File size: 1,291 Bytes
b011288 d7fc545 b011288 d7fc545 b011288 06d376e b011288 |
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 |
---
tags:
- trl
- orpo
- generated_from_trainer
model-index:
- name: results
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. -->
# results
This model was trained from scratch on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1578
- Rewards/chosen: -0.0609
- Rewards/rejected: -0.0483
- Rewards/accuracies: 0.5
- Rewards/margins: -0.0126
- Logps/rejected: -0.9658
- Logps/chosen: -1.2181
- Logits/rejected: -1.9100
- Logits/chosen: -2.3033
- Nll Loss: 1.1151
- Log Odds Ratio: -0.8532
- Log Odds Chosen: -0.2694
## 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-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
### Framework versions
- Transformers 4.43.3
- Pytorch 2.4.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1
|