metadata
license: gemma
library_name: peft
tags:
- alignment-handbook
- trl
- sft
- generated_from_trainer
base_model: google/gemma-7b
datasets:
- chansung/no_robots_only_coding
model-index:
- name: gemma-7b-sft-qlora-1
results: []
gemma-7b-sft-qlora-1
This model is a fine-tuned version of google/gemma-7b on the chansung/no_robots_only_coding dataset. It achieves the following results on the evaluation set:
- Loss: 2.2095
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: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- total_eval_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 23.6212 | 0.91 | 5 | 8.0020 |
| 14.6688 | 2.0 | 11 | 6.8099 |
| 10.8277 | 2.91 | 16 | 6.4585 |
| 10.965 | 4.0 | 22 | 5.2759 |
| 8.3233 | 4.91 | 27 | 1.6939 |
| 2.2795 | 6.0 | 33 | 1.4540 |
| 1.5047 | 6.91 | 38 | 1.3612 |
| 1.3243 | 8.0 | 44 | 1.2886 |
| 1.1264 | 8.91 | 49 | 1.2783 |
| 0.9122 | 10.0 | 55 | 1.2740 |
| 0.8184 | 10.91 | 60 | 1.2854 |
| 0.6918 | 12.0 | 66 | 1.3135 |
| 0.6194 | 12.91 | 71 | 1.3431 |
| 0.5176 | 14.0 | 77 | 1.4737 |
| 0.4514 | 14.91 | 82 | 1.7112 |
| 0.3759 | 16.0 | 88 | 1.8429 |
| 0.3464 | 16.91 | 93 | 1.8994 |
| 0.2681 | 18.0 | 99 | 1.9583 |
| 0.2487 | 18.91 | 104 | 2.1623 |
| 0.2122 | 20.0 | 110 | 2.2136 |
| 0.2036 | 20.91 | 115 | 2.2150 |
| 0.2098 | 22.0 | 121 | 2.2189 |
| 0.1955 | 22.73 | 125 | 2.2095 |
Framework versions
- PEFT 0.7.1
- Transformers 4.39.3
- Pytorch 2.2.2+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2