Datasets:

Modalities:
Tabular
Text
Formats:
csv
Languages:
English
ArXiv:
Libraries:
Datasets
pandas
License:
File size: 940 Bytes
f1b5d0c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
{
    "model" : "cardiffnlp/twitter-roberta-base",
    "max_length":128,
    "checkpoint_filename" : "./checkpoints/[US]Twitter-RoBERTa-base_CREHate_checkpoint.pth.tar",
    "best_filename" : "./checkpoints/[US]Twitter-RoBERTa-base_CREHate_best.pth.tar",
    "additional_tokens" : ["[US]","[AU]","[GB]","[ZA]","[SG]","@USER","URL"],
    "remove_special_tokens" :true,

    "train_data" : "./data_splits/CREHate_train.csv",
    "val_data" : "./data_splits/CREHate_valid.csv",
    "test_data" : "./data_splits/CREHate_test_res.csv",
    "test_res" : "./data_splits/CREHate_test_res.csv",

    "sent_col" : "Text" ,
    "label_col" : "United_States_Hate" ,
    "num_labels" : 2,

    "test_col" : "Text",
    "test_res_col" : "Twitter-RoBERTa-CREHate-US",
    

    "batch_size" : 32 ,
    "num_workers" : 4,
    "distributed":true,

   "train" : true,
    "evaluate" : true,
    "test" : true,

    "resume" : false,
    "resume_model" : ""
}