TensorFlow published a blog on the roadmap they expect to take to scale their existing business.
TensorFlow is a free and open-source software library for machine learning and artificial intelligence. The project was authored by Google engineers, and the tech giant remains responsible for its development. The team recently announced future plans in a blog post.
The TensorFlow team says it understands the importance of customer feedback and has committed to listening to users to answer their needs. They also want to follow trends and iterate their existing IPs while scaling up. According to the blog post, TensorFlow bases its success on four pillars: fast and scalable growth, utilization of machine learning, securing deployment and maintaining simplicity.
Their first step towards fast-paced scaling is the introduction of their new API, DTensor. “DTensor unlocks the future of ultra-large model training and deployment and allows you to develop your model as if you were training on a single device, even while using multiple clients”, the team said.
Furthermore, the company is working towards new tools for natural language processing. “We are investing in our ecosystem for applied ML, in particular via the KerasCV and KerasNLP packages which offer modular and composable components for applied CV and NLP use cases, including a large array of state-of-the-art pretrained models.”
In line with TensorFlow’s goal to value and process customer feedback, the team is adding more code examples, guides, and documentation for popular and emerging applied ML use cases. “We aim to increasingly reduce the barrier to entry of ML and turn it into a tool in the hands of every developer.”
When are these innovations expected?
While there is no fixed release date for the changes, the new capabilities can be expected in the second quarter of 2023. Customers may need to wait a bit longer for the production version, which is expected later in the same year. TensorFlow clearly sees what they want to do in the coming years, and the roadmap is well-defined.