metadata

inference: false
license: mit
widget: null
language:
  - en
metrics:
  - mrr
datasets:
  - augmented_codesearchnet

🔥 Augmented Code Model 🔥

This is Augmented Code Model which is a fined-tune model of CodeBERT for processing of similarity between given docstring and code. This model is fined-model based on Augmented Code Corpus with ACS=4.

How to use the model ?

Similar to other huggingface model, you may load the model as follows.


from transformers import AutoTokenizer, AutoModelForSequenceClassification

tokenizer = AutoTokenizer.from_pretrained("Fujitsu/AugCode")

model = AutoModelForSequenceClassification.from_pretrained("Fujitsu/AugCode")

Then you may use model to infer the similarity between a given docstring and code.

Citation

    title={AugmentedCode: Examining the Effects of Natural Language Resources in Code Retrieval Models},
    author={Mehdi Bahrami, N. C. Shrikanth, Yuji Mizobuchi, Lei Liu, Masahiro Fukuyori, Wei-Peng Chen, Kazuki Munakata},
    year={2021},
    eprint={TBA},
    archivePrefix={TBA},
    primaryClass={cs.CL}
}