|
--- |
|
license: apache-2.0 |
|
library_name: peft |
|
tags: |
|
- trl |
|
- sft |
|
- generated_from_trainer |
|
base_model: teknium/OpenHermes-2.5-Mistral-7B |
|
model-index: |
|
- name: paper_imp_new_format_1 |
|
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. --> |
|
|
|
# paper_imp_new_format_1 |
|
|
|
This model is a fine-tuned version of [teknium/OpenHermes-2.5-Mistral-7B](https://huggingface.co/teknium/OpenHermes-2.5-Mistral-7B) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.6201 |
|
|
|
## 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: 0.0002 |
|
- train_batch_size: 1 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: reduce_lr_on_plateau |
|
- lr_scheduler_warmup_steps: 200 |
|
- num_epochs: 10 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 0.99 | 0.15 | 4 | 1.0009 | |
|
| 1.0287 | 0.3 | 8 | 0.9440 | |
|
| 0.9432 | 0.44 | 12 | 0.8927 | |
|
| 0.829 | 0.59 | 16 | 0.8395 | |
|
| 0.7702 | 0.74 | 20 | 0.7773 | |
|
| 0.6886 | 0.89 | 24 | 0.7180 | |
|
| 0.5844 | 1.04 | 28 | 0.6774 | |
|
| 0.6016 | 1.19 | 32 | 0.6593 | |
|
| 0.5237 | 1.33 | 36 | 0.6484 | |
|
| 0.5505 | 1.48 | 40 | 0.6418 | |
|
| 0.669 | 1.63 | 44 | 0.6350 | |
|
| 0.4595 | 1.78 | 48 | 0.6320 | |
|
| 0.5486 | 1.93 | 52 | 0.6290 | |
|
| 0.539 | 2.07 | 56 | 0.6242 | |
|
| 0.5089 | 2.22 | 60 | 0.6188 | |
|
| 0.6012 | 2.37 | 64 | 0.6195 | |
|
| 0.4286 | 2.52 | 68 | 0.6147 | |
|
| 0.4335 | 2.67 | 72 | 0.6154 | |
|
| 0.4584 | 2.81 | 76 | 0.6154 | |
|
| 0.4597 | 2.96 | 80 | 0.6115 | |
|
| 0.421 | 3.11 | 84 | 0.6076 | |
|
| 0.4541 | 3.26 | 88 | 0.6131 | |
|
| 0.3533 | 3.41 | 92 | 0.6199 | |
|
| 0.4007 | 3.56 | 96 | 0.6142 | |
|
| 0.3876 | 3.7 | 100 | 0.6104 | |
|
| 0.4258 | 3.85 | 104 | 0.6073 | |
|
| 0.4537 | 4.0 | 108 | 0.6059 | |
|
| 0.3486 | 4.15 | 112 | 0.6098 | |
|
| 0.3399 | 4.3 | 116 | 0.6172 | |
|
| 0.3826 | 4.44 | 120 | 0.6215 | |
|
| 0.3813 | 4.59 | 124 | 0.6210 | |
|
| 0.3835 | 4.74 | 128 | 0.6192 | |
|
| 0.3864 | 4.89 | 132 | 0.6173 | |
|
| 0.3216 | 5.04 | 136 | 0.6147 | |
|
| 0.3923 | 5.19 | 140 | 0.6151 | |
|
| 0.3757 | 5.33 | 144 | 0.6158 | |
|
| 0.2861 | 5.48 | 148 | 0.6164 | |
|
| 0.3838 | 5.63 | 152 | 0.6174 | |
|
| 0.3669 | 5.78 | 156 | 0.6180 | |
|
| 0.3405 | 5.93 | 160 | 0.6188 | |
|
| 0.3355 | 6.07 | 164 | 0.6188 | |
|
| 0.3533 | 6.22 | 168 | 0.6193 | |
|
| 0.4245 | 6.37 | 172 | 0.6191 | |
|
| 0.3257 | 6.52 | 176 | 0.6196 | |
|
| 0.3361 | 6.67 | 180 | 0.6198 | |
|
| 0.3595 | 6.81 | 184 | 0.6198 | |
|
| 0.3075 | 6.96 | 188 | 0.6198 | |
|
| 0.4133 | 7.11 | 192 | 0.6199 | |
|
| 0.3341 | 7.26 | 196 | 0.6197 | |
|
| 0.385 | 7.41 | 200 | 0.6197 | |
|
| 0.3254 | 7.56 | 204 | 0.6198 | |
|
| 0.385 | 7.7 | 208 | 0.6199 | |
|
| 0.3332 | 7.85 | 212 | 0.6197 | |
|
| 0.2648 | 8.0 | 216 | 0.6198 | |
|
| 0.3602 | 8.15 | 220 | 0.6203 | |
|
| 0.4129 | 8.3 | 224 | 0.6200 | |
|
| 0.3716 | 8.44 | 228 | 0.6196 | |
|
| 0.3152 | 8.59 | 232 | 0.6200 | |
|
| 0.3463 | 8.74 | 236 | 0.6199 | |
|
| 0.3357 | 8.89 | 240 | 0.6198 | |
|
| 0.3029 | 9.04 | 244 | 0.6199 | |
|
| 0.3679 | 9.19 | 248 | 0.6200 | |
|
| 0.3911 | 9.33 | 252 | 0.6201 | |
|
| 0.2844 | 9.48 | 256 | 0.6199 | |
|
| 0.2584 | 9.63 | 260 | 0.6199 | |
|
| 0.3966 | 9.78 | 264 | 0.6200 | |
|
| 0.3423 | 9.93 | 268 | 0.6201 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.7.1 |
|
- Transformers 4.36.2 |
|
- Pytorch 2.1.2+cu121 |
|
- Datasets 2.14.6 |
|
- Tokenizers 0.15.0 |
