Sequence Model Examples

Runnable sequence-model examples, organized by what each file is meant to teach:

The “main” entrypoints most people should look at first:

• CharGpt (torchlean chargpt): Karpathy-style char-level GPT on a single text file (Tiny Shakespeare). This is the simplest end-to-end path: read text, tokenize by characters, train, sample.
• Gpt2 (torchlean gpt2): byte-level GPT-2-style causal Transformer with a small, local-friendly config. Use this when you want to see masked self-attention + LayerNorm + FFN wiring, and a save/reload path via Gpt2Saved.
• TextGpt2 (torchlean text_gpt2): CUDA-only corpus trainer (byte-level by default, optional GPT-2 BPE). This is the “serious” trainer interface for bigger text runs.
• Mamba (torchlean mamba): compact text walkthrough for the Mamba-style model.

Other sequence examples:

• Rnn and Lstm: compact real-text recurrent smoke tests over the shared SimpleText runner.
• Transformer: one-block encoder example for attention/norm/FFN wiring.
• GptAdder: synthetic algorithmic curriculum (addition), runnable as torchlean gpt_adder.

For supervised time-series forecasting with an LSTM, see NN.Examples.Models.Supervised.LstmRegression (torchlean lstm_regression).