|
--- |
|
license: apache-2.0 |
|
base_model: riotu-lab/ArabianGPT-03B |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: model_outputs |
|
results: [] |
|
language: |
|
- ar |
|
datasets: |
|
- arcd |
|
pipeline_tag: question-answering |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# model_outputs |
|
|
|
This model is a fine-tuned version of [riotu-lab/ArabianGPT-03B](https://huggingface.co/riotu-lab/ArabianGPT-03B) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 4.8610 |
|
|
|
## 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: 0.0003 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 32 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- num_epochs: 3 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 6.4515 | 0.25 | 10 | 4.5855 | |
|
| 4.8912 | 0.49 | 20 | 4.1608 | |
|
| 4.3524 | 0.74 | 30 | 4.0509 | |
|
| 4.1537 | 0.99 | 40 | 4.0484 | |
|
| 3.6716 | 1.23 | 50 | 4.0211 | |
|
| 3.4284 | 1.48 | 60 | 4.1357 | |
|
| 3.5215 | 1.73 | 70 | 4.2520 | |
|
| 3.4336 | 1.98 | 80 | 4.0270 | |
|
| 2.8886 | 2.22 | 90 | 4.9232 | |
|
| 2.6176 | 2.47 | 100 | 5.0723 | |
|
| 2.5867 | 2.72 | 110 | 4.8623 | |
|
| 2.6076 | 2.96 | 120 | 4.8610 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.39.3 |
|
- Pytorch 2.1.2 |
|
- Datasets 2.18.0 |
|
- Tokenizers 0.15.2 |
