File size: 3,426 Bytes
dfdee00 a70c0a7 dfdee00 a70c0a7 dfdee00 a70c0a7 dfdee00 a70c0a7 dfdee00 |
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 |
---
base_model: defog/llama-3-sqlcoder-8b
library_name: peft
license: cc-by-sa-4.0
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. -->
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/sert121/huggingface/runs/ubrsk8hu)
# results
This model is a fine-tuned version of [defog/llama-3-sqlcoder-8b](https://huggingface.co/defog/llama-3-sqlcoder-8b) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2568
- Rewards/chosen: -0.0126
- Rewards/rejected: -0.0217
- Rewards/accuracies: 0.8944
- Rewards/margins: 0.0091
- Logps/rejected: -0.2167
- Logps/chosen: -0.1258
- Logits/rejected: 0.1307
- Logits/chosen: 0.1283
- Nll Loss: 0.2132
- Log Odds Ratio: -0.4354
- Log Odds Chosen: 0.6768
## 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: 8e-06
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 10
- 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 | Nll Loss | Log Odds Ratio | Log Odds Chosen |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|:--------:|:--------------:|:---------------:|
| 0.9481 | 0.2 | 72 | 0.9541 | -0.0776 | -0.0797 | 0.7143 | 0.0021 | -0.7975 | -0.7765 | -0.3031 | -0.3167 | 0.8875 | -0.6703 | 0.0480 |
| 0.7313 | 0.4 | 144 | 0.7089 | -0.0551 | -0.0596 | 0.8292 | 0.0045 | -0.5962 | -0.5513 | -0.1005 | -0.1135 | 0.6459 | -0.6312 | 0.1330 |
| 0.547 | 0.6 | 216 | 0.4407 | -0.0292 | -0.0367 | 0.8882 | 0.0075 | -0.3670 | -0.2924 | -0.0064 | -0.0109 | 0.3866 | -0.5408 | 0.3609 |
| 0.2547 | 0.8 | 288 | 0.3018 | -0.0164 | -0.0250 | 0.8882 | 0.0085 | -0.2498 | -0.1644 | 0.0633 | 0.0592 | 0.2551 | -0.4664 | 0.5805 |
| 0.3407 | 1.0 | 360 | 0.2568 | -0.0126 | -0.0217 | 0.8944 | 0.0091 | -0.2167 | -0.1258 | 0.1307 | 0.1283 | 0.2132 | -0.4354 | 0.6768 |
### Framework versions
- PEFT 0.11.1
- Transformers 4.42.1
- Pytorch 2.3.1+cu118
- Datasets 2.20.0
- Tokenizers 0.19.1 |