distilbert_add_GLUE_Experiment_mnli_96

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: 1.0256
• Accuracy: 0.5004

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
1.0987	1.0	1534	1.0980	0.3545
1.0979	2.0	3068	1.0942	0.3580
1.0897	3.0	4602	1.0896	0.3706
1.0817	4.0	6136	1.0769	0.3991
1.072	5.0	7670	1.0680	0.4146
1.0603	6.0	9204	1.0700	0.4174
1.0515	7.0	10738	1.0655	0.4179
1.0441	8.0	12272	1.0546	0.4335
1.038	9.0	13806	1.0751	0.4059
1.0344	10.0	15340	1.0554	0.4363
1.0275	11.0	16874	1.0736	0.4207
1.0225	12.0	18408	1.0662	0.4295
1.0169	13.0	19942	1.0544	0.4421
1.0111	14.0	21476	1.0635	0.4411
1.0043	15.0	23010	1.0505	0.4567
0.9986	16.0	24544	1.0402	0.4643
0.9925	17.0	26078	1.0531	0.4545
0.9861	18.0	27612	1.0431	0.4675
0.9781	19.0	29146	1.0361	0.4801
0.9673	20.0	30680	1.0301	0.4879
0.9552	21.0	32214	1.0327	0.4908
0.9467	22.0	33748	1.0248	0.5013
0.9396	23.0	35282	1.0297	0.4977
0.9328	24.0	36816	1.0237	0.5025
0.9277	25.0	38350	1.0384	0.5010
0.9228	26.0	39884	1.0374	0.5037
0.918	27.0	41418	1.0242	0.5006
0.9128	28.0	42952	1.0248	0.5060
0.9087	29.0	44486	1.0283	0.5027

Framework versions

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