Exporting an HuggingFace pipeline

Learn how to export an HuggingFace pipeline

Last updated 12th August, 2020.


HuggingFace is a popular machine learning library supported by OVHcloud ML Serving. This tutorial will cover how to export an HuggingFace pipeline.


Save HuggingFace pipeline

Let's take an example of an HuggingFace pipeline to illustrate, this script leverages PyTorch based models:

import transformers
import json

# Sentiment analysis pipeline
pipeline = transformers.pipeline('sentiment-analysis')

# OR: Question answering pipeline, specifying the checkpoint identifier
pipeline = transformers.pipeline('question-answering', model='distilbert-base-cased-distilled-squad', tokenizer='bert-base-cased')

# OR: Named entity recognition pipeline, passing in a specific model and tokenizer
model = transformers.AutoModelForTokenClassification.from_pretrained('dbmdz/bert-large-cased-finetuned-conll03-english')
tokenizer = transformers.AutoTokenizer.from_pretrained('bert-base-cased')
pipeline = transformers.pipeline('ner', model=model, tokenizer=tokenizer)

# Save pipeline
path = 'my_model_dir'
# Save manifest (needed by OVHcloud ML Serving to load your pipeline)
with open(path + '/manifest.json', 'w') as file:
        'type': 'huggingface_pipeline',
        'pipeline_class': type(pipeline).__name__,
        'tokenizer_class': type(pipeline.tokenizer).__name__,
        'model_class': type(pipeline.model).__name__,
    }, file, indent=2)

Your model is now serialized on your local file system in the my_model_dir directory.

The manifest.json should look like:

  "type": "huggingface_pipeline",
  "pipeline_class": "FeatureExtractionPipeline",
  "tokenizer_class": "DistilBertTokenizer",
  "model_class": "DistilBertModel"

Going further

Did you find this guide useful?

Please feel free to give any suggestions in order to improve this documentation.

Whether your feedback is about images, content, or structure, please share it, so that we can improve it together.

Your support requests will not be processed via this form. To do this, please use the "Create a ticket" form.

Thank you. Your feedback has been received.

These guides might also interest you...

OVHcloud Community

Access your community space. Ask questions, search for information, post content, and interact with other OVHcloud Community members.

Discuss with the OVHcloud community