mobilebert_sa_GLUE_Experiment_logit_kd_mnli

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

• Loss: 1.1966
• Accuracy: 0.6173

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: 128
• eval_batch_size: 128
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.6232	1.0	3068	1.3870	0.5505
1.4341	2.0	6136	1.3186	0.5834
1.3724	3.0	9204	1.2819	0.5943
1.3249	4.0	12272	1.2702	0.5982
1.2788	5.0	15340	1.2359	0.6031
1.2302	6.0	18408	1.2008	0.6193
1.1842	7.0	21476	1.1991	0.6222
1.1441	8.0	24544	1.1839	0.6202
1.1057	9.0	27612	1.1861	0.6244
1.0715	10.0	30680	1.1755	0.6250
1.0386	11.0	33748	1.1972	0.6313
1.0066	12.0	36816	1.2149	0.6276
0.9767	13.0	39884	1.2187	0.6193
0.9482	14.0	42952	1.2004	0.6226
0.921	15.0	46020	1.2093	0.6194

Framework versions

• Transformers 4.26.0
• Pytorch 1.14.0a0+410ce96
• Datasets 2.9.0
• Tokenizers 0.13.2