Downloads · 30 days
2
6% of all-time downloads
model_hh_shp2_dpo5
guoyu-zhang/model_hh_shp2_dpo5
model_hh_shp2_dpo5 is a machine learning model from guoyu-zhang. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft.
should probably proofread and complete it, then remove this comment. --
All-time downloads
33
Public
Repo size
12.6 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors25.2 MB · 93%
At a glance
- Library
- peft
- Access
- Public
- Created
- Apr 16, 2024
- Updated
- Apr 16, 2024
- SHA
- 15436de9
Base models
Model card
From the Hugging Face model README
<!-- 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_hh_shp2_dpo5
This model is a fine-tuned version of meta-llama/Llama-2-7b-chat-hf on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 2.6851
- Rewards/chosen: -2.9461
- Rewards/rejected: -3.5981
- Rewards/accuracies: 0.5100
- Rewards/margins: 0.6520
- Logps/rejected: -231.1039
- Logps/chosen: -240.5692
- Logits/rejected: -1.0755
- Logits/chosen: -1.0796
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: 0.0005
- train_batch_size: 4
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 1000
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.0775 | 2.67 | 100 | 1.5649 | -1.7873 | -2.0584 | 0.5400 | 0.2711 | -228.0247 | -238.2517 | -0.8122 | -0.8281 |
| 0.0049 | 5.33 | 200 | 1.6753 | -0.0539 | -1.0216 | 0.5400 | 0.9678 | -225.9511 | -234.7848 | -1.0685 | -1.0774 |
| 0.0012 | 8.0 | 300 | 2.5069 | -2.8261 | -2.9419 | 0.4900 | 0.1158 | -229.7917 | -240.3293 | -1.1074 | -1.1154 |
| 0.0 | 10.67 | 400 | 2.8043 | -3.0513 | -3.6756 | 0.5 | 0.6243 | -231.2590 | -240.7796 | -1.0512 | -1.0557 |
| 0.0 | 13.33 | 500 | 2.7025 | -2.9535 | -3.5803 | 0.5100 | 0.6268 | -231.0683 | -240.5840 | -1.0760 | -1.0802 |
| 0.0 | 16.0 | 600 | 2.6581 | -2.9364 | -3.5947 | 0.5100 | 0.6583 | -231.0972 | -240.5500 | -1.0747 | -1.0789 |
| 0.0 | 18.67 | 700 | 2.6744 | -2.9415 | -3.6188 | 0.5200 | 0.6773 | -231.1454 | -240.5601 | -1.0743 | -1.0785 |
| 0.0 | 21.33 | 800 | 2.6936 | -2.9559 | -3.6011 | 0.5100 | 0.6452 | -231.1101 | -240.5889 | -1.0752 | -1.0793 |
| 0.0 | 24.0 | 900 | 2.6867 | -2.9162 | -3.5792 | 0.5200 | 0.6629 | -231.0661 | -240.5095 | -1.0752 | -1.0798 |
| 0.0 | 26.67 | 1000 | 2.6851 | -2.9461 | -3.5981 | 0.5100 | 0.6520 | -231.1039 | -240.5692 | -1.0755 | -1.0796 |
Framework versions
- PEFT 0.10.0
- Transformers 4.39.1
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2
- Library
- peft
- Created
- Apr 16, 2024
- Updated
- Apr 16, 2024