metadata
license: llama3
base_model: meta-llama/Meta-Llama-3-8B
tags:
- generated_from_trainer
model-index:
- name: MSc_llama3_finetuned_model_secondData
results: []
library_name: peft
MSc_llama3_finetuned_model_secondData
This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.5993
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
The following bitsandbytes
quantization config was used during training:
- quant_method: bitsandbytes
- _load_in_8bit: False
- _load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: bfloat16
- load_in_4bit: True
- load_in_8bit: False
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.03
- training_steps: 250
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
3.329 | 1.33 | 10 | 1.8003 |
1.296 | 2.67 | 20 | 1.0774 |
0.9489 | 4.0 | 30 | 0.9022 |
0.7167 | 5.33 | 40 | 0.7270 |
0.552 | 6.67 | 50 | 0.7372 |
0.4766 | 8.0 | 60 | 0.7281 |
0.4153 | 9.33 | 70 | 0.7673 |
0.3614 | 10.67 | 80 | 0.8597 |
0.3238 | 12.0 | 90 | 0.8915 |
0.2923 | 13.33 | 100 | 0.9281 |
0.2648 | 14.67 | 110 | 1.0239 |
0.2483 | 16.0 | 120 | 1.0198 |
0.2311 | 17.33 | 130 | 1.1314 |
0.2196 | 18.67 | 140 | 1.2578 |
0.2109 | 20.0 | 150 | 1.3155 |
0.1997 | 21.33 | 160 | 1.2602 |
0.1927 | 22.67 | 170 | 1.4758 |
0.191 | 24.0 | 180 | 1.4080 |
0.1834 | 25.33 | 190 | 1.4783 |
0.1799 | 26.67 | 200 | 1.5217 |
0.1796 | 28.0 | 210 | 1.5525 |
0.1738 | 29.33 | 220 | 1.5714 |
0.1725 | 30.67 | 230 | 1.5953 |
0.1727 | 32.0 | 240 | 1.5980 |
0.172 | 33.33 | 250 | 1.5993 |
Framework versions
- PEFT 0.4.0
- Transformers 4.38.2
- Pytorch 2.3.1+cu121
- Datasets 2.13.1
- Tokenizers 0.15.2