spaCy/website/docs/usage/101/_serialization.md

30 lines
1.6 KiB
Markdown
Raw Normal View History

If you've been modifying the pipeline, vocabulary, vectors and entities, or made
updates to the model, you'll eventually want to **save your progress** for
example, everything that's in your `nlp` object. This means you'll have to
translate its contents and structure into a format that can be saved, like a
file or a byte string. This process is called serialization. spaCy comes with
**built-in serialization methods** and supports the
[Pickle protocol](https://www.diveinto.org/python3/serializing.html#dump).
> #### What's pickle?
>
> Pickle is Python's built-in object persistence system. It lets you transfer
> arbitrary Python objects between processes. This is usually used to load an
> object to and from disk, but it's also used for distributed computing, e.g.
> with
> [PySpark](https://spark.apache.org/docs/0.9.0/python-programming-guide.html)
> or [Dask](http://dask.pydata.org/en/latest/). When you unpickle an object,
> you're agreeing to execute whatever code it contains. It's like calling
> `eval()` on a string so don't unpickle objects from untrusted sources.
All container classes, i.e. [`Language`](/api/language) (`nlp`),
[`Doc`](/api/doc), [`Vocab`](/api/vocab) and [`StringStore`](/api/stringstore)
have the following methods available:
| Method | Returns | Example |
| ------------ | ------- | ------------------------ |
| `to_bytes` | bytes | `data = nlp.to_bytes()` |
| `from_bytes` | object | `nlp.from_bytes(data)` |
| `to_disk` | - | `nlp.to_disk("/path")` |
| `from_disk` | object | `nlp.from_disk("/path")` |