BT5153-kaggle-sentiment-model-3000-samples

This model is a fine-tuned version of distilbert-base-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6160
• Accuracy: 0.9270
• F1: 0.9288

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: 2e-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
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.2851	1.0	625	0.2058	0.9216	0.9231
0.1735	2.0	1250	0.2257	0.9244	0.9258
0.121	3.0	1875	0.2907	0.9232	0.9251
0.0525	4.0	2500	0.3607	0.9194	0.9219
0.0381	5.0	3125	0.4109	0.9216	0.9233
0.0257	6.0	3750	0.4142	0.9232	0.9244
0.0192	7.0	4375	0.4321	0.9230	0.9233
0.0126	8.0	5000	0.4745	0.9250	0.9278
0.01	9.0	5625	0.5053	0.9240	0.9246
0.0091	10.0	6250	0.5256	0.9240	0.9267
0.0062	11.0	6875	0.5798	0.9246	0.9255
0.0033	12.0	7500	0.5935	0.9242	0.9262
0.0019	13.0	8125	0.5891	0.9286	0.9303
0.0018	14.0	8750	0.6176	0.9266	0.9287
0.0001	15.0	9375	0.6160	0.9270	0.9288

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.0
• Datasets 2.1.0
• Tokenizers 0.13.2