hBERTv2_new_pretrain_48_emb_com_sst2

This model is a fine-tuned version of gokuls/bert_12_layer_model_v2_complete_training_new_emb_compress_48 on the GLUE SST2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4789
• Accuracy: 0.8050

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: 4e-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
0.4943	1.0	527	0.5216	0.7454
0.3354	2.0	1054	0.5366	0.7913
0.274	3.0	1581	0.5091	0.7982
0.2347	4.0	2108	0.5886	0.7970
0.2094	5.0	2635	0.4789	0.8050
0.1944	6.0	3162	0.5025	0.7993
0.1826	7.0	3689	0.6511	0.7901
0.1642	8.0	4216	0.5241	0.7993
0.1516	9.0	4743	0.6334	0.8016
0.1462	10.0	5270	0.6750	0.7913

Framework versions

• Transformers 4.30.2
• Pytorch 1.14.0a0+410ce96
• Datasets 2.12.0
• Tokenizers 0.13.3