distilbert_add_GLUE_Experiment_mrpc_256

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

• Loss: 0.5932
• Accuracy: 0.7108
• F1: 0.8234
• Combined Score: 0.7671

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: 256
• eval_batch_size: 256
• seed: 10
• distributed_type: multi-GPU
• 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	F1	Combined Score
0.637	1.0	15	0.6242	0.6838	0.8122	0.7480
0.629	2.0	30	0.6240	0.6838	0.8122	0.7480
0.6302	3.0	45	0.6248	0.6838	0.8122	0.7480
0.63	4.0	60	0.6241	0.6838	0.8122	0.7480
0.6323	5.0	75	0.6240	0.6838	0.8122	0.7480
0.6299	6.0	90	0.6243	0.6838	0.8122	0.7480
0.6325	7.0	105	0.6239	0.6838	0.8122	0.7480
0.6301	8.0	120	0.6239	0.6838	0.8122	0.7480
0.6324	9.0	135	0.6240	0.6838	0.8122	0.7480
0.6293	10.0	150	0.6240	0.6838	0.8122	0.7480
0.6307	11.0	165	0.6239	0.6838	0.8122	0.7480
0.6302	12.0	180	0.6240	0.6838	0.8122	0.7480
0.6338	13.0	195	0.6237	0.6838	0.8122	0.7480
0.6281	14.0	210	0.6225	0.6838	0.8122	0.7480
0.6263	15.0	225	0.6183	0.6838	0.8122	0.7480
0.6017	16.0	240	0.5932	0.7108	0.8234	0.7671
0.5213	17.0	255	0.6146	0.6642	0.7540	0.7091
0.4383	18.0	270	0.6405	0.6912	0.7842	0.7377
0.3903	19.0	285	0.6910	0.6912	0.7872	0.7392
0.363	20.0	300	0.7221	0.6544	0.7374	0.6959
0.3306	21.0	315	0.7583	0.6863	0.7808	0.7335

Framework versions

• Transformers 4.26.0
• Pytorch 1.14.0a0+410ce96
• Datasets 2.8.0
• Tokenizers 0.13.2