"Value Error: bytes object is too large" when using to_disk on large model.

Hi,

I'm attempting to initialize the [gpt2-xl huggingface model](https://huggingface.co/gpt2-xl) in SpaCy using the following code provided in [examples/init_model.py](https://github.com/explosion/spacy-transformers/blob/master/examples/init_model.py):

```
#!/usr/bin/env python
import plac
from wasabi import Printer
from spacy_transformers import TransformersLanguage, TransformersWordPiecer
from spacy_transformers import TransformersTok2Vec

@plac.annotations(
path=("Output path", "positional", None, str),
name=("Name of pre-trained model", "option", "n", str),
lang=("Language code to use", "option", "l", str),
)
def main(path, name="gpt2-xl", lang="en"):
msg = Printer()
msg.info(f"Creating model for '{name}' ({lang})")
with msg.loading(f"Setting up the pipeline..."):
nlp = TransformersLanguage(trf_name=name, meta={"lang": lang})
nlp.add_pipe(nlp.create_pipe("sentencizer"))
nlp.add_pipe(TransformersWordPiecer.from_pretrained(nlp.vocab, name))
nlp.add_pipe(TransformersTok2Vec.from_pretrained(nlp.vocab, name))
msg.good("Initialized the model pipeline")
nlp.to_disk(path)
msg.good(f"Saved '{name}' ({lang})")
msg.text(f"Pipeline: {nlp.pipe_names}")
msg.text(f"Location: {path}")
with msg.loading("Verifying model loads..."):
nlp.from_disk(path)
msg.good("Model loads!")

if __name__ == "__main__":
plac.call(main)
```

After downloading the model, the function nlp.to_disk(path) raises an exception:

```
$ python3 test.py models/gpt2-xl/
ℹ Creating model for 'gpt2-xl' (en)
✔ Initialized the model pipeline
Traceback (most recent call last):
File "test.py", line 49, in
plac.call(main)
File "/home/jaronmm2/.local/lib/python3.6/site-packages/plac_core.py", line 367, in call
cmd, result = parser.consume(arglist)
File "/home/jaronmm2/.local/lib/python3.6/site-packages/plac_core.py", line 232, in consume
return cmd, self.func(*(args + varargs + extraopts), **kwargs)
File "test.py", line 39, in main
init_gpt2(path, name, lang)
File "test.py", line 23, in init_gpt2
nlp.to_disk(path)
File "/home/jaronmm2/.local/lib/python3.6/site-packages/spacy/language.py", line 924, in to_disk
util.to_disk(path, serializers, exclude)
File "/home/jaronmm2/.local/lib/python3.6/site-packages/spacy/util.py", line 677, in to_disk
writer(path / key)
File "/home/jaronmm2/.local/lib/python3.6/site-packages/spacy/language.py", line 922, in
serializers[name] = lambda p, proc=proc: proc.to_disk(p, exclude=["vocab"])
File "pipes.pyx", line 208, in spacy.pipeline.pipes.Pipe.to_disk
File "/home/jaronmm2/.local/lib/python3.6/site-packages/spacy/util.py", line 677, in to_disk
writer(path / key)
File "pipes.pyx", line 206, in spacy.pipeline.pipes.Pipe.to_disk.lambda6
File "/home/jaronmm2/.local/lib/python3.6/site-packages/thinc/neural/_classes/model.py", line 405, in to_disk
file_.write(self.to_bytes())
File "/home/jaronmm2/.local/lib/python3.6/site-packages/thinc/neural/_classes/model.py", line 372, in to_bytes
return srsly.msgpack_dumps({b"weights": weights})
File "/home/jaronmm2/.local/lib/python3.6/site-packages/srsly/_msgpack_api.py", line 16, in msgpack_dumps
return msgpack.dumps(data, use_bin_type=True)
File "/home/jaronmm2/.local/lib/python3.6/site-packages/srsly/msgpack/__init__.py", line 40, in packb
return Packer(**kwargs).pack(o)
File "_packer.pyx", line 285, in srsly.msgpack._packer.Packer.pack
File "_packer.pyx", line 291, in srsly.msgpack._packer.Packer.pack
File "_packer.pyx", line 288, in srsly.msgpack._packer.Packer.pack
File "_packer.pyx", line 235, in srsly.msgpack._packer.Packer._pack
File "_packer.pyx", line 264, in srsly.msgpack._packer.Packer._pack
File "_packer.pyx", line 206, in srsly.msgpack._packer.Packer._pack
ValueError: bytes object is too large
```