metadata
license: llama3
base_model: tsavage68/Summary_L3_1000steps_1e7rate_SFT2
tags:
- trl
- dpo
- generated_from_trainer
model-index:
- name: Summary_L3_300steps_1e7rate_01beta_CSFTDPO
results: []
Summary_L3_300steps_1e7rate_01beta_CSFTDPO
This model is a fine-tuned version of tsavage68/Summary_L3_1000steps_1e7rate_SFT2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6344
- Rewards/chosen: 0.0119
- Rewards/rejected: -0.1548
- Rewards/accuracies: 0.1400
- Rewards/margins: 0.1666
- Logps/rejected: -16.8115
- Logps/chosen: -9.2641
- Logits/rejected: -1.0932
- Logits/chosen: -1.0948
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-07
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 300
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.6896 | 0.2004 | 50 | 0.6887 | 0.0011 | -0.0081 | 0.1300 | 0.0092 | -15.3448 | -9.3720 | -1.0951 | -1.0966 |
| 0.6884 | 0.4008 | 100 | 0.6748 | 0.0029 | -0.0369 | 0.1400 | 0.0397 | -15.6323 | -9.3540 | -1.0944 | -1.0960 |
| 0.6604 | 0.6012 | 150 | 0.6466 | 0.0088 | -0.1100 | 0.1400 | 0.1188 | -16.3635 | -9.2949 | -1.0927 | -1.0944 |
| 0.6427 | 0.8016 | 200 | 0.6367 | 0.0112 | -0.1460 | 0.1400 | 0.1572 | -16.7239 | -9.2705 | -1.0934 | -1.0951 |
| 0.6476 | 1.0020 | 250 | 0.6341 | 0.0133 | -0.1549 | 0.1400 | 0.1682 | -16.8126 | -9.2497 | -1.0935 | -1.0951 |
| 0.6481 | 1.2024 | 300 | 0.6344 | 0.0119 | -0.1548 | 0.1400 | 0.1666 | -16.8115 | -9.2641 | -1.0932 | -1.0948 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.0.0+cu117
- Datasets 2.20.0
- Tokenizers 0.19.1