track:nlp-llm

NLP / LLM

A focused route through sequence models, transformers, adaptation, evaluation, and language-model serving.

10

Lessons

Formal explanations, examples, simulations, and checkpoints.

Sequence ML: RNNs, GRUs & LSTMs

Sequences are not IID

Use sequences are not iid to move the sequence models production brief toward a defensible release.

Sequence ML: RNNs, GRUs & LSTMs

Recurrent state and BPTT

Use recurrent state and bptt to move the sequence models production brief toward a defensible release.

Sequence ML: RNNs, GRUs & LSTMs

Gates that remember: LSTM and GRU

Use gates that remember: lstm and gru to move the sequence models production brief toward a defensible release.

Sequence ML: RNNs, GRUs & LSTMs

Many-to-one and many-to-many training

Use many-to-one and many-to-many training to move the sequence models production brief toward a defensible release.

Sequence ML: RNNs, GRUs & LSTMs

Stateful streaming inference without future leakage

Use stateful streaming inference without future leakage to move the sequence models production brief toward a defensible release.

Build a transformer from scratch

Text to tensors: tokenization, embeddings, position

Use text to tensors: tokenization, embeddings, position to move the transformer from scratch production brief toward a defensible release.

Build a transformer from scratch

Scaled dot-product self-attention by hand

Use scaled dot-product self-attention by hand to move the transformer from scratch production brief toward a defensible release.

Build a transformer from scratch

Heads, norms, residuals, and MLPs

Use heads, norms, residuals, and mlps to move the transformer from scratch production brief toward a defensible release.

Build a transformer from scratch

Train a causal language model

Use train a causal language model to move the transformer from scratch production brief toward a defensible release.

Build a transformer from scratch

Sampling, KV cache, and efficient generation

Use sampling, kv cache, and efficient generation to move the transformer from scratch production brief toward a defensible release.