Downloads · 30 days
1
4% of all-time downloads
model_hh_shp4_dpo9
guoyu-zhang/model_hh_shp4_dpo9
model_hh_shp4_dpo9 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
26
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
- e6d20930
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_dpo9
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: 4.3544
- Rewards/chosen: -5.3458
- Rewards/rejected: -4.3841
- Rewards/accuracies: 0.4600
- Rewards/margins: -0.9617
- Logps/rejected: -254.8011
- Logps/chosen: -257.0686
- Logits/rejected: -0.5056
- Logits/chosen: -0.4831
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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.0182 | 2.67 | 100 | 1.9417 | 0.0715 | 0.3404 | 0.4200 | -0.2689 | -249.5517 | -251.0494 | -0.5768 | -0.5403 |
| 0.0554 | 5.33 | 200 | 3.7620 | -6.0611 | -5.3910 | 0.4500 | -0.6701 | -255.9199 | -257.8634 | -0.5245 | -0.4755 |
| 0.0475 | 8.0 | 300 | 5.9980 | -19.5304 | -17.4313 | 0.4600 | -2.0991 | -269.2980 | -272.8293 | -0.7793 | -0.7367 |
| 0.0009 | 10.67 | 400 | 4.6897 | 0.9315 | 2.1688 | 0.4300 | -1.2373 | -247.5201 | -250.0939 | -0.5994 | -0.5636 |
| 0.0 | 13.33 | 500 | 4.3775 | -5.3412 | -4.3620 | 0.4600 | -0.9792 | -254.7766 | -257.0635 | -0.5053 | -0.4827 |
| 0.0 | 16.0 | 600 | 4.3248 | -5.2899 | -4.3683 | 0.4600 | -0.9216 | -254.7836 | -257.0065 | -0.5057 | -0.4828 |
| 0.0 | 18.67 | 700 | 4.3972 | -5.3148 | -4.3036 | 0.4600 | -1.0112 | -254.7117 | -257.0341 | -0.5053 | -0.4828 |
| 0.0 | 21.33 | 800 | 4.3504 | -5.3214 | -4.3905 | 0.4500 | -0.9309 | -254.8082 | -257.0416 | -0.5056 | -0.4828 |
| 0.0 | 24.0 | 900 | 4.3824 | -5.3434 | -4.3767 | 0.4600 | -0.9667 | -254.7929 | -257.0659 | -0.5055 | -0.4829 |
| 0.0 | 26.67 | 1000 | 4.3544 | -5.3458 | -4.3841 | 0.4600 | -0.9617 | -254.8011 | -257.0686 | -0.5056 | -0.4831 |
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