metadata
license: mit
base_model: mikhail-panzo/malay_full_checkpoint
tags:
- generated_from_trainer
model-index:
- name: fil_checkpoint
results: []
fil_checkpoint
This model is a fine-tuned version of mikhail-panzo/malay_full_checkpoint on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.3935
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: 0.0001
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 2000
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.4871 | 6.76 | 500 | 0.4428 |
| 0.4518 | 13.51 | 1000 | 0.4171 |
| 0.4426 | 20.27 | 1500 | 0.4133 |
| 0.4355 | 27.03 | 2000 | 0.4077 |
| 0.425 | 33.78 | 2500 | 0.4013 |
| 0.4167 | 40.54 | 3000 | 0.4020 |
| 0.4037 | 47.3 | 3500 | 0.4051 |
| 0.3933 | 54.05 | 4000 | 0.3945 |
| 0.3875 | 60.81 | 4500 | 0.3928 |
| 0.3828 | 67.57 | 5000 | 0.3935 |
Framework versions
- Transformers 4.40.0.dev0
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2