Implement Full Transformer Model in Rust

[JakubSchwenkbeck]

Implement Full Transformer Model in Rust

Resolves #4

Description

This pull request introduces a complete implementation of a Transformer model based on the "Attention is All You Need" paper. The model includes:

1. Separate Encoder and Decoder Inputs:

   • Handles two distinct input sequences, one for the encoder and one for the decoder.

2. Embedding Layer:

   • Converts tokenized input sequences into dense vector embeddings.

3. Transformer Encoder:

   • Processes the encoder input to produce encoded representations.

4. Transformer Decoder:

   • Utilizes the encoded representation and decoder input to generate context-aware outputs.

5. Output Projection and Softmax:

   • Projects the decoder output into vocabulary space and applies softmax to generate token probabilities.

[JakubSchwenkbeck]

MERGE ON WORKING TESTS