bert_uncased_L-12_H-128_A-2-mlm-multi-emails-hq

This model is a fine-tuned version of google/bert_uncased_L-12_H-128_A-2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.4235
• Accuracy: 0.5780

Model description

This is a 40 MB version of BERT that does surprisingly well!

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 16
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 8.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.9421	0.99	141	2.7769	0.5330
2.772	1.99	282	2.6669	0.5487
2.6997	2.99	423	2.5486	0.5621
2.6281	3.99	564	2.4865	0.5704
2.5626	4.99	705	2.4385	0.5766
2.5504	5.99	846	2.4421	0.5772
2.5434	6.99	987	2.4094	0.5818
2.5174	7.99	1128	2.4235	0.5780

Framework versions

• Transformers 4.27.0.dev0
• Pytorch 2.0.0.dev20230129+cu118
• Datasets 2.8.0
• Tokenizers 0.13.1