metadata
library_name: peft
license: other
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
tags:
- base_model:adapter:meta-llama/Meta-Llama-3.1-8B-Instruct
- llama-factory
- lora
- transformers
metrics:
- accuracy
pipeline_tag: text-generation
model-index:
- name: belebele_npi_Deva
results: []
belebele_npi_Deva
This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the belebele_npi_Deva_train dataset. It achieves the following results on the evaluation set:
- Loss: 0.0808
- Accuracy: 0.9740
- Mcq Accuracy: 0.7111
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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 0.1
- num_epochs: 5.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Mcq Accuracy |
|---|---|---|---|---|---|
| 0.0552 | 1.0 | 40 | 0.0544 | 0.9722 | 0.6778 |
| 0.0226 | 2.0 | 80 | 0.0656 | 0.9713 | 0.6667 |
| 0.0034 | 3.0 | 120 | 0.0894 | 0.9744 | 0.7111 |
| 0.0027 | 4.0 | 160 | 0.0852 | 0.9756 | 0.7222 |
| 0.0003 | 5.0 | 200 | 0.0808 | 0.9740 | 0.7111 |
Framework versions
- PEFT 0.18.1
- Transformers 5.2.0
- Pytorch 2.10.0+cu128
- Datasets 4.0.0
- Tokenizers 0.22.2