Subscribe
Sign in
DataTorch — Build and share datasets to experiment with ML
DataTorch is a platform for building datasets and experimenting with machine learning algorithms. Organize and annotate data across multiple cloud storages, automate tasks with agents and pipelines, and share your models and data with the wider community!
11
Replies
Hi, I'm Michael Nguyen, co-founder of DataTorch. One of the biggest problems with trying to apply machine learning to real-world problems is the lack of training datasets.
Like many others, I taught myself ML through online resources and communities, but when it came time to actually use my knowledge in a research project, I quickly realized that building the algorithm wasn't the hard part; it was gathering the data. My PI (a doctor, not a computer scientist) was appalled to learn that the very classification work that we had set out to automate actually had to be done anyways, if not many times over, to be fed into the model to get the best results possible! If we had DataTorch back then, things would have been so much easier.
We built DataTorch to be a community and a developer tool put together. A "Github for datasets", if you will. A place where people can collaborate on gathering and annotating data, experiment with machine learning, automate tasks and deployments, and share or discuss their models. We are used by companies, universities, and hobbyists around the world to help make their ML ideas come to fruition.
It's easy to join us and get started. Go to:
https://datatorch.io/
and create an account! And please don't hesitate to leave a comment or ask questions here on ProductHunt. We are happy to let you know more.
In November I stumbled upon DataTorch when I needed to annotate images for a computer vision model for the final project in my data science bootcamp. I was looking for a program that was quick to get started, relatively intuitive to use, and easy to share the work with my project partners. DataTorch checked all those boxes.
Since finishing the bootcamp I continue to use DataTorch and have been impressed with the ever expanding suite of tools available. I still have a lot to learn to put all the tools to their best use but it makes working in a new career less intimidating when I have a great place to start.
@lvparkinson Thanks so much Lindsey! We really enjoy working with you as well.
As a software eng and founder without any ML exp, I never leveraged any ML in my products. When I found DataTorch, I was able to start leveraging ML in my apps to exponentially grow. DataTorch has been extremely helpful, especially with growth and retention experiments. The pricing is great too. I upgraded to Pro and it's incredible what I get for only $10/mo.
Geneva