Downloads · 30 days
11
22% of all-time downloads
UTI2_M2_250steps_1e8rate_01beta_CSFTDPO
tsavage68/UTI2_M2_250steps_1e8rate_01beta_CSFTDPO
UTI2_M2_250steps_1e8rate_01beta_CSFTDPO is a text generation model from tsavage68. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as apache-2.0.
should probably proofread and complete it, then remove this comment. --
All-time downloads
50
Public
Parameters
7.2B
14.5 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors29 GB · 100%
At a glance
- Task
- Text Generation
- Library
- transformers
- License
- apache-2.0
- Model type
- mistral
- Access
- Public
- Created
- Jun 7, 2024
- Updated
- Jun 7, 2024
- SHA
- 086ae07b
Try a prompt
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. -->
UTI2_M2_250steps_1e8rate_01beta_CSFTDPO
This model is a fine-tuned version of tsavage68/UTI_M2_1000steps_1e7rate_SFT on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6844
- Rewards/chosen: 0.0019
- Rewards/rejected: -0.0159
- Rewards/accuracies: 0.6700
- Rewards/margins: 0.0178
- Logps/rejected: -39.5152
- Logps/chosen: -19.9022
- Logits/rejected: -2.6813
- Logits/chosen: -2.6787
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: 1e-08
- train_batch_size: 2
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 250
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.6931 | 0.3333 | 25 | 0.6929 | 0.0006 | 0.0000 | 0.1900 | 0.0006 | -39.3560 | -19.9153 | -2.6832 | -2.6806 |
| 0.6898 | 0.6667 | 50 | 0.6940 | -0.0024 | -0.0009 | 0.4700 | -0.0015 | -39.3651 | -19.9455 | -2.6823 | -2.6798 |
| 0.6943 | 1.0 | 75 | 0.6915 | 0.0010 | -0.0025 | 0.5200 | 0.0035 | -39.3809 | -19.9114 | -2.6828 | -2.6803 |
| 0.6889 | 1.3333 | 100 | 0.6916 | 0.0004 | -0.0029 | 0.5100 | 0.0033 | -39.3855 | -19.9178 | -2.6827 | -2.6801 |
| 0.6885 | 1.6667 | 125 | 0.6866 | 0.0008 | -0.0126 | 0.6500 | 0.0134 | -39.4820 | -19.9136 | -2.6819 | -2.6793 |
| 0.6878 | 2.0 | 150 | 0.6861 | 0.0013 | -0.0129 | 0.7000 | 0.0142 | -39.4852 | -19.9089 | -2.6822 | -2.6797 |
| 0.6872 | 2.3333 | 175 | 0.6830 | 0.0027 | -0.0180 | 0.7200 | 0.0207 | -39.5362 | -19.8949 | -2.6821 | -2.6795 |
| 0.6869 | 2.6667 | 200 | 0.6840 | 0.0017 | -0.0169 | 0.7800 | 0.0186 | -39.5253 | -19.9045 | -2.6811 | -2.6785 |
| 0.6888 | 3.0 | 225 | 0.6844 | 0.0019 | -0.0159 | 0.6700 | 0.0178 | -39.5152 | -19.9022 | -2.6813 | -2.6787 |
| 0.6877 | 3.3333 | 250 | 0.6844 | 0.0019 | -0.0159 | 0.6700 | 0.0178 | -39.5152 | -19.9022 | -2.6813 | -2.6787 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.0.0+cu117
- Datasets 2.19.2
- Tokenizers 0.19.1
- Task
- Text Generation
- Library
- transformers
- Type
- mistral
- License
- apache-2.0
- Created
- Jun 7, 2024
- Updated
- Jun 7, 2024