distilbert_add_GLUE_Experiment_mnli

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

• Loss: 0.9474
• Accuracy: 0.5358

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
1.0947	1.0	1534	1.0842	0.3763
1.0512	2.0	3068	1.0320	0.4501
0.9934	3.0	4602	0.9839	0.4935
0.9689	4.0	6136	0.9703	0.4942
0.953	5.0	7670	0.9731	0.5038
0.9377	6.0	9204	0.9563	0.5152
0.9191	7.0	10738	0.9544	0.5311
0.9014	8.0	12272	0.9629	0.5164
0.883	9.0	13806	0.9817	0.5301
0.865	10.0	15340	0.9691	0.5209
0.8452	11.0	16874	0.9606	0.5456
0.8227	12.0	18408	0.9846	0.5341

Framework versions

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