mobilebert_add_GLUE_Experiment_logit_kd_cola_256

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

• Loss: 0.6707
• Matthews Correlation: 0.0

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	Matthews Correlation
0.8133	1.0	67	0.6864	0.0
0.797	2.0	134	0.6866	0.0
0.7964	3.0	201	0.6832	0.0
0.7948	4.0	268	0.6827	0.0
0.7934	5.0	335	0.6787	0.0
0.7759	6.0	402	0.6742	0.0
0.761	7.0	469	0.6777	0.0
0.756	8.0	536	0.6754	0.0
0.7471	9.0	603	0.6772	0.0
0.7457	10.0	670	0.6760	0.0
0.7419	11.0	737	0.6707	0.0
0.741	12.0	804	0.6729	0.0
0.7351	13.0	871	0.6836	0.0758
0.7349	14.0	938	0.6769	0.0795
0.7357	15.0	1005	0.6715	0.0315
0.7333	16.0	1072	0.6813	0.0894

Framework versions

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