The model is trained considering the annotation of annotators from France only. The annotations from these annotators 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 France from the MultiPICo dataset. 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
0.0044	1.0	72	0.0042	0.7953	0.0	0.0	0.0
0.0042	2.0	144	0.0041	0.8058	0.5417	0.3333	0.4127
0.0041	3.0	216	0.0039	0.5958	0.3	0.7308	0.4254
0.0037	4.0	288	0.0036	0.5486	0.2824	0.7821	0.4150
0.0034	5.0	360	0.0031	0.7192	0.3782	0.5769	0.4569
0.0033	6.0	432	0.0031	0.6352	0.3129	0.6538	0.4232
0.0027	7.0	504	0.0028	0.6089	0.3216	0.8205	0.4621
0.002	8.0	576	0.0028	0.6614	0.3265	0.6154	0.4267
0.0017	9.0	648	0.0023	0.6562	0.3536	0.8205	0.4942
0.0017	10.0	720	0.0024	0.5906	0.3178	0.8718	0.4658
0.0012	11.0	792	0.0028	0.7192	0.3910	0.6667	0.4929

Framework versions

Transformers 4.34.1
Pytorch 2.0.1+cu117
Datasets 2.14.5
Tokenizers 0.14.1

Multilingual-Perspectivist-NLU/irony_fr_France

irony_fr_France

Model description

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Multilingual-Perspectivist-NLU/
irony_fr_France