distilbert-base-uncased__sst2__train-32-0

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: 0.8558
• Accuracy: 0.7183

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.7088	1.0	13	0.6819	0.6154
0.635	2.0	26	0.6318	0.7692
0.547	3.0	39	0.5356	0.7692
0.3497	4.0	52	0.4456	0.6923
0.1979	5.0	65	0.3993	0.7692
0.098	6.0	78	0.3613	0.7692
0.0268	7.0	91	0.3561	0.9231
0.0137	8.0	104	0.3755	0.9231
0.0083	9.0	117	0.4194	0.7692
0.0065	10.0	130	0.4446	0.7692
0.005	11.0	143	0.4527	0.7692
0.0038	12.0	156	0.4645	0.7692
0.0033	13.0	169	0.4735	0.7692
0.0033	14.0	182	0.4874	0.7692
0.0029	15.0	195	0.5041	0.7692
0.0025	16.0	208	0.5148	0.7692
0.0024	17.0	221	0.5228	0.7692

Framework versions

• Transformers 4.15.0
• Pytorch 1.10.2+cu102
• Datasets 1.18.2
• Tokenizers 0.10.3