metadata
datasets:
- barilan/blog_authorship_corpus
language:
- en
pipeline_tag: text-classification
base_model:
- microsoft/deberta-v3-small
DeBERTaV3 (small) finetuned on the Blog Authorship Corpus for authorship attribution with 10 authors using the author10 dataset from:
Meisenbacher, Stephen, and Florian Matthes. "Thinking Outside of the Differential Privacy Box: A Case Study in Text Privatization with Language Model Prompting." arXiv preprint arXiv:2410.00751 (2024).
Found in: https://github.com/sjmeis/DPNONDP
{
"epoch": 5.0,
"eval_accuracy": 0.639,
"eval_loss": 0.9551867842674255,
"eval_macro_f1": 0.6359876614042939,
"eval_macro_precision": 0.6469646011112227,
"eval_macro_recall": 0.639,
"eval_micro_f1": 0.639,
"eval_runtime": 282.9465,
"eval_samples_per_second": 7.068,
"eval_steps_per_second": 0.884,
"step": 1875
}