File size: 4,070 Bytes
b3953fc |
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 81 |
---
license: apache-2.0
base_model: tsavage68/UTI_M2_1000steps_1e7rate_SFT
tags:
- trl
- dpo
- generated_from_trainer
model-index:
- name: UTI2_M2_275steps_1e8rate_05beta_CSFTDPO
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. -->
# UTI2_M2_275steps_1e8rate_05beta_CSFTDPO
This model is a fine-tuned version of [tsavage68/UTI_M2_1000steps_1e7rate_SFT](https://huggingface.co/tsavage68/UTI_M2_1000steps_1e7rate_SFT) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6867
- Rewards/chosen: -0.0017
- Rewards/rejected: -0.0154
- Rewards/accuracies: 0.1700
- Rewards/margins: 0.0138
- Logps/rejected: -9.4048
- Logps/chosen: -4.5458
- Logits/rejected: -2.7057
- Logits/chosen: -2.7050
## 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-08
- train_batch_size: 2
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 2
- 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: 275
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
| 0.6931 | 0.3333 | 25 | 0.6941 | 0.0016 | 0.0029 | 0.0700 | -0.0013 | -9.3682 | -4.5393 | -2.7069 | -2.7061 |
| 0.6849 | 0.6667 | 50 | 0.6964 | -0.0083 | -0.0028 | 0.1100 | -0.0055 | -9.3796 | -4.5591 | -2.7057 | -2.7049 |
| 0.6934 | 1.0 | 75 | 0.6896 | -0.0050 | -0.0129 | 0.1300 | 0.0079 | -9.3998 | -4.5524 | -2.7063 | -2.7056 |
| 0.6902 | 1.3333 | 100 | 0.6901 | -0.0010 | -0.0078 | 0.1400 | 0.0068 | -9.3896 | -4.5445 | -2.7066 | -2.7058 |
| 0.6942 | 1.6667 | 125 | 0.6876 | 0.0031 | -0.0090 | 0.1400 | 0.0121 | -9.3920 | -4.5362 | -2.7061 | -2.7053 |
| 0.6823 | 2.0 | 150 | 0.6875 | 0.0028 | -0.0094 | 0.1500 | 0.0122 | -9.3928 | -4.5369 | -2.7062 | -2.7055 |
| 0.6846 | 2.3333 | 175 | 0.6803 | 0.0047 | -0.0227 | 0.1700 | 0.0273 | -9.4193 | -4.5331 | -2.7064 | -2.7057 |
| 0.6766 | 2.6667 | 200 | 0.6874 | -0.0018 | -0.0138 | 0.1600 | 0.0120 | -9.4015 | -4.5461 | -2.7058 | -2.7050 |
| 0.6896 | 3.0 | 225 | 0.6873 | 0.0001 | -0.0126 | 0.1500 | 0.0127 | -9.3992 | -4.5423 | -2.7057 | -2.7050 |
| 0.6895 | 3.3333 | 250 | 0.6867 | -0.0017 | -0.0154 | 0.1700 | 0.0138 | -9.4048 | -4.5458 | -2.7057 | -2.7050 |
| 0.687 | 3.6667 | 275 | 0.6867 | -0.0017 | -0.0154 | 0.1700 | 0.0138 | -9.4048 | -4.5458 | -2.7057 | -2.7050 |
### Framework versions
- Transformers 4.41.2
- Pytorch 2.0.0+cu117
- Datasets 2.19.2
- Tokenizers 0.19.1
|