metadata
base_model: aubmindlab/araelectra-base-generator
tags:
- generated_from_trainer
model-index:
- name: QA_FineTuned_AraElectra
results: []
language:
- ar
QA_FineTuned_AraElectra
This model is a fine-tuned version of aubmindlab/araelectra-base-generator on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 3.3206
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: 0.0002
- train_batch_size: 12
- eval_batch_size: 12
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 48
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 5.5622 | 0.46 | 10 | 4.5956 |
| 4.4125 | 0.92 | 20 | 3.9095 |
| 3.9487 | 1.38 | 30 | 3.7421 |
| 3.7229 | 1.84 | 40 | 3.5886 |
| 3.3851 | 2.3 | 50 | 3.5666 |
| 3.1301 | 2.76 | 60 | 3.4475 |
| 2.9588 | 3.22 | 70 | 3.4111 |
| 2.7213 | 3.68 | 80 | 3.3688 |
| 2.5743 | 4.14 | 90 | 3.3205 |
| 2.3191 | 4.6 | 100 | 3.3206 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.2+cpu
- Datasets 2.18.0
- Tokenizers 0.15.2