Sentencizer read model from disk

[roelvanderburg]

I re-trained the sentence tagger by way of #6382.

However I do get a:

RegistryError: Cant't find './output_dir/model-best/senter' in registry spacy -> architectures. Available names: spacy-transformers.Tok2VecTransformer.v1

Error when trying to read in the model. I'm trying to load the model from disk with below code.

nlp = spacy.load('nl_core_news_sm')

# Construction via create_pipe with custom model
# config = {"model": {"@architectures": "./output_dir/model-best/senter/model"}}
# senter = nlp.add_pipe(factory_name='sentencizer',name="senter_custom_roel", config=config)

senter = nlp.add_pipe(factory_name='sentencizer', name="senter_custom",config=config)
senter.from_disk("./output_dir/model-best/senter/model")

Your Environment

• spaCy version: 3.0.0rc2
• Platform: macOS-10.16-x86_64-i386-64bit
• Python version: 3.8.5
• Pipelines: nl_core_news_sm (3.0.0a0)

[svlandeg]

The @architectures keyword in a config dictionary typically refers to a registered architecture, and not a file location.

You should be able to just load the whole nlp model back from file with

nlp = spacy.load("outputdir/model-best")

[github-actions]

This issue has been automatically closed because it was answered and there was no follow-up discussion.

[github-actions]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.