Researchers love PyTorch and TensorFlow

In a recent survey—AI Adoption in the Enterprise, which drew more than 1,300 respondents—we found significant usage of several machine learning (ML) libraries and frameworks. About half indicated they used TensorFlow or scikit-learn, and a third reported they were using PyTorch or Keras.

I recently attended an interesting RISELab presentation delivered by Caroline Lemieux describing recent work on AutoPandas and automation tools that rely on program synthesis. In the course of her presentation, Lemieux reviewed usage statistics they had gathered on different deep learning frameworks and data science libraries. She kindly shared some of that data with me, which I used to draw this chart:

The numbers are based on simple full-text searches of papers posted on the popular e-print service arXiv.org. Specifically, they reflect the number of papers which mention (in a full-text search) each of the frameworks. Using this metric, the two most popular deep learning frameworks among researchers are TensorFlow and PyTorch. From January to the end of June 2019, about 1,800 papers mentioned TensorFlow and a comparable number mentioned PyTorch. Most notably, interest in PyTorch among researchers is growing rapidly: it grew 194% year-over-year (Jan-Jun 2018 compared to Jan-Jun 2019).