Hugging Face's logo

IMSyPP
/
hate_speech_multilingual

Text Classification
Safetensors
xlm-roberta
hate-speech
Youtube
Twitter
Model card Files Community
Bojan commited on
Commit
2750f3b
·
verified ·
1 Parent(s): b397691

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +8 -6
README.md CHANGED
@@ -33,16 +33,18 @@ The model classifies each input into one of four distinct classes:
33
  * 2 - offensive
34
  * 3 - violent
35
 
36
- **Training data**
37
- * 103k English Youtube comments
38
- * 119k Italian Youtube comments
39
  * 50k Slovenian Twitter comments
40
 
41
- **Evaluation data**
42
- * 20k English Youtube comments
43
- * 21k Italian Youtube comments
44
  * 10k Slovenian Twitter comments
45
 
 
 
46
  **Fine-tuning hyperparameters**
47
 
48
  num_train_epochs=3,
 
33
  * 2 - offensive
34
  * 3 - violent
35
 
36
+ **Training data***
37
+ * 51k English Youtube comments
38
+ * 60k Italian Youtube comments
39
  * 50k Slovenian Twitter comments
40
 
41
+ **Evaluation data***
42
+ * 10k English Youtube comments
43
+ * 10k Italian Youtube comments
44
  * 10k Slovenian Twitter comments
45
 
46
+ \* each comment is manually labeled by two different annotators
47
+
48
  **Fine-tuning hyperparameters**
49
 
50
  num_train_epochs=3,