LayoutMap
LayoutMap is a Keras API class that maps model-variable paths to sharding or replication layouts across a device mesh.
Deep Learning with Python, Third Edition
A Keras API for assigning variable sharding or replication layouts across a device mesh.
- How it is used
- Configures variable sharding and replication for Keras model parallelism across a named device mesh.
- How deep it goes
- Explains per-dimension layouts and shows assignments, while stopping short of broader compiler implementation details.
- Examples
- Uses a 32×64 variable, dense-layer kernels and biases, variable paths, and a 2×4 data/model device mesh.
- What you can do after
- Operate/configure Keras model parallelism by mapping variable paths to per-dimension sharding layouts.
Read how Deep Learning with Python, Third Edition explains LayoutMap →
Deep Learning with R, Third Edition
LayoutMap is a Keras API class that maps model-variable paths to sharding or replication specifications over a device mesh.
- How it is used
- Configures how model variables are sharded or replicated across a device mesh for distributed training.
- How deep it goes
- Defines path-to-tuple mappings and demonstrates assignments, without detailing compiler internals.
- Examples
- Uses Dense-layer kernel and bias paths with tuple(NULL, "model") and tuple("model") layouts.
- What you can do after
- Operate Keras distribution configuration by assigning sharding layouts to model-variable paths.
Read how Deep Learning with R, Third Edition explains LayoutMap →
Related concepts
- dataset splitconfigures · A dataset split is the division of a dataset into distinct subsets assigned different purposes, such as training, validation, or testing.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.