model_265165561_coca_tiny.py

Model Overview

A tiny-scale implementation of the coca architecture, built for generation tasks.

Architecture

• Architecture: coca
• Scale: tiny
• Attention: standard
• Fusion strategy: bilinear
• Task head: generation
• Activation: gelu tanh
• Normalization: rmsnorm
• Initialization: orthogonal

Training

• Optimizer: sgd
• LR scheduler: exponential

Files

• model_265165561_coca_tiny.py — main artifact of this repository

License

See the license field above.