bert-tiny-distilled

This model is a fine-tuned version of google/bert_uncased_L-2_H-128_A-2 on the hate_speech18 dataset. It achieves the following results on the evaluation set:

• eval_loss: 0.4745
• eval_accuracy: 0.8556
• eval_runtime: 20.4101
• eval_samples_per_second: 107.202
• eval_steps_per_second: 0.882
• epoch: 1.0
• step: 479

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: 7.474176155523857e-05
• train_batch_size: 16
• eval_batch_size: 128
• seed: 23
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.1.1
• Datasets 2.15.0
• Tokenizers 0.15.0