Update v2-2 docs

website/docs/usage/v2-2.md

@@ -98,9 +98,10 @@ on disk**.
 
 > #### Example
 >
-> ```python
-> scorer = nlp.evaluate(dev_data)
-> print(scorer.textcat_scores, scorer.textcats_per_cat)
+> ```bash
+> spacy train en /path/to/output /path/to/train /path/to/dev \
+>   --pipeline textcat \
+>   --textcat-arch simple_cnn --textcat-multilabel
 > ```
 
 When training your models using the `spacy train` command, you can now also
@@ -117,6 +118,34 @@ classification.
 
 </Infobox>
 
+### New DocPallet class to efficiently Doc collections
+
+> #### Example
+> 
+> ```python
+> from spacy.tokens import DocPallet
+> pallet = DocPallet(attrs=["LEMMA", "ENT_IOB", "ENT_TYPE"], store_user_data=False)
+> for doc in nlp.pipe(texts):
+>     pallet.add(doc)
+> byte_data = pallet.to_bytes()
+> # Deserialize later, e.g. in a new process
+> nlp = spacy.blank("en")
+> pallet = DocPallet()
+> docs = list(pallet.get_docs(nlp.vocab))
+> ```
+
+If you're working with lots of data, you'll probably need to pass analyses
+between machines, either to use something like Dask or Spark, or even just to
+save out work to disk. Often it's sufficient to use the doc.to_array()
+functionality for this, and just serialize the numpy arrays --- but other times
+you want a more general way to save and restore `Doc` objects.
+
+The new `DocPallet` class makes it easy to serialize and deserialize
+a collection of `Doc` objects together, and is much more efficient than
+calling `doc.to_bytes()` on each individual `Doc` object. You can also control
+what data gets saved, and you can merge pallets together for easy
+map/reduce-style processing.
+
 ### CLI command to debug and validate training data {#debug-data}
 
 > #### Example