Implementing higher parallelism using the WebGL backend

WebGL is a backend implementation that uses the WebGL API. WebGL is a standard API that's used for graphical processing on the web. Most web browser vendors implement this API in their web browsers. The WebGL backend implements kernel operations by using shader programs so that we can make use of the capabilities of higher parallelism. The WebGL backend stores tensors as textures that can be uploaded to GPU memory. Each element in the tensor is treated as a point of the texture. Shader programs execute these operations by the point in the texture in parallel. In general, we can get up to 100 x faster performance than a CPU backend.

The kernel operation in the WebGL backend is written ...

Get Hands-On Machine Learning with TensorFlow.js now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.