Monthly Product Updates March 2024
Hey there! This is Eftal from 1019 with some exciting updates! We launched our closed beta version on February 1st, and it's been a bustling month of additions and improvements. We thrive in the hustle and bustle, and amid all the excitement, we've been building some fantastic new features for 1019!
1- Stripe is Now Our Go-To for All Things Billing!
Goodbye to our old-school billing system, hello Stripe! As of now, all payments and subscription fees are handled through the secure Stripe billing platform.
2- Prioritizing User Experience and Minimizing Complexity
We've been busy bees working on our user interface, and we're thrilled to share some major revamps:
2.1 The Model Execution Page Gets a Makeover
We gave the Model Execution Page a makeover with a new feature we're calling "Execution Modes". Each model has required or paramount inputs which we've streamlined into the 'Basic' mode. For the more adventurous users who love tinkering with specifics, switch over to 'Expert' mode and change any parameter to create the content you desire!
2.2 Speedy Uploads for Speedy Results!
We're talking about a whopping 50% decrease in storage service latency. We took our storage microservice back to the drawing board and supercharged it for speedier results. Got any models with image, voice, or video inputs? On the execution detail page, you can now upload your file and get your results in half the time!
2.3 Filter Models by Collection
As our Model Garden keeps flourishing with the addition of new AI models each day, we realized it can get a bit overwhelming to find a specific model. Enter our brand-new category filtering on the explore page. You can select a collection and instantly get related models from our Garden!
2.4 Welcome Aboard, GitHub Users!
We've opened the doors to GitHub users with our new authentication. And just like that, you can effortlessly sign in with your GitHub account.
2.5 Sleek New Design for the Execution Detail Page
Our revamped Execution Detail Page now lets you historize your inputs and outputs. Have any files to include in your output/inputs? Not a problem - we're storing your execution files for a full 24 hours. So you can revisit your data on the Execution Detail Page whenever you'd like.
3- Bring Your Squad to 1019!
We firmly believe in the power of collaboration. That's why we're thrilled to introduce the Squad plan – an innovative platform for teams and startups to develop, test, and use AI models as a group! Just head over to Settings -> Team page to invite your team members.
4- Faster Inference and Quicker Boot Time - We're in the Fast Lane Now!
We've revved up our inference and training architecture, leading to a significant decrease in inference time. How did we do it? By deploying our stack on AWS and Azure and sharing model volumes. The result? A slick 10% faster inference time is now available on 1019!
5- OpenAI SDK-Compatible API for a Full-Powered LLM
For those teams diving into research and development and needing to fine-tune more parameters from the base models, we've got good news! We've rolled out Nous Hermes Mistral 7B DPO, fully compatible with Open AI API. You can now run Nous Hermes's Mistral model on our Model Garden or via our API
6- Keep an Eye on your Execution Logs!
We've woven log infrastructure into our dashboard. You can check how your execution went and see your prints on the execution detail page.
7- A New Arrival - GPU-Based Cloud Development Environments
Exciting news! Our GPU-based cloud development environment is ready and waiting for you! Select any type of GPU on the CDE page and have your GPU-based Jupyter Notebook up and running in no time!
A blog post explaining the Cloud Development Environment in more detail is coming soon!
Stay tuned for more updates. Exciting times ahead!
