Llama3.1-8b-instruct-SFT-2024-09-20_LoRAs_r256
This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.8291
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-05
- train_batch_size: 6
- eval_batch_size: 1
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- num_epochs: 1.5
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
1.4481 | 0.0793 | 1000 | 1.2518 |
1.2848 | 0.1586 | 2000 | 1.2200 |
1.2297 | 0.2380 | 3000 | 1.1614 |
1.1872 | 0.3173 | 4000 | 1.1351 |
1.152 | 0.3966 | 5000 | 1.1096 |
1.113 | 0.4759 | 6000 | 1.0702 |
1.0937 | 0.5552 | 7000 | 1.0237 |
1.0591 | 0.6346 | 8000 | 1.0028 |
1.0422 | 0.7139 | 9000 | 0.9935 |
1.0197 | 0.7932 | 10000 | 0.9761 |
0.9956 | 0.8725 | 11000 | 0.9550 |
0.9743 | 0.9519 | 12000 | 0.9363 |
0.9069 | 1.0312 | 13000 | 0.9297 |
0.8075 | 1.1105 | 14000 | 0.9187 |
0.8066 | 1.1898 | 15000 | 0.9032 |
0.7871 | 1.2691 | 16000 | 0.8891 |
0.7801 | 1.3485 | 17000 | 0.8747 |
0.7773 | 1.4278 | 18000 | 0.8291 |
Framework versions
- PEFT 0.12.0
- Transformers 4.44.2
- Pytorch 2.0.1+cu118
- Datasets 3.0.0
- Tokenizers 0.19.1
- Downloads last month
- 6
Model tree for ccibeekeoc42/Llama3.1-8b-instruct-SFT-2024-09-20_LoRAs_r256
Base model
meta-llama/Llama-3.1-8B
Finetuned
meta-llama/Llama-3.1-8B-Instruct