metadata
license: gemma
library_name: peft
tags:
- generated_from_trainer
base_model: google/gemma-2b
model-index:
- name: Gemma2b-APPS
results: []
Gemma2b-APPS
This model is a fine-tuned version of google/gemma-2b on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.2944
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: 5e-06
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 2500
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.2046 | 0.04 | 100 | 1.2184 |
| 1.0393 | 0.08 | 200 | 1.1178 |
| 1.064 | 0.12 | 300 | 1.1963 |
| 1.0558 | 0.16 | 400 | 1.2121 |
| 0.9774 | 0.2 | 500 | 1.1823 |
| 1.0325 | 0.24 | 600 | 1.1829 |
| 1.0572 | 0.28 | 700 | 1.1875 |
| 0.9972 | 0.32 | 800 | 1.1901 |
| 0.9901 | 0.36 | 900 | 1.2049 |
| 1.0587 | 0.4 | 1000 | 1.2187 |
| 1.0643 | 0.44 | 1100 | 1.2270 |
| 1.1095 | 0.48 | 1200 | 1.2266 |
| 1.0179 | 0.52 | 1300 | 1.2468 |
| 1.0581 | 0.56 | 1400 | 1.2480 |
| 1.0722 | 0.6 | 1500 | 1.2642 |
| 1.1467 | 0.64 | 1600 | 1.2622 |
| 1.0536 | 0.68 | 1700 | 1.2816 |
| 1.0757 | 0.72 | 1800 | 1.2840 |
| 1.1504 | 0.76 | 1900 | 1.2884 |
| 1.1494 | 0.8 | 2000 | 1.2870 |
| 1.1206 | 0.84 | 2100 | 1.2888 |
| 1.0454 | 0.88 | 2200 | 1.2938 |
| 1.1487 | 0.92 | 2300 | 1.2944 |
| 1.1521 | 0.96 | 2400 | 1.2945 |
| 1.1155 | 1.0 | 2500 | 1.2944 |
Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2