Integrating Hasty in our workflow


For about a week I have been using your labeling service. I must say I am amazed by the results. I wanted to train my own model, however yours seems to do really well. This made me think, once the result is perfected can I use your model instead of training my own? Being able to run locally would be great, however uploading batches and getting result back would be great as well. Is this already an option or do you have plans to offer this in the future?

I am training a Deeplab v3+ but it’s not easy on limited hardware. I am able to train but not yet able to get the loss down. Do you have recommendations on what to use?

Hi Maarten!

Thanks for reaching out; I’m thrilled to hear that Hasty is working out so great for you. Yes, it is possible to use our models. We also did that for several edge use-cases in the past already, so if you have hardware limitations, it still should work.

In the near future, this functionality will be part of our app. But for now, we still have to do some custom work. So, to help you here, we’d need to know some more details. I suggest having a call with me and our Co-Founder Tristan. I’ve sent you a Calendly link as a private message.

Talk to you soon,

Hi Tobias, thanks! I got the PM and will schedule the call.

Great! Looking forward to talking to you.