A developer wants to write their own neural network code, train it, and then run the same code on a laptop or on TPU pods. Which Google-originated tool is designed for this?
Show answer and explanation
Answer: C
TensorFlow is an open-source machine learning framework originally developed at Google. Developers use it to define neural networks, train them and export them for deployment, and it runs on CPUs, GPUs and TPUs. Its portable design is why the same code can move between a laptop and TPU hardware.
Why the other options are wrong
- A. A pre-built vision endpoint offers fixed models accessed through an API and does not let a developer author and train their own network code.
- B. A managed Kubernetes control plane orchestrates containers. It does not provide APIs for defining and training neural networks.
- D. A serverless ETL service moves and transforms data between systems. It is unrelated to building and training neural network models.