DistilRoBERTa-Base-Go-Emotion

Model description:

Training Parameters:

Num examples = 27,135
Num Epochs = 8
Instantaneous batch size per device = 8
Gradient Accumulation steps = 1
Total optimization steps = 43,416

TrainOutput:

'train_loss': 0.09555230289697647, 

Evaluation Output:

 'eval_f1_macro': 0.5073628477609134,
 'eval_f1_weighted': 0.5804294635834008