The model is trained considering the annotation of annotators from India only, on instances in Hindi. The annotations are aggregated using majority voting and then used to train the model.

Training and evaluation data

The model has been trained on the annotation from annotators from India from the MultiPICo dataset (instances in Hindi). The data has been randomly split into a train and a validation set.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e-06
train_batch_size: 16
eval_batch_size: 64
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.0045	1.0	34	0.0042	0.6278	0.3433	0.5	0.4071
0.0041	2.0	68	0.0041	0.5389	0.3168	0.6957	0.4354
0.004	3.0	102	0.0039	0.5333	0.3137	0.6957	0.4324
0.0038	4.0	136	0.0037	0.6	0.3556	0.6957	0.4706
0.0039	5.0	170	0.0038	0.5833	0.3210	0.5652	0.4094
0.0036	6.0	204	0.0032	0.6111	0.3605	0.6739	0.4697
0.003	7.0	238	0.0030	0.6722	0.4	0.5652	0.4685
0.0024	8.0	272	0.0030	0.6778	0.4118	0.6087	0.4912
0.0023	9.0	306	0.0030	0.5	0.3103	0.7826	0.4444

Framework versions

Transformers 4.34.1
Pytorch 2.0.1+cu117
Datasets 2.14.5
Tokenizers 0.14.1

Multilingual-Perspectivist-NLU/irony_hi_India

irony_hi_India

Model description

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Multilingual-Perspectivist-NLU/
irony_hi_India