hBERTv2_sst2

This model is a fine-tuned version of gokuls/bert_12_layer_model_v2 on the GLUE SST2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.6964
• Accuracy: 0.5092

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
0.6916	1.0	264	0.6999	0.5092
0.6885	2.0	528	0.6978	0.5092
0.6871	3.0	792	0.6984	0.5092
0.6869	4.0	1056	0.6990	0.5092
0.6868	5.0	1320	0.6974	0.5092
0.6869	6.0	1584	0.6980	0.5092
0.6867	7.0	1848	0.6984	0.5092
0.6868	8.0	2112	0.6975	0.5092
0.6868	9.0	2376	0.6964	0.5092
0.6865	10.0	2640	0.6978	0.5092
0.6868	11.0	2904	0.6980	0.5092
0.6865	12.0	3168	0.7001	0.5092
0.6867	13.0	3432	0.6966	0.5092
0.6867	14.0	3696	0.6980	0.5092

Framework versions

• Transformers 4.26.1
• Pytorch 1.14.0a0+410ce96
• Datasets 2.10.1
• Tokenizers 0.13.2