Mara_DistilBert_Pretrained
DistilBERT, a variant of BERT, was employed to pre-trained a Marathi language model from scratch using one million sentences. This compact yet powerful model utilizes a distilled version of BERT's transformer architecture
- Loss: 7.4249
Examples
माझं प्रिय मित्र [MASK] आहे
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Examples
माझं प्रिय मित्र [MASK] आहे
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 512
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 7.9421 | 0.84 | 1000 | 7.4249 |
Framework versions
- Transformers 4.18.0
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.12.1