Frequently Asked Questions (FAQs)
Vilvik is a cloud optimization platform. It runs your problem through an optimization algorithm and returns the best solution it can find, along with the data you need to understand how it got there. Today the supported algorithm is the genetic algorithm, built on an extended version of the open-source PyGAD library, with more algorithms and problem types on the roadmap. Vilvik handles single- and multi-objective problems and works with popular libraries like Keras, PyTorch, and scikit-learn, so you can train machine-learning models and tune hyperparameters with minimal setup in a secure environment. It also ships with ready-to-use applications that need no coding, from hyperparameter optimization to the travelling salesman problem and 2D clustering, and it supports custom code for problems that are not covered out of the box.
Yes. Vilvik offers a trial that includes some limits compared to the full version, so you can explore its features. If your trial has ended and you need more time to evaluate Vilvik, reach out to us and we will be happy to discuss extending it.
Yes. Vilvik offers a REST API so you can create and manage submissions, fetch results, and automate runs from your own code, plus a Python SDK (the vilvik package) that wraps the API with a friendlier interface. You can try endpoints in the API playground and find full reference in the API and SDK sections of the documentation.
To continue an already-run submission for more generations, the 'Save Code' option must have been enabled in the app settings before you ran the original submission. When it was, an option to continue the submission is available.
Yes. The 'Save Code' option in the app settings controls whether your code is saved. It is enabled by default, but when disabled any code you provide is used only to run the submission and is then discarded.
You can cancel your subscription at any time from the Billing page by clicking 'Cancel Subscription' to schedule the cancellation.
Once you cancel, the subscription is scheduled for cancellation but stays active until the end of your current billing period. After that period ends the subscription is canceled and no further payments are taken. During this grace period you can reactivate at any time before the period ends, which resumes normal payment processing.
No, we do not store your card or bank information. All payments are processed securely through the Stripe platform.
We aim to include a wide range of Python libraries so you can import them directly into your code. Some libraries might not be pre-installed in our environment. If you need one that is not available, contact us and we will look into adding it.
Vilvik has two billing models, and a single account can use both at once. The website runs on a monthly subscription: you pick a tier and get a set number of submissions per day, and you can cancel at any time. The API runs on credits, which is pay as you go: you top up a balance once and each run deducts credits, with no monthly fee. Your subscription tier covers what you run on the website, and your credit balance covers what you run through the API. New accounts also start with a free trial and a free monthly credit allowance, so you can begin without paying anything.
API runs are charged when the run finishes, and the cost reflects the compute the run actually used, the processing time and the memory, not the parameters you submitted. Reading a result is free, and a run that is rejected before it starts costs nothing. A run that fails partway through, for example by reaching a memory or time limit, still costs for the time it ran. You can see every charge and your remaining balance on the billing tab of your profile.
It depends on your problem. The main factors are the number of generations, the population size, and how long your fitness function takes to evaluate a single solution, since that function runs many times. Small problems can finish in seconds, while larger searches take longer. Each tier also caps the maximum time a single run may use, and a run that reaches that cap stops automatically.
Yes. Your submissions run in a secure, isolated environment, and your results stay private to your account unless you choose to share them. Your custom code is only stored when the 'Save Code' option is enabled, which is what lets you continue a run later; with it disabled, the code is used for the run and then discarded. As a precaution, do not put passwords or other secrets in the code you submit.
Continuing resumes a finished run and searches for more generations, building on the solutions it already found, which requires the 'Save Code' option to have been enabled on the original run. Re-executing runs the same submission again from the start. Editing and re-running lets you change the parameters or code first and then run it. Each one is a fresh run and is billed like any other run.
We use cookies. Essential cookies keep you signed in. With your consent, optional cookies help us improve the site. Privacy policy.
