metadata
base_model: meta-llama/Llama-2-7b-hf
tags:
- generated_from_trainer
datasets:
- samsum
model-index:
- name: finetuned-test
results: []
finetuned-test
This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on the samsum dataset. It achieves the following results on the evaluation set:
- Loss: 1.8801
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: 0.0005
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- lr_scheduler_warmup_ratio: 0.03
- training_steps: 50
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.9324 | 0.01 | 10 | 1.8002 |
| 1.7211 | 0.02 | 20 | 1.7225 |
| 1.6751 | 0.03 | 30 | 1.7174 |
| 1.5841 | 0.04 | 40 | 1.7560 |
| 1.4824 | 0.05 | 50 | 1.8801 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.1.0+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1