|
--- |
|
license: mit |
|
base_model: roberta-base |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
model-index: |
|
- name: default |
|
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. --> |
|
|
|
# default |
|
|
|
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.2050 |
|
- Accuracy: 0.9350 |
|
|
|
## 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: 128 |
|
- eval_batch_size: 512 |
|
- seed: 42 |
|
- distributed_type: multi-GPU |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 512 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 10000 |
|
- num_epochs: 3 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:| |
|
| 1.0664 | 0.06 | 20 | 1.0651 | 0.4448 | |
|
| 1.0423 | 0.12 | 40 | 1.0188 | 0.5034 | |
|
| 1.0137 | 0.18 | 60 | 0.9871 | 0.5279 | |
|
| 1.0027 | 0.24 | 80 | 0.9889 | 0.5308 | |
|
| 0.9914 | 0.3 | 100 | 0.9763 | 0.5308 | |
|
| 0.9826 | 0.36 | 120 | 0.9713 | 0.5388 | |
|
| 0.9788 | 0.42 | 140 | 0.9766 | 0.5313 | |
|
| 0.984 | 0.48 | 160 | 0.9590 | 0.5398 | |
|
| 0.9694 | 0.54 | 180 | 0.9535 | 0.5423 | |
|
| 0.9676 | 0.6 | 200 | 0.9274 | 0.5672 | |
|
| 0.9753 | 0.66 | 220 | 0.9126 | 0.5736 | |
|
| 0.9557 | 0.72 | 240 | 0.9053 | 0.5760 | |
|
| 0.9508 | 0.78 | 260 | 0.9179 | 0.5767 | |
|
| 0.9355 | 0.84 | 280 | 0.8937 | 0.5892 | |
|
| 0.9 | 0.9 | 300 | 0.8469 | 0.6130 | |
|
| 0.993 | 0.96 | 320 | 0.8615 | 0.6047 | |
|
| 0.8527 | 1.02 | 340 | 0.7896 | 0.6439 | |
|
| 0.966 | 1.08 | 360 | 1.0124 | 0.5316 | |
|
| 0.8441 | 1.14 | 380 | 0.7911 | 0.6489 | |
|
| 0.8226 | 1.2 | 400 | 0.7472 | 0.6700 | |
|
| 0.7948 | 1.26 | 420 | 0.7664 | 0.6581 | |
|
| 0.7428 | 1.32 | 440 | 0.6994 | 0.6992 | |
|
| 0.7109 | 1.38 | 460 | 0.6511 | 0.7284 | |
|
| 0.6882 | 1.44 | 480 | 0.5988 | 0.7577 | |
|
| 0.7296 | 1.5 | 500 | 0.5993 | 0.7564 | |
|
| 0.5677 | 1.57 | 520 | 0.5068 | 0.8126 | |
|
| 0.5096 | 1.63 | 540 | 0.4273 | 0.8520 | |
|
| 0.4452 | 1.69 | 560 | 0.3796 | 0.8722 | |
|
| 0.3836 | 1.75 | 580 | 0.3855 | 0.8757 | |
|
| 0.3783 | 1.81 | 600 | 0.3586 | 0.8894 | |
|
| 0.3496 | 1.87 | 620 | 0.3210 | 0.8972 | |
|
| 0.3585 | 1.93 | 640 | 0.3006 | 0.9035 | |
|
| 0.345 | 1.99 | 660 | 0.3054 | 0.9014 | |
|
| 0.3327 | 2.05 | 680 | 0.3174 | 0.8913 | |
|
| 0.2962 | 2.11 | 700 | 0.2770 | 0.9122 | |
|
| 0.3032 | 2.17 | 720 | 0.2979 | 0.9062 | |
|
| 0.27 | 2.23 | 740 | 0.2973 | 0.8998 | |
|
| 0.2912 | 2.29 | 760 | 0.2467 | 0.9222 | |
|
| 0.2412 | 2.35 | 780 | 0.2761 | 0.9113 | |
|
| 0.2746 | 2.41 | 800 | 0.2410 | 0.9260 | |
|
| 0.2637 | 2.47 | 820 | 0.2447 | 0.9221 | |
|
| 0.2605 | 2.53 | 840 | 0.2475 | 0.9237 | |
|
| 0.2517 | 2.59 | 860 | 0.2590 | 0.9265 | |
|
| 0.2453 | 2.65 | 880 | 0.2248 | 0.9300 | |
|
| 0.2247 | 2.71 | 900 | 0.2285 | 0.9273 | |
|
| 0.2402 | 2.77 | 920 | 0.2304 | 0.9306 | |
|
| 0.2033 | 2.83 | 940 | 0.2228 | 0.9319 | |
|
| 0.2315 | 2.89 | 960 | 0.2275 | 0.9271 | |
|
| 0.2231 | 2.95 | 980 | 0.2115 | 0.9343 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.34.1 |
|
- Pytorch 2.1.0 |
|
- Datasets 2.14.5 |
|
- Tokenizers 0.14.1 |
|
|
