tiny-vanilla-target-glue-cola

This model is a fine-tuned version of google/bert_uncased_L-2_H-128_A-2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.8745
• Matthews Correlation: 0.0651

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	0.6204	0.0
0.603	3.73	1000	0.6181	0.0
0.5927	5.6	1500	0.6231	0.0194
0.5707	7.46	2000	0.6366	0.0149
0.5399	9.33	2500	0.6549	0.0646
0.5169	11.19	3000	0.6769	0.0736
0.4954	13.06	3500	0.6856	0.0742
0.4724	14.93	4000	0.7246	0.0777
0.4552	16.79	4500	0.7517	0.0767
0.434	18.66	5000	0.7868	0.0832
0.4192	20.52	5500	0.8144	0.0664
0.4092	22.39	6000	0.8025	0.0670
0.3874	24.25	6500	0.8745	0.0651

Framework versions

• Transformers 4.26.0.dev0
• Pytorch 1.13.0+cu116
• Datasets 2.8.1.dev0
• Tokenizers 0.13.2