distilbert_add_GLUE_Experiment_qqp_96

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

• Loss: 0.4726
• Accuracy: 0.7906
• F1: 0.7104
• Combined Score: 0.7505

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
0.5993	1.0	1422	0.5459	0.7243	0.6353	0.6798
0.5167	2.0	2844	0.5176	0.7471	0.6481	0.6976
0.4956	3.0	4266	0.5036	0.7588	0.6463	0.7025
0.4849	4.0	5688	0.5056	0.7546	0.6610	0.7078
0.4762	5.0	7110	0.5127	0.7530	0.6705	0.7118
0.4689	6.0	8532	0.5218	0.7476	0.6754	0.7115
0.4622	7.0	9954	0.4935	0.7661	0.6571	0.7116
0.4554	8.0	11376	0.5039	0.7605	0.6537	0.7071
0.4483	9.0	12798	0.5009	0.7625	0.6732	0.7178
0.4393	10.0	14220	0.4991	0.7594	0.6857	0.7226
0.4293	11.0	15642	0.4857	0.7761	0.6548	0.7155
0.4162	12.0	17064	0.4897	0.7735	0.6935	0.7335
0.4021	13.0	18486	0.4758	0.7822	0.6881	0.7352
0.3871	14.0	19908	0.4801	0.7815	0.7050	0.7433
0.3714	15.0	21330	0.4846	0.7827	0.7111	0.7469
0.3556	16.0	22752	0.4726	0.7906	0.7104	0.7505
0.341	17.0	24174	0.4787	0.7942	0.7047	0.7494
0.3269	18.0	25596	0.4914	0.7884	0.7198	0.7541
0.3127	19.0	27018	0.4774	0.7950	0.7156	0.7553
0.3	20.0	28440	0.4862	0.7965	0.7253	0.7609
0.2885	21.0	29862	0.4982	0.7939	0.7300	0.7620

Framework versions

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