kinjal1234/transformer-backend repository overview

⁠Attention-Seeker

Attention-Seeker is a comprehensive Transformer Model trained for Translation. It is a from-scratch implementation of the Transformer architecture as proposed in "Attention Is All You Need."

Unlike standard implementations that utilize high-level abstractions, this project is a ground-up reconstruction of the core architecture. This includes the manual implementation of Multi-Head Attention, Positional Encodings, and specialized masking logic.

View on GitHub⁠

View model weights on HuggingFace⁠