deep networks automatically vs. making you choose, re-train automatically, abstract-away non-essential ML jargon, etc. In addition, we don't limit ourselves to only vision - we'd like to be the one stop shop for ML as a service. We also have developer-friendly pricing with quick and easy signup.
We benchmarked ourselves against Google AutoML and HuggingFace, looking at both user experience and model performance, and wrote it up in a blog post that may interest you: https://www.nyckel.com/blog/automl-benchmark-nyckel-google-h...
Re: Launch HN: Nyckel (YC W22) – Train and deploy ML classifiers in minutes
#45Looks great, good luck! Bonus points for XKCD. I'm a big fan of technologies that bring new capabilities to everyone, not just specialists of that field. Minor clarification, I assume those 'includes X invocations' are per month, not a one time credit?
Thank you! Everything just clicked when we saw that XKCD strip.
Yes, you are right - 'includes X invocations' are per month.
Re: Launch HN: Nyckel (YC W22) – Train and deploy ML classifiers in minutes
#46Interesting product strategy, but I can't help laugh at some of the product examples like "barcodeless scanner" or "quality inspection" (I can only comment on imaging as that's my background). First of all the idea of replacing a perfectly functional barcode scanner with an ML model is not a great sell (having how much I pay at the register dependent on the lighting at the grocery store is not going to be a fun time)…
Barcodeless scanner is a great example, this is something that major supermarkets are already rolling out as a complement to their self-serve checkouts.
Re: Launch HN: Nyckel (YC W22) – Train and deploy ML classifiers in minutes
#47This is very cool. A few quick questions: 1.) Would it be possible to buy the model and integrate/host it on my own machines? 2.) Would you consider making solutions for embedded ML in the future?
We do think "model export" is important, but we're still getting our heads around how to do it in the most non-ML-expert friendly way. We don't think the persona we're building for wants a weights file dropped in their lap. What output / format would be ideal from your perspective?
As a SWE, my preferred export format would be a folder with a Dockerfile. I'm not sure how GPUs work with docker, but if I could just run the container on a machine with a gpu, or deploy it into a k8s cluster with an affinity for a node with a gpu, that would be my ideal. I'm not sure how I'd want model drift to be addressed, so being able to pull down and deploy new versions would be something to consider as well.
Re: Launch HN: Nyckel (YC W22) – Train and deploy ML classifiers in minutes
#48Earlier quoted context omitted.
We do think "model export" is important, but we're still getting our heads around how to do it in the most non-ML-expert friendly way. We don't think the persona we're building for wants a weights file dropped in their lap. What output / format would be ideal from your perspective?
As a SWE, my preferred export format would be a folder with a Dockerfile. I'm not sure how GPUs work with docker, but if I could just run the container on a machine with a gpu, or deploy it into a k8s cluster with an affinity for a node with a gpu, that would be my ideal. I'm not sure how I'd want model drift to be addressed, so being able to pull down and deploy new versions would be something to consider as well.
Thanks for the input - that is useful to know.
Re: Launch HN: Nyckel (YC W22) – Train and deploy ML classifiers in minutes
#49Hi, exactly what I am looking for. Do you handle multiple languages for text categorization (other than English)?
Re: Launch HN: Nyckel (YC W22) – Train and deploy ML classifiers in minutes
#50Hi, exactly what I am looking for. Do you handle multiple languages for text categorization (other than English)?
Yes! We have several language-specific models we train against to see which performs the best for your data. Which language (or languages) do you need? (feel free to email me at dan at nyckel dot com)