model_700769099_dino_huge.py

Model Overview

A huge-scale implementation of the dino architecture, built for generation tasks.

Architecture

• Architecture: dino
• Scale: huge
• Attention: grouped query
• Fusion strategy: low rank
• Task head: generation
• Activation: approx gelu
• Normalization: rmsnorm
• Initialization: xavier uniform

Training

• Optimizer: rmsprop
• LR scheduler: step

Files

• model_700769099_dino_huge.py — main artifact of this repository

License

See the license field above.