update model card README.md
Browse files
README.md
CHANGED
|
@@ -16,8 +16,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
-
- Loss: 0.
|
| 20 |
-
- F1: 0.
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
@@ -37,54 +37,23 @@ More information needed
|
|
| 37 |
|
| 38 |
The following hyperparameters were used during training:
|
| 39 |
- learning_rate: 1e-05
|
| 40 |
-
- train_batch_size:
|
| 41 |
-
- eval_batch_size:
|
| 42 |
- seed: 42
|
| 43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
- lr_scheduler_type: linear
|
| 45 |
- lr_scheduler_warmup_steps: 5
|
| 46 |
-
- num_epochs:
|
| 47 |
- mixed_precision_training: Native AMP
|
| 48 |
|
| 49 |
### Training results
|
| 50 |
|
| 51 |
| Training Loss | Epoch | Step | Validation Loss | F1 |
|
| 52 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
| 53 |
-
| 0.
|
| 54 |
-
| 0.
|
| 55 |
-
| 0.
|
| 56 |
-
| 0.
|
| 57 |
-
| 0.2817 | 1.43 | 500 | 0.2579 | 0.8650 |
|
| 58 |
-
| 0.2997 | 1.71 | 600 | 0.2089 | 0.8911 |
|
| 59 |
-
| 0.2784 | 2.0 | 700 | 0.2069 | 0.8818 |
|
| 60 |
-
| 0.2231 | 2.29 | 800 | 0.2233 | 0.8872 |
|
| 61 |
-
| 0.2261 | 2.57 | 900 | 0.1598 | 0.9215 |
|
| 62 |
-
| 0.238 | 2.86 | 1000 | 0.1524 | 0.9137 |
|
| 63 |
-
| 0.2014 | 3.14 | 1100 | 0.1155 | 0.9441 |
|
| 64 |
-
| 0.1669 | 3.43 | 1200 | 0.1203 | 0.9436 |
|
| 65 |
-
| 0.1691 | 3.71 | 1300 | 0.0957 | 0.9566 |
|
| 66 |
-
| 0.1787 | 4.0 | 1400 | 0.0763 | 0.9709 |
|
| 67 |
-
| 0.1277 | 4.29 | 1500 | 0.0696 | 0.9717 |
|
| 68 |
-
| 0.1359 | 4.57 | 1600 | 0.0654 | 0.9734 |
|
| 69 |
-
| 0.1138 | 4.86 | 1700 | 0.0542 | 0.9788 |
|
| 70 |
-
| 0.1057 | 5.14 | 1800 | 0.0587 | 0.9747 |
|
| 71 |
-
| 0.1055 | 5.43 | 1900 | 0.0420 | 0.9843 |
|
| 72 |
-
| 0.0908 | 5.71 | 2000 | 0.0386 | 0.9866 |
|
| 73 |
-
| 0.1094 | 6.0 | 2100 | 0.0328 | 0.9890 |
|
| 74 |
-
| 0.0845 | 6.29 | 2200 | 0.0320 | 0.9885 |
|
| 75 |
-
| 0.0697 | 6.57 | 2300 | 0.0322 | 0.9893 |
|
| 76 |
-
| 0.083 | 6.86 | 2400 | 0.0260 | 0.9912 |
|
| 77 |
-
| 0.0659 | 7.14 | 2500 | 0.0259 | 0.9923 |
|
| 78 |
-
| 0.0745 | 7.43 | 2600 | 0.0304 | 0.9900 |
|
| 79 |
-
| 0.0623 | 7.71 | 2700 | 0.0284 | 0.9912 |
|
| 80 |
-
| 0.0825 | 8.0 | 2800 | 0.0215 | 0.9933 |
|
| 81 |
-
| 0.0414 | 8.29 | 2900 | 0.0222 | 0.9939 |
|
| 82 |
-
| 0.0477 | 8.57 | 3000 | 0.0231 | 0.9940 |
|
| 83 |
-
| 0.0606 | 8.86 | 3100 | 0.0211 | 0.9937 |
|
| 84 |
-
| 0.0616 | 9.14 | 3200 | 0.0190 | 0.9947 |
|
| 85 |
-
| 0.0413 | 9.43 | 3300 | 0.0182 | 0.9950 |
|
| 86 |
-
| 0.0462 | 9.71 | 3400 | 0.0181 | 0.9949 |
|
| 87 |
-
| 0.0473 | 10.0 | 3500 | 0.0182 | 0.9951 |
|
| 88 |
|
| 89 |
|
| 90 |
### Framework versions
|
|
|
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 0.3848
|
| 20 |
+
- F1: 0.8133
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
|
|
| 37 |
|
| 38 |
The following hyperparameters were used during training:
|
| 39 |
- learning_rate: 1e-05
|
| 40 |
+
- train_batch_size: 64
|
| 41 |
+
- eval_batch_size: 64
|
| 42 |
- seed: 42
|
| 43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
- lr_scheduler_type: linear
|
| 45 |
- lr_scheduler_warmup_steps: 5
|
| 46 |
+
- num_epochs: 8
|
| 47 |
- mixed_precision_training: Native AMP
|
| 48 |
|
| 49 |
### Training results
|
| 50 |
|
| 51 |
| Training Loss | Epoch | Step | Validation Loss | F1 |
|
| 52 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
| 53 |
+
| 0.4116 | 1.14 | 400 | 0.3611 | 0.7973 |
|
| 54 |
+
| 0.2742 | 2.29 | 800 | 0.3676 | 0.8053 |
|
| 55 |
+
| 0.2216 | 3.43 | 1200 | 0.3799 | 0.8047 |
|
| 56 |
+
| 0.1832 | 4.57 | 1600 | 0.3848 | 0.8133 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 57 |
|
| 58 |
|
| 59 |
### Framework versions
|