eh train¶
Train a new eigenspace model from a corpus directory.
Usage¶
Options¶
| Option | Default | Description |
|---|---|---|
--corpus PATH |
required | Path to synced corpus directory |
--language LANG |
required | Target language |
--output PATH |
required | Output .npz model path |
--corpus-class {A,B,C} |
A |
Corpus quality class |
--min-files N |
10 | Minimum files required to train |
Example¶
# Sync a corpus from manifest
eh corpus sync corpora/my-corpus.toml
# Train
eh train --corpus corpus/my-corpus --language python --output models/my-model.npz
# Verify
eh inspect models/my-model.npz
Corpus classes¶
| Class | Description |
|---|---|
| A | Elite — curated, reviewed, exemplary code |
| B | Good — production code from quality projects |
| C | Mixed — broad collection, mixed quality |
Higher-class corpora produce more discriminating models.