metadata
library_name: peft
license: other
base_model: ibm-granite/granite-3.3-8b-instruct
tags:
- llama-factory
- lora
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: factory_granite_results
results: []
factory_granite_results
This model is a fine-tuned version of ibm-granite/granite-3.3-8b-instruct on the train dataset. It achieves the following results on the evaluation set:
- Loss: 0.2523
- Accuracy: 0.9475
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: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- total_eval_batch_size: 8
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.03
- num_epochs: 9.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.4051 | 1.0 | 32 | 0.4305 | 0.8978 |
| 0.2886 | 2.0 | 64 | 0.3232 | 0.9206 |
| 0.2289 | 3.0 | 96 | 0.2742 | 0.9323 |
| 0.1925 | 4.0 | 128 | 0.2514 | 0.9387 |
| 0.1079 | 5.0 | 160 | 0.2456 | 0.9420 |
| 0.0968 | 6.0 | 192 | 0.2410 | 0.9454 |
| 0.0835 | 7.0 | 224 | 0.2464 | 0.9466 |
| 0.0716 | 8.0 | 256 | 0.2516 | 0.9472 |
| 0.0611 | 9.0 | 288 | 0.2523 | 0.9475 |
Framework versions
- PEFT 0.15.2
- Transformers 4.52.4
- Pytorch 2.7.0
- Datasets 3.6.0
- Tokenizers 0.21.1