there is a download link in the fine-tune page

We're trying to streamline the process and explain it. Running the infra for this is hard. We are running on expensive A100 machines which are not easily available and not part of Google Colab infra. There are lots of hyper-parameters one could toy with, and we're trying to hide those and give optimal results automatically. Obviously this could all just be the start of it, and we might continue to develop on different directions where we see it's most fun :)

This seems possibly almost perfect for me as an API but I can't use it for my service without knowing what the pricing will be for generation after the model is trained.

Like I can't launch with the assumption that will stay free forever unless you really pro.ises that, and can't know if it is feasible to use this instead of setting up everything myself on Google Cloud without knowing that pricing.

$3 sounds pretty good for the training though. How long does it take? Also so you have an option for using a different number of instances to train faster (for more money)?

Hey @ilaksh Let's talk over email. We have a lot of different options and details re the API that we're still considering and trying to improve with our partners

We're relying on Google's Dreambooth paper which differently from Text-Inversion, it does full model fine-tuning. This should help achieve higher fidelity results if you're trying to replicate the distinct features of your subject

Indeed we're working on that and explaining the technology behind is one of the things we're working on. Currently we're training a Stable Diffusion image-generation model, to the custom images that the user provide so that he can generate images of his specific subject consistently. This relies on a paper by Google called Dreambooth. So thank you mainly to Nathaniel Ruiz

Hi all, great to see someone posted our service :) We started working on this 3 weeks ago and are having real fun with the response. People are creating amazing stuff, and doing great experiments.

Why is this needed if you implement SCA 3D secure?

Thanks for your comment/question. 3D secure is helpful and would be redundant. We found that 3D secure support is a bit nebulous across platforms and also dependent upon the implementation. (Javascript vs API vs your ecommerce platform.)

UpUp also lets ecommerce teams verify customers without requiring their engineering teams to immediately implement 3D secure.

We do expect we'll end up using 3D secure as part of our flow in the future.

It's not, and SCA 3D Secure is a better solution.

Especially when the card providers move the risk, and therefore part of the cost of the transaction onto the bank. It's up to the bank to make 3D Secure, secure by authenticating the user in it's own way.

Yep. I'm looking forward to finish our implementation using WebDNN. It's super fast

Indeed I had a problem converting the latest model to KerasJS

Here are my thoughts

1) Free p2p money transfers / gateway to bitcoin or other crypto-currency so that it's more widely adopted

2) Better open bank accounts - allowing open transparent accounting for organizations and companies

3) Solve democracy - better analytical tools for mass discussion, arguments and decision making which will encourage use of facts and science, and discourage politics

4) Human-Machine interfaces - memory augmentation

5) Solve the common cold and influenza

6) Robotics - better batteries, finer motors and sensors - possibly through the usage of biological systems

7) Public access to satellites - realtime security monitoring, crops analytics and forecasting

8) Solve weather or create private air-conditioned jackets ;)

For 3) Solve democracy, a friend showed me pol.is (neat demo of it: https://pol.is/demo/2demo).

Anyway, it's description is..."pol.is brings AI & machine learning to participatory democracy. Scale up outreach in online consultation & get powerful insights that can shape and legitimize policy." It would be amazing if U.S. politics could be grounded in legitimate understanding of each other.

The problem is not making the tools, it's getting everybody to learn and use them.

I'm wondering if some kind of mind-map/graph mashed-up with a wikipedia model and "likes" on each node, can help drive decisions or is it just too naive/simplistic and low adoption

It starts with the issue that most people don't especially like to use computers and will avoid them if they can.

7) there are several companies in this space, https://www.planet.com/ being one, and they're acquiring another one, Terra Bella (formerly Skybox), the deal is on the front page right now: https://news.ycombinator.com/item?id=13573523). There might be more that I don't remember.

Hasn't Johnny Mnemonic taught you anything???

BoardingBot https://boardingbot.boorgle.com Is a tool I made few weeks ago over 4 days. I started it as a Facebook chat bot which sends TestFlight invitations and then realized it's also worth turning it into a more complete tool which creates a website and invitation request form which automatically sends TestFlight invitations. It was featured in ProductHunt and I was generally surprised with the positive feedback. I'm wondering if it's worth investing more time into it...