Based on our record, Neuro should be more popular than Scale Nucleus. It has been mentiond 4 times since March 2021. We are tracking product recommendations and mentions on various public social media platforms and blogs. They can help you identify which product is more popular and what people think of it.
Projects are definitely the best way to learn models. Build things for fun that do things in topics/fields that you care about or think is cool. a few years ago when I was getting into ML stuff I build fantasy football things that weren't even useful but provided an actual use case. Then I did more complicated stuff with photography and lighting because I did real estate photography. As far as ML libraries go,... Source: almost 3 years ago
So far I’ve seen AWS Sagemaker kind of allows for a situation like this, but would rather not deal with all that config. Algorithmia and Nuclio are too enterprise focused. Neuro is new and looks great, but from my understanding I would still need to create a lambda instance myself that then calls neuro’s servers - too indirect. Is there a total solution out there for this? Source: almost 3 years ago
A couple of weeks ago I put out a post on DeepSpeech running on the serverless setup at Neuro (https://getneuro.ai), and I've now got Silero running there as well. I've found this model is a lot faster than DS and way more accurate. Seeing around 300ms per request at the moment, hopefully will be closer to 100ms soon but this is a pretty decent speed in this application already. Source: about 3 years ago
I just made a streaming script connecting Deepspeech to serverless GPUs at Neuro (https://getneuro.ai). Was a fun piece of work, and cool to play around with. You can find the source here: https://github.com/neuro-ai-dev/npu_examples/tree/main/deepspeech. Source: about 3 years ago
At Scale we built a tool for model debugging in computer vision called Nucleus (scale.com/nucleus) designed exactly for this, which is free try out if you're curious to see where your model predictions are most at odds with your ground truth. Source: over 2 years ago
To address your point about gathering edge cases, which can also be defined as cases of low model fidelity for our use cases, there is active learning and tools such as Aquarium Learning and Scale Nucleus which make it easy to implement into workflows. Source: almost 3 years ago
Lobe - Visual tool for building custom deep learning models
Aquarium - Improve ML models by improving datasets they’re trained on
Opta - Opta is a new kind of Infrastructure-As-Code framework designed for fast moving startups.
PerceptiLabs - A tool to build your machine learning model at warp speed.
mlblocks - A no-code Machine Learning solution. Made by teenagers.
ML Image Classifier - Quickly train custom machine learning models in your browser