TajaKuzman commited on
Commit
fc64d8f
1 Parent(s): a6bd902

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -126,7 +126,7 @@ base_model:
126
  # Multilingual IPTC Media Topic Classifier
127
 
128
  News topic classification model based on [`xlm-roberta-large`](https://huggingface.co/FacebookAI/xlm-roberta-large)
129
- and fine-tuned on a news corpus in 4 languages (Croatian, Slovenian, Catalan and Greek), annotated with the [top-level IPTC
130
  Media Topic NewsCodes labels](https://www.iptc.org/std/NewsCodes/treeview/mediatopic/mediatopic-en-GB.html).
131
 
132
  The model can be used for classification into topic labels from the
@@ -215,7 +215,7 @@ and enriched with information which specific subtopics belong to the top-level t
215
 
216
  ## Training data
217
 
218
- The model was fine-tuned on a training dataset consisting of 15,000 news in four languages (Croatian, Slovenian, Catalan and Greek).
219
  The news texts were extracted from the [MaCoCu-Genre web corpora](http://hdl.handle.net/11356/1969) based on the "News" genre label, predicted with the [X-GENRE classifier](https://huggingface.co/classla/xlm-roberta-base-multilingual-text-genre-classifier).
220
  The training dataset was automatically annotated with the IPTC Media Topic labels by
221
  the [GPT-4o](https://platform.openai.com/docs/models/gpt-4o) model (yielding 0.72 micro-F1 and 0.73 macro-F1 on the test dataset).
 
126
  # Multilingual IPTC Media Topic Classifier
127
 
128
  News topic classification model based on [`xlm-roberta-large`](https://huggingface.co/FacebookAI/xlm-roberta-large)
129
+ and fine-tuned on a [news corpus in 4 languages](http://hdl.handle.net/11356/1991) (Croatian, Slovenian, Catalan and Greek), annotated with the [top-level IPTC
130
  Media Topic NewsCodes labels](https://www.iptc.org/std/NewsCodes/treeview/mediatopic/mediatopic-en-GB.html).
131
 
132
  The model can be used for classification into topic labels from the
 
215
 
216
  ## Training data
217
 
218
+ The model was fine-tuned on the training split of the [EMMediaTopic 1.0 dataset](http://hdl.handle.net/11356/1991) consisting of 15,000 news in four languages (Croatian, Slovenian, Catalan and Greek).
219
  The news texts were extracted from the [MaCoCu-Genre web corpora](http://hdl.handle.net/11356/1969) based on the "News" genre label, predicted with the [X-GENRE classifier](https://huggingface.co/classla/xlm-roberta-base-multilingual-text-genre-classifier).
220
  The training dataset was automatically annotated with the IPTC Media Topic labels by
221
  the [GPT-4o](https://platform.openai.com/docs/models/gpt-4o) model (yielding 0.72 micro-F1 and 0.73 macro-F1 on the test dataset).