This model achieves an accuracy of 64% on specific metaphor data of small dataset.
This is build upon bert-base-uncased
this is not yet too much reliable to use in full scale as the metaphors meanings
varies over the context of the sentences and the trained data is specific for only some cases.
If you train it over VU Amsterdam Metaphor Corpus you can get better results in overall.
Usage
You can load this model in your Hugging Face code as follows:
