metadata
license: apache-2.0
library_name: peft
tags:
- trl
- sft
- unsloth
- generated_from_trainer
base_model: unsloth/llama-2-7b-bnb-4bit
model-index:
- name: llama2-envi-tran-10000
results: []
llama2-envi-tran-10000
This model is a fine-tuned version of unsloth/llama-2-7b-bnb-4bit on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.0906
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: 8
- eval_batch_size: 4
- seed: 3407
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 5
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.1238 | 0.33 | 104 | 0.9544 |
| 1.0108 | 0.67 | 208 | 0.9355 |
| 1.0058 | 1.0 | 312 | 0.9277 |
| 0.9212 | 1.33 | 416 | 0.9363 |
| 0.9255 | 1.66 | 520 | 0.9323 |
| 0.9222 | 2.0 | 624 | 0.9275 |
| 0.7975 | 2.33 | 728 | 0.9634 |
| 0.802 | 2.66 | 832 | 0.9622 |
| 0.7979 | 3.0 | 936 | 0.9654 |
| 0.6747 | 3.33 | 1040 | 1.0305 |
| 0.6646 | 3.66 | 1144 | 1.0320 |
| 0.6697 | 3.99 | 1248 | 1.0257 |
| 0.5873 | 4.33 | 1352 | 1.0879 |
| 0.5792 | 4.66 | 1456 | 1.0902 |
| 0.5819 | 4.99 | 1560 | 1.0906 |
Framework versions
- PEFT 0.10.0
- Transformers 4.39.3
- Pytorch 2.2.2+cu121
- Datasets 2.16.0
- Tokenizers 0.15.2