Atlas record: https://www.aitimeline.dev/en/events/evt-0230/
Availability · November 9, 2015
TensorFlow
Google open-sourced its own system for training models: computation is described as a graph and can run on a processor, a graphics accelerator or a cluster.
Why it matters
An industrial-grade tool became free, and the barrier to deep learning fell to knowing Python.
Before this every laboratory had its own half-written tooling and reproducing someone else's result was hard. TensorFlow gave a shared foundation and with it a shared way of describing models. A year later PyTorch appeared with a different execution philosophy, and the rivalry between the two shaped the field's tooling for a decade.
Event record
- Event date
- November 9, 2015
- Timeline date
- Event date
- Verification
- Sources gathered automatically · September 17, 2026
- Lines
- ID
- evt-0230
The system was open-sourced on 9 November 2015; the white paper appeared in March 2016.
Related events
Earlier
- Related Keras 2015links to this record
It later became the official high-level interface to TensorFlow.
- Related Theano 2012links to this record
The direct forerunner of TensorFlow in design: a symbolic graph with automatic differentiation.
- Related Deep networks on graphics processors 2009
A tool that makes distributed training on accelerators available outside large companies.
Later