Extended-mBERT / README.md
morten-j's picture
Update README.md
5b00399 verified
metadata
license: apache-2.0
base_model: google-bert/bert-base-multilingual-uncased
tags:
  - generated_from_trainer
model-index:
  - name: MEHDIE_mBERT
    results: []

MEHDIE_mBERT

This model is a fine-tuned version of google-bert/bert-base-multilingual-uncased on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 1.0261
  • Perplexity: 2.79

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: 64
  • eval_batch_size: 64
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 8
  • total_train_batch_size: 512
  • total_eval_batch_size: 512
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 5.0

Training results

Training Loss Epoch Step Validation Loss
1.5096 1.0 51630 1.2459
1.2498 2.0 103260 1.1339
1.1693 3.0 154890 1.0784
1.1233 4.0 206520 1.0425
1.0951 5.0 258150 1.0263

Framework versions

  • Transformers 4.38.2
  • Pytorch 2.3.0a0+ebedce2
  • Datasets 2.17.1
  • Tokenizers 0.15.2