CSV loader tutorial

This tutorial mirrors the "data first" workflow people expect from PyTorch:

1. Load a dataset from disk (CSV).
2. Turn it into a fixed-size batched dataset.
3. Train through the public Trainer API.

Generate a small deterministic regression dataset with python3 NN/Examples/Data/generate_small_data.py:

• NN/Examples/Data/small_regression.csv with rows x1,x2,y (25 samples).

Build:

• lake build NN.Examples.Data.Loaders.Csv

The tutorial code is compiled with the rest of TorchLean and is directly runnable as lake exe torchlean data_csv. It keeps data loading and training together so the parsed CSV shape can be checked before the first optimizer step.

Optional flags (tutorial-specific):

• --data-dir PATH (default: NN/Examples/Data)
• --csv PATH (override the CSV file)
• --seed S (controls shuffling and model initialization)
• --batch N
• --steps N

Public API used here:

• Data.fromCsv
• Trainer.new
• Trainer.RunConfig
• Trainer.TrainOptions
• trainer.train

def NN.Examples.Data.Loaders.Csv.exeName : String

Command name used in diagnostics and by the top-level example runner.

def NN.Examples.Data.Loaders.Csv.missingCsvHint : String

Printed when the CSV is absent, so the reader knows how to produce it rather than just seeing a file-not-found error.

def NN.Examples.Data.Loaders.Csv.inputWidth : ℕ

Two input features, matching the generated regression CSV.

def NN.Examples.Data.Loaders.Csv.hiddenWidth : ℕ

Hidden width; small enough that the printed parameter tensors fit on a screen.

def NN.Examples.Data.Loaders.Csv.outputWidth : ℕ

One regression target.

def NN.Examples.Data.Loaders.Csv.model {batchSize : ℕ} : TorchLean.nn.Builder (TorchLean.nn.Sequential [batchSize, inputWidth] [batchSize, outputWidth])

A small 2-layer batched MLP 2 -> 8 -> 1.

def NN.Examples.Data.Loaders.Csv.usage : String

Command-line help for the CSV loader tutorial.

def NN.Examples.Data.Loaders.Csv.main (args : List String) : IO Unit

Entry point: read the CSV, build the loader, then train the 2 -> 8 -> 1 MLP on it.