metadata
license: other
library_name: peft
tags:
- trl
- sft
- generated_from_trainer
base_model: NousResearch/Meta-Llama-3-8B-Instruct
datasets:
- generator
model-index:
- name: llama38binstruct_summarize
results: []
llama38binstruct_summarize
This model is a fine-tuned version of NousResearch/Meta-Llama-3-8B-Instruct on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 1.8328
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: 1
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- lr_scheduler_warmup_steps: 0.03
- training_steps: 100
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.3955 | 1.3158 | 25 | 1.3893 |
| 0.4287 | 2.6316 | 50 | 1.5275 |
| 0.2387 | 3.9474 | 75 | 1.6572 |
| 0.0876 | 5.2632 | 100 | 1.8328 |
Framework versions
- PEFT 0.11.1
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1