metadata
library_name: peft
tags:
- alignment-handbook
- trl
- dpo
- generated_from_trainer
base_model: NbAiLab/nb-gpt-j-6B-v2
datasets:
- hugodk-sch/aftonposten_title_prefs
model-index:
- name: aftonposten-6b-align-scan
results: []
aftonposten-6b-align-scan
This model is a fine-tuned version of data/ap-gpt-j-6b-sft-qlora-04-08 on the hugodk-sch/aftonposten_title_prefs dataset. It achieves the following results on the evaluation set:
- Loss: 0.4934
- Rewards/chosen: 0.2139
- Rewards/rejected: 0.1872
- Rewards/accuracies: 0.5457
- Rewards/margins: 0.0267
- Logps/rejected: -37.2826
- Logps/chosen: -33.7672
- Logits/rejected: -2.2262
- Logits/chosen: -2.2310
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: 8
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 2
- total_train_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: 1
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.4749 | 0.26 | 100 | 0.4963 | 0.1467 | 0.1303 | 0.5336 | 0.0164 | -37.3537 | -33.8512 | -2.2327 | -2.2375 |
| 0.4376 | 0.52 | 200 | 0.4956 | 0.1959 | 0.1769 | 0.5486 | 0.0191 | -37.2955 | -33.7896 | -2.2291 | -2.2339 |
| 0.3835 | 0.78 | 300 | 0.4950 | 0.2045 | 0.1836 | 0.5245 | 0.0210 | -37.2872 | -33.7789 | -2.2264 | -2.2312 |
Framework versions
- PEFT 0.8.2
- Transformers 4.37.2
- Pytorch 2.1.2+cu121
- Datasets 2.17.0
- Tokenizers 0.15.1