File size: 3,756 Bytes
f4fc65a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 |
---
license: mit
base_model: EleutherAI/gpt-neo-125M
tags:
- trl
- dpo
- generated_from_trainer
model-index:
- name: model
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# model
This model is a fine-tuned version of [EleutherAI/gpt-neo-125M](https://huggingface.co/EleutherAI/gpt-neo-125M) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6955
- Rewards/chosen: -0.0079
- Rewards/rejected: -0.0080
- Rewards/accuracies: 0.4813
- Rewards/margins: 0.0001
- Logps/rejected: -478.8612
- Logps/chosen: -494.2958
- Logits/rejected: -18.3633
- Logits/chosen: -18.4819
## 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-07
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- 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: 0.1
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
| 0.6955 | 0.2992 | 100 | 0.6958 | -0.0017 | -0.0008 | 0.4701 | -0.0008 | -478.7900 | -494.2336 | -18.3637 | -18.4824 |
| 0.6906 | 0.5984 | 200 | 0.6962 | -0.0028 | -0.0016 | 0.4744 | -0.0013 | -478.7974 | -494.2453 | -18.3625 | -18.4806 |
| 0.6985 | 0.8975 | 300 | 0.6959 | -0.0222 | -0.0214 | 0.4738 | -0.0008 | -478.9952 | -494.4388 | -18.3624 | -18.4809 |
| 0.6946 | 1.1967 | 400 | 0.6955 | 0.0015 | 0.0015 | 0.4753 | 0.0000 | -478.7664 | -494.2018 | -18.3628 | -18.4811 |
| 0.6946 | 1.4959 | 500 | 0.6960 | -0.0046 | -0.0040 | 0.4791 | -0.0006 | -478.8223 | -494.2634 | -18.3631 | -18.4816 |
| 0.6952 | 1.7951 | 600 | 0.6951 | -0.0047 | -0.0057 | 0.4882 | 0.0011 | -478.8391 | -494.2639 | -18.3636 | -18.4821 |
| 0.6947 | 2.0942 | 700 | 0.6955 | -0.0053 | -0.0056 | 0.4822 | 0.0003 | -478.8379 | -494.2701 | -18.3634 | -18.4820 |
| 0.6995 | 2.3934 | 800 | 0.6948 | -0.0060 | -0.0076 | 0.4918 | 0.0015 | -478.8574 | -494.2774 | -18.3632 | -18.4818 |
| 0.6932 | 2.6926 | 900 | 0.6952 | -0.0080 | -0.0087 | 0.4837 | 0.0008 | -478.8692 | -494.2970 | -18.3633 | -18.4817 |
| 0.6964 | 2.9918 | 1000 | 0.6955 | -0.0079 | -0.0080 | 0.4813 | 0.0001 | -478.8612 | -494.2958 | -18.3633 | -18.4819 |
### Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1
|