|
--- |
|
license: mit |
|
base_model: FacebookAI/roberta-large |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: left_as_train_context_roberta-large_20e |
|
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. --> |
|
|
|
# left_as_train_context_roberta-large_20e |
|
|
|
This model is a fine-tuned version of [FacebookAI/roberta-large](https://huggingface.co/FacebookAI/roberta-large) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 3.0530 |
|
- Val Accuracy: 0.7598 |
|
- Val Precision Macro: 0.7129 |
|
- Val Recall Macro: 0.7027 |
|
- Val F1 Macro: 0.7066 |
|
- Val Precision Weighted: 0.7605 |
|
- Val Recall Weighted: 0.7598 |
|
- Val F1 Weighted: 0.7595 |
|
|
|
## 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: 5e-06 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 20 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Val Accuracy | Val Precision Macro | Val Recall Macro | Val F1 Macro | Val Precision Weighted | Val Recall Weighted | Val F1 Weighted | |
|
|:-------------:|:-----:|:-----:|:---------------:|:------------:|:-------------------:|:----------------:|:------------:|:----------------------:|:-------------------:|:---------------:| |
|
| 0.4664 | 1.0 | 3630 | 0.6205 | 0.7544 | 0.7032 | 0.7108 | 0.7050 | 0.7625 | 0.7544 | 0.7564 | |
|
| 0.3597 | 2.0 | 7260 | 0.7307 | 0.7556 | 0.6982 | 0.7237 | 0.7093 | 0.7639 | 0.7556 | 0.7587 | |
|
| 0.2864 | 3.0 | 10890 | 0.8032 | 0.7509 | 0.6944 | 0.7157 | 0.7035 | 0.7605 | 0.7509 | 0.7542 | |
|
| 0.2149 | 4.0 | 14520 | 1.0851 | 0.7581 | 0.7066 | 0.7070 | 0.7061 | 0.7609 | 0.7581 | 0.7588 | |
|
| 0.182 | 5.0 | 18150 | 1.3747 | 0.7503 | 0.6907 | 0.7128 | 0.7004 | 0.7590 | 0.7503 | 0.7535 | |
|
| 0.1306 | 6.0 | 21780 | 1.7668 | 0.7444 | 0.7013 | 0.6941 | 0.6936 | 0.7534 | 0.7444 | 0.7456 | |
|
| 0.1116 | 7.0 | 25410 | 1.7892 | 0.7631 | 0.7199 | 0.6947 | 0.7046 | 0.7617 | 0.7631 | 0.7612 | |
|
| 0.0915 | 8.0 | 29040 | 2.0678 | 0.7565 | 0.7064 | 0.6918 | 0.6979 | 0.7551 | 0.7565 | 0.7553 | |
|
| 0.0696 | 9.0 | 32670 | 2.2576 | 0.7554 | 0.7103 | 0.6981 | 0.7019 | 0.7582 | 0.7554 | 0.7553 | |
|
| 0.0427 | 10.0 | 36300 | 2.2779 | 0.7588 | 0.7117 | 0.6998 | 0.7046 | 0.7589 | 0.7588 | 0.7582 | |
|
| 0.046 | 11.0 | 39930 | 2.4922 | 0.7580 | 0.7066 | 0.7004 | 0.7030 | 0.7581 | 0.7580 | 0.7578 | |
|
| 0.0242 | 12.0 | 43560 | 2.6629 | 0.7623 | 0.7150 | 0.7034 | 0.7085 | 0.7612 | 0.7623 | 0.7615 | |
|
| 0.0251 | 13.0 | 47190 | 2.7028 | 0.7527 | 0.7031 | 0.6977 | 0.6997 | 0.7538 | 0.7527 | 0.7528 | |
|
| 0.0214 | 14.0 | 50820 | 2.7458 | 0.7572 | 0.7104 | 0.7021 | 0.7046 | 0.7599 | 0.7572 | 0.7574 | |
|
| 0.0256 | 15.0 | 54450 | 2.7886 | 0.7552 | 0.7045 | 0.7036 | 0.7032 | 0.7582 | 0.7552 | 0.7560 | |
|
| 0.0134 | 16.0 | 58080 | 2.9100 | 0.7583 | 0.7077 | 0.7005 | 0.7036 | 0.7582 | 0.7583 | 0.7580 | |
|
| 0.0109 | 17.0 | 61710 | 2.8942 | 0.7599 | 0.7137 | 0.6963 | 0.7038 | 0.7580 | 0.7599 | 0.7584 | |
|
| 0.0087 | 18.0 | 65340 | 2.9562 | 0.7602 | 0.7146 | 0.7019 | 0.7072 | 0.7599 | 0.7602 | 0.7595 | |
|
| 0.0019 | 19.0 | 68970 | 3.0273 | 0.7589 | 0.7145 | 0.6999 | 0.7051 | 0.7602 | 0.7589 | 0.7584 | |
|
| 0.0043 | 20.0 | 72600 | 3.0530 | 0.7598 | 0.7129 | 0.7027 | 0.7066 | 0.7605 | 0.7598 | 0.7595 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.38.2 |
|
- Pytorch 2.1.2 |
|
- Datasets 2.18.0 |
|
- Tokenizers 0.15.2 |
|
|
