Incompatibility of spacy models as ray reference

## How to reproduce the behaviour

When running spacy with ray, for inference usecase, an error `ValueError: buffer source array is read-only` is encountered. Full stack is shown below:

```
File "~/.venv/lib/python3.8/site-packages/ray/_private/worker.py", line 2380, in get
raise value.as_instanceof_cause()
ray.exceptions.RayTaskError(ValueError): ray::generate() (pid=596931, ip=172.16.147.156)
File "test.py", line 9, in generate
doc = list(nlp.pipe([corpus]))[0]
File "~/.venv/lib/python3.8/site-packages/spacy/language.py", line 1618, in pipe
for doc in docs:
File "~/.venv/lib/python3.8/site-packages/spacy/util.py", line 1703, in _pipe
yield from proc.pipe(docs, **kwargs)
File "spacy/pipeline/transition_parser.pyx", line 245, in pipe
File "~/.venv/lib/python3.8/site-packages/spacy/util.py", line 1650, in minibatch
batch = list(itertools.islice(items, int(batch_size)))
File "~/.venv/lib/python3.8/site-packages/spacy/util.py", line 1703, in _pipe
yield from proc.pipe(docs, **kwargs)
File "spacy/pipeline/transition_parser.pyx", line 245, in pipe
File "~/.venv/lib/python3.8/site-packages/spacy/util.py", line 1650, in minibatch
batch = list(itertools.islice(items, int(batch_size)))
File "~/.venv/lib/python3.8/site-packages/spacy/util.py", line 1703, in _pipe
yield from proc.pipe(docs, **kwargs)
File "spacy/pipeline/pipe.pyx", line 55, in pipe
File "~/.venv/lib/python3.8/site-packages/spacy/util.py", line 1703, in _pipe
yield from proc.pipe(docs, **kwargs)
File "spacy/pipeline/pipe.pyx", line 55, in pipe
File "~/.venv/lib/python3.8/site-packages/spacy/util.py", line 1703, in _pipe
yield from proc.pipe(docs, **kwargs)
File "spacy/pipeline/trainable_pipe.pyx", line 73, in pipe
File "~/.venv/lib/python3.8/site-packages/spacy/util.py", line 1650, in minibatch
batch = list(itertools.islice(items, int(batch_size)))
File "~/.venv/lib/python3.8/site-packages/spacy/util.py", line 1703, in _pipe
yield from proc.pipe(docs, **kwargs)
File "spacy/pipeline/trainable_pipe.pyx", line 79, in pipe
File "~/.venv/lib/python3.8/site-packages/spacy/util.py", line 1722, in raise_error
raise e
File "spacy/pipeline/trainable_pipe.pyx", line 75, in spacy.pipeline.trainable_pipe.TrainablePipe.pipe
File "~/.venv/lib/python3.8/site-packages/spacy/pipeline/tok2vec.py", line 126, in predict
tokvecs = self.model.predict(docs)
File "~/.venv/lib/python3.8/site-packages/thinc/model.py", line 334, in predict
return self._func(self, X, is_train=False)[0]
File "~/.venv/lib/python3.8/site-packages/thinc/layers/chain.py", line 54, in forward
Y, inc_layer_grad = layer(X, is_train=is_train)
File "~/.venv/lib/python3.8/site-packages/thinc/model.py", line 310, in __call__
return self._func(self, X, is_train=is_train)
File "~/.venv/lib/python3.8/site-packages/thinc/layers/chain.py", line 54, in forward
Y, inc_layer_grad = layer(X, is_train=is_train)
File "~/.venv/lib/python3.8/site-packages/thinc/model.py", line 310, in __call__
return self._func(self, X, is_train=is_train)
File "~/.venv/lib/python3.8/site-packages/thinc/layers/with_array.py", line 36, in forward
return cast(Tuple[SeqT, Callable], _ragged_forward(model, Xseq, is_train))
File "~/.venv/lib/python3.8/site-packages/thinc/layers/with_array.py", line 91, in _ragged_forward
Y, get_dX = layer(Xr.dataXd, is_train)
File "~/.venv/lib/python3.8/site-packages/thinc/model.py", line 310, in __call__
return self._func(self, X, is_train=is_train)
File "~/.venv/lib/python3.8/site-packages/thinc/layers/concatenate.py", line 57, in forward
Ys, callbacks = zip(*[layer(X, is_train=is_train) for layer in model.layers])
File "~/.venv/lib/python3.8/site-packages/thinc/layers/concatenate.py", line 57, in
Ys, callbacks = zip(*[layer(X, is_train=is_train) for layer in model.layers])
File "~/.venv/lib/python3.8/site-packages/thinc/model.py", line 310, in __call__
return self._func(self, X, is_train=is_train)
File "~/.venv/lib/python3.8/site-packages/thinc/layers/chain.py", line 54, in forward
Y, inc_layer_grad = layer(X, is_train=is_train)
File "~/.venv/lib/python3.8/site-packages/thinc/model.py", line 310, in __call__
return self._func(self, X, is_train=is_train)
File "~/.venv/lib/python3.8/site-packages/thinc/layers/hashembed.py", line 72, in forward
output = model.ops.gather_add(vectors, keys)
File "thinc/backends/numpy_ops.pyx", line 460, in thinc.backends.numpy_ops.NumpyOps.gather_add
File "stringsource", line 660, in View.MemoryView.memoryview_cwrapper
File "stringsource", line 350, in View.MemoryView.memoryview.__cinit__
ValueError: buffer source array is read-only
```
Code to reproduce is:
```
from typing import List

import ray
import spacy

@ray.remote
def generate(nlp, corpus: str) -> List[str]:
doc = list(nlp.pipe([corpus]))[0]
return doc.noun_chunks

if __name__ == "__main__":
ray.init()
nlp = spacy.load(name="en_core_sci_sm")
nlp_ref = ray.put(nlp)
texts = ["sfdfdl?", "dgfhgfhjgj"]
ref_ids = [generate.remote(nlp_ref, text) for text in texts]

while len(ref_ids):
processed, unprocessed = ray.wait(ref_ids)
ref_ids = unprocessed
if processed:
print(ray.get(processed))

```

Note: The error disappears when the model is initialised in action instead of as a ray reference.

I expected it to just work?

## Your Environment
* Operating System: Linux/mac
* Python Version Used: 3.8.10
* spaCy Version Used: '3.7.4'
* Environment Information: