distilbert-base-cased-emotion

Training: The model has been trained using the script provided in the following repository https://github.com/MorenoLaQuatra/transformers-tasks-templates

This model is a fine-tuned version of distilbert-base-cased on emotion dataset. It achieves the following results on the evaluation set:

• Loss: 0.3272
• Accuracy: 0.9235
• F1: 0.9217
• Precision: 0.9224
• Recall: 0.9235

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.2776	1.0	500	0.2954	0.9	0.8957	0.9031	0.9
0.1887	2.0	1000	0.1716	0.934	0.9344	0.9370	0.934
0.119	3.0	1500	0.1614	0.9345	0.9342	0.9377	0.9345
0.1001	4.0	2000	0.2018	0.936	0.9353	0.9359	0.936
0.0704	5.0	2500	0.1925	0.935	0.9349	0.9354	0.935
0.0471	6.0	3000	0.2369	0.938	0.9373	0.9377	0.938
0.0322	7.0	3500	0.2693	0.938	0.9382	0.9392	0.938
0.0137	8.0	4000	0.2926	0.937	0.9371	0.9372	0.937
0.0099	9.0	4500	0.2964	0.9365	0.9362	0.9362	0.9365
0.0114	10.0	5000	0.3044	0.935	0.9349	0.9350	0.935

Framework versions

• Transformers 4.22.1
• Pytorch 1.11.0+cu113
• Datasets 2.0.0
• Tokenizers 0.11.6