model_308967884_tiny_transformer_small.py

Model Overview

A small-scale implementation of the tiny transformer architecture, built for classification tasks.

Architecture

• Architecture: tiny transformer
• Scale: small
• Attention: sliding window
• Fusion strategy: co attention
• Task head: classification
• Activation: approx gelu
• Normalization: rmsnorm
• Initialization: kaiming normal

Training

• Optimizer: lamb
• LR scheduler: linear warmup

Files

• model_308967884_tiny_transformer_small.py — main artifact of this repository

License

See the license field above.