UTI_L3_1000steps_1e8rate_SFT
This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 2.4667
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: 1e-08
- train_batch_size: 2
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 1000
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
2.4483 | 0.3333 | 25 | 2.4675 |
2.4751 | 0.6667 | 50 | 2.4674 |
2.4867 | 1.0 | 75 | 2.4684 |
2.493 | 1.3333 | 100 | 2.4674 |
2.4343 | 1.6667 | 125 | 2.4683 |
2.3951 | 2.0 | 150 | 2.4668 |
2.4952 | 2.3333 | 175 | 2.4687 |
2.5018 | 2.6667 | 200 | 2.4667 |
2.4918 | 3.0 | 225 | 2.4681 |
2.4504 | 3.3333 | 250 | 2.4672 |
2.4035 | 3.6667 | 275 | 2.4667 |
2.4595 | 4.0 | 300 | 2.4669 |
2.5324 | 4.3333 | 325 | 2.4669 |
2.4547 | 4.6667 | 350 | 2.4671 |
2.5054 | 5.0 | 375 | 2.4661 |
2.4749 | 5.3333 | 400 | 2.4669 |
2.4974 | 5.6667 | 425 | 2.4666 |
2.4473 | 6.0 | 450 | 2.4670 |
2.4679 | 6.3333 | 475 | 2.4667 |
2.4151 | 6.6667 | 500 | 2.4659 |
2.5285 | 7.0 | 525 | 2.4676 |
2.5264 | 7.3333 | 550 | 2.4675 |
2.4917 | 7.6667 | 575 | 2.4665 |
2.4537 | 8.0 | 600 | 2.4658 |
2.4891 | 8.3333 | 625 | 2.4674 |
2.4612 | 8.6667 | 650 | 2.4666 |
2.5735 | 9.0 | 675 | 2.4666 |
2.4547 | 9.3333 | 700 | 2.4668 |
2.4897 | 9.6667 | 725 | 2.4670 |
2.5413 | 10.0 | 750 | 2.4668 |
2.4772 | 10.3333 | 775 | 2.4666 |
2.4411 | 10.6667 | 800 | 2.4667 |
2.5307 | 11.0 | 825 | 2.4667 |
2.5068 | 11.3333 | 850 | 2.4667 |
2.4636 | 11.6667 | 875 | 2.4667 |
2.5263 | 12.0 | 900 | 2.4667 |
2.4637 | 12.3333 | 925 | 2.4667 |
2.4442 | 12.6667 | 950 | 2.4667 |
2.4893 | 13.0 | 975 | 2.4667 |
2.4725 | 13.3333 | 1000 | 2.4667 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.0.0+cu117
- Datasets 2.19.2
- Tokenizers 0.19.1
- Downloads last month
- 3