VisionMambaModelSize

Enums · AiDotNet.Enums

Defines the backbone size variants for Vision Mamba (Vim) segmentation.

How It Works

Reference: Zhu et al., “Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Model”, ICML 2024.

Fields

Field Summary
VisionMambaModelSize Small Vim-S: width 384, 24 layers, with a 2D rotary position embedding in the official segmentation configuration.
VisionMambaModelSize Tiny Vim-Ti: width 192, 24 layers.
int value__