|
--- |
|
license: apache-2.0 |
|
base_model: hZzy/qwen2.5-0.5b-sft-news-IFT |
|
tags: |
|
- alignment-handbook |
|
- ndcg |
|
- trl |
|
- expo |
|
- generated_from_trainer |
|
- trl |
|
- expo |
|
- generated_from_trainer |
|
datasets: |
|
- hZzy/train_pairwise |
|
model-index: |
|
- name: qwen2.5-0.5b-expo-DPO-EXPERIMENT-100-5e6 |
|
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/zhiyuzha-university-of-florida/huggingface/runs/ptp2yd12) |
|
# qwen2.5-0.5b-expo-DPO-EXPERIMENT-100-5e6 |
|
|
|
This model is a fine-tuned version of [hZzy/qwen2.5-0.5b-sft-news-IFT](https://huggingface.co/hZzy/qwen2.5-0.5b-sft-news-IFT) on the hZzy/train_pairwise dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 153.9577 |
|
- Logps: -79.3234 |
|
- Logits: -1.1891 |
|
- Objective: 152.3114 |
|
- Dpo Loss: 152.3114 |
|
- Regularize: 152.3114 |
|
- Ranking Simple: 0.5227 |
|
- Ranking Idealized: 0.5093 |
|
- Ranking Idealized Expo: 0.5093 |
|
|
|
## 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-06 |
|
- train_batch_size: 4 |
|
- eval_batch_size: 4 |
|
- seed: 42 |
|
- distributed_type: multi-GPU |
|
- num_devices: 6 |
|
- gradient_accumulation_steps: 12 |
|
- total_train_batch_size: 288 |
|
- total_eval_batch_size: 24 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_ratio: 0.1 |
|
- num_epochs: 2 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Logps | Logits | Objective | Dpo Loss | Regularize | Ranking Simple | Ranking Idealized | Ranking Idealized Expo | |
|
|:-------------:|:------:|:----:|:---------------:|:--------:|:-------:|:---------:|:--------:|:----------:|:--------------:|:-----------------:|:----------------------:| |
|
| 89.5677 | 0.2834 | 50 | 97.0098 | -93.4757 | -1.4670 | 103.5481 | 103.5481 | 103.5481 | 0.5072 | 0.5093 | 0.5093 | |
|
| 102.7372 | 0.5668 | 100 | 164.4481 | -79.3850 | -1.4159 | 169.0837 | 169.0837 | 169.0837 | 0.5238 | 0.5093 | 0.5093 | |
|
| 86.6457 | 0.8503 | 150 | 159.7297 | -80.3621 | -1.2164 | 155.2103 | 155.2103 | 155.2103 | 0.5279 | 0.5093 | 0.5093 | |
|
| 40.1205 | 1.1337 | 200 | 164.8019 | -78.8446 | -1.1758 | 161.0171 | 161.0171 | 161.0171 | 0.5248 | 0.5093 | 0.5093 | |
|
| 40.2475 | 1.4171 | 250 | 156.8958 | -80.0693 | -1.2420 | 156.9776 | 156.9776 | 156.9776 | 0.5279 | 0.5093 | 0.5093 | |
|
| 24.0056 | 1.7005 | 300 | 154.3221 | -79.4678 | -1.1971 | 153.7111 | 153.7111 | 153.7111 | 0.5238 | 0.5093 | 0.5093 | |
|
| 25.1496 | 1.9839 | 350 | 153.9577 | -79.3234 | -1.1891 | 152.3116 | 152.3116 | 152.3116 | 0.5227 | 0.5093 | 0.5093 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.42.0 |
|
- Pytorch 2.3.0+cu121 |
|
- Datasets 2.19.1 |
|
- Tokenizers 0.19.1 |
|
|
