We explore the geometric properties of features learned by image models. Through the intrinsic dimensionality of feature spaces, we seek to understand the structure of feature manifolds and their interrelationships. This investigation aims to illuminate how neural networks represent and transform visual information across their layers.
sun-wendy/intrinsic-dim
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|
Repository files navigation
Releases
No releases published
Generated from eliahuhorwitz/Academic-project-page-template