bert-tiny-target-cola

This model is a fine-tuned version of google/bert_uncased_L-2_H-128_A-2 on the glue dataset. It achieves the following results on the evaluation set:

• Loss: 0.6626
• Matthews Correlation: 0.1458

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: 3e-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: constant
• num_epochs: 200

Training results

Training Loss	Epoch	Step	Validation Loss	Matthews Correlation
0.6124	1.87	500	0.6192	0.0
0.6016	3.73	1000	0.6167	0.0
0.5838	5.6	1500	0.6166	0.0149
0.5555	7.46	2000	0.6344	0.0465
0.5272	9.33	2500	0.6542	0.1399
0.5058	11.19	3000	0.6626	0.1458
0.4791	13.06	3500	0.6868	0.1192
0.4577	14.93	4000	0.7215	0.1230
0.4425	16.79	4500	0.7322	0.1243

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.0+cu116
• Datasets 2.8.0
• Tokenizers 0.13.2