Running serverless Runpod in a production-level Gen Art service Hey all -- my startup, Distillery, runs 100% on Runpod serverless, using network storage and A6000s. We run ComfyUI on the backend with a custom connector we created to do it, and we open sourced both it and the Runpod worker codebase. If you'd like to know more about how to deploy it, fire me a DM and I'll be glad to talk about it. I had some tough times with the lack of servers, but this issue has reduced considerably in the past weeks, as promised by the Runpod guys. We launched on ProductHunt today, and thankfully people seem to be interested in it; we're currently #6 in the product launches today with about 200 upvotes. We are now approaching 2k members in our server, and this very busy day has been running very smoothly so far using our Runpod backend. Thanks Runpod guys for the ongoing support.

Up and running in 60 seconds Our team took a little time to test out RunPod to see how it stacks up against AWS, Azure, Google Cloud, and other cloud infrastructure providers we've used in the past. What can I say? The prices are an offer I can't refuse! We were able to spin up our first RunPod instances within about a minute using their custom docker deployment interface. Definitely recommend this service to anyone that needs GPUs in a hurry.

Told me no refund Told me no refund. (Teutontech) even though their documentation is not good for one and things kept crashing. And when you upload anything you cannot run it from the terminal. I SAW YOUR REPLY AND IT MADE ME EVEN MORE MAD. DO NOT WORRY I'll CONTINUE TO TELL PEOPLE TO STAY AWAY!

Very positive experience Very positive experience, the service and the price are unmatched. RunPod being very reactive and involved in the ML and AI Art communities makes them a great choice for people who want to tinker with machine learning without breaking the bank. The very helpful and friendly support from the team is also a big plus when you're trying to get things set up.

Please don't buy. Won't recommend using it. I used for training 7 models. For all of them, the cells (cpu process) took longer than other service like google colab. For example, loading a pretrained model from hf. It took around 20 sec for me to load and finish running the cell in colab. The runpod paused. Used an L40. It it pauses for around 5 min. Happens almost all the time. Had to delete the pod and recreate another pod. Which again adds unnecessary cost. Overall, they are theif.