Downloads · 30 days
1
3% of all-time downloads
model_hh_shp4_dpo5
guoyu-zhang/model_hh_shp4_dpo5
model_hh_shp4_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
29
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
- 80bb2579
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_shp4_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: 2.7112
- Rewards/chosen: -9.5668
- Rewards/rejected: -9.8462
- Rewards/accuracies: 0.5400
- Rewards/margins: 0.2794
- Logps/rejected: -266.4585
- Logps/chosen: -255.0426
- Logits/rejected: -0.8301
- Logits/chosen: -0.8200
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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.0505 | 2.67 | 100 | 1.9463 | -7.2393 | -6.9165 | 0.4500 | -0.3228 | -260.5990 | -250.3877 | -0.9466 | -0.9197 |
| 0.0379 | 5.33 | 200 | 3.2214 | -9.5352 | -9.9687 | 0.5 | 0.4335 | -266.7035 | -254.9795 | -0.9120 | -0.8842 |
| 0.0013 | 8.0 | 300 | 2.6505 | -10.4459 | -11.0078 | 0.5300 | 0.5620 | -268.7817 | -256.8008 | -0.7232 | -0.7080 |
| 0.0 | 10.67 | 400 | 2.7047 | -9.5022 | -9.7523 | 0.5300 | 0.2501 | -266.2706 | -254.9136 | -0.8290 | -0.8187 |
| 0.0 | 13.33 | 500 | 2.7070 | -9.5225 | -9.7846 | 0.5400 | 0.2621 | -266.3353 | -254.9541 | -0.8296 | -0.8194 |
| 0.0 | 16.0 | 600 | 2.7121 | -9.5623 | -9.8180 | 0.5400 | 0.2557 | -266.4020 | -255.0337 | -0.8298 | -0.8197 |
| 0.0 | 18.67 | 700 | 2.7255 | -9.5794 | -9.8090 | 0.5300 | 0.2296 | -266.3841 | -255.0679 | -0.8301 | -0.8200 |
| 0.0 | 21.33 | 800 | 2.7158 | -9.5789 | -9.8343 | 0.5300 | 0.2555 | -266.4347 | -255.0669 | -0.8298 | -0.8196 |
| 0.0 | 24.0 | 900 | 2.7341 | -9.5811 | -9.8243 | 0.5300 | 0.2432 | -266.4146 | -255.0712 | -0.8300 | -0.8197 |
| 0.0 | 26.67 | 1000 | 2.7112 | -9.5668 | -9.8462 | 0.5400 | 0.2794 | -266.4585 | -255.0426 | -0.8301 | -0.8200 |
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