kapilchauhan
commited on
Commit
•
4735c8a
1
Parent(s):
869fcf1
update model card README.md
Browse files
README.md
CHANGED
@@ -18,7 +18,7 @@ model-index:
|
|
18 |
metrics:
|
19 |
- name: Matthews Correlation
|
20 |
type: matthews_correlation
|
21 |
-
value: 0.
|
22 |
---
|
23 |
|
24 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -28,8 +28,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
28 |
|
29 |
This model is a fine-tuned version of [textattack/bert-base-uncased-CoLA](https://huggingface.co/textattack/bert-base-uncased-CoLA) on the glue dataset.
|
30 |
It achieves the following results on the evaluation set:
|
31 |
-
- Loss:
|
32 |
-
- Matthews Correlation: 0.
|
33 |
|
34 |
## Model description
|
35 |
|
@@ -48,43 +48,21 @@ More information needed
|
|
48 |
### Training hyperparameters
|
49 |
|
50 |
The following hyperparameters were used during training:
|
51 |
-
- learning_rate:
|
52 |
-
- train_batch_size:
|
53 |
-
- eval_batch_size:
|
54 |
- seed: 42
|
55 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
56 |
- lr_scheduler_type: linear
|
57 |
-
- num_epochs:
|
58 |
|
59 |
### Training results
|
60 |
|
61 |
| Training Loss | Epoch | Step | Validation Loss | Matthews Correlation |
|
62 |
|:-------------:|:-----:|:----:|:---------------:|:--------------------:|
|
63 |
-
|
|
64 |
-
|
|
65 |
-
|
|
66 |
-
| 0.1706 | 4.0 | 536 | 0.7076 | 0.5905 |
|
67 |
-
| 0.1706 | 5.0 | 670 | 0.7160 | 0.5930 |
|
68 |
-
| 0.1706 | 6.0 | 804 | 0.9217 | 0.5397 |
|
69 |
-
| 0.1706 | 7.0 | 938 | 0.9129 | 0.5725 |
|
70 |
-
| 0.0562 | 8.0 | 1072 | 1.0039 | 0.5687 |
|
71 |
-
| 0.0562 | 9.0 | 1206 | 1.1554 | 0.5558 |
|
72 |
-
| 0.0562 | 10.0 | 1340 | 1.0369 | 0.5668 |
|
73 |
-
| 0.0562 | 11.0 | 1474 | 1.0916 | 0.5829 |
|
74 |
-
| 0.027 | 12.0 | 1608 | 1.2173 | 0.5931 |
|
75 |
-
| 0.027 | 13.0 | 1742 | 1.2788 | 0.5684 |
|
76 |
-
| 0.027 | 14.0 | 1876 | 1.3221 | 0.5619 |
|
77 |
-
| 0.0113 | 15.0 | 2010 | 1.3900 | 0.5567 |
|
78 |
-
| 0.0113 | 16.0 | 2144 | 1.4093 | 0.5533 |
|
79 |
-
| 0.0113 | 17.0 | 2278 | 1.3916 | 0.5527 |
|
80 |
-
| 0.0113 | 18.0 | 2412 | 1.4848 | 0.5539 |
|
81 |
-
| 0.0081 | 19.0 | 2546 | 1.5134 | 0.5518 |
|
82 |
-
| 0.0081 | 20.0 | 2680 | 1.4968 | 0.5748 |
|
83 |
-
| 0.0081 | 21.0 | 2814 | 1.5581 | 0.5573 |
|
84 |
-
| 0.0081 | 22.0 | 2948 | 1.6561 | 0.5478 |
|
85 |
-
| 0.004 | 23.0 | 3082 | 1.5850 | 0.5648 |
|
86 |
-
| 0.004 | 24.0 | 3216 | 1.6088 | 0.5619 |
|
87 |
-
| 0.004 | 25.0 | 3350 | 1.6116 | 0.5617 |
|
88 |
|
89 |
|
90 |
### Framework versions
|
|
|
18 |
metrics:
|
19 |
- name: Matthews Correlation
|
20 |
type: matthews_correlation
|
21 |
+
value: 0.5755298089385917
|
22 |
---
|
23 |
|
24 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
28 |
|
29 |
This model is a fine-tuned version of [textattack/bert-base-uncased-CoLA](https://huggingface.co/textattack/bert-base-uncased-CoLA) on the glue dataset.
|
30 |
It achieves the following results on the evaluation set:
|
31 |
+
- Loss: 0.8318
|
32 |
+
- Matthews Correlation: 0.5755
|
33 |
|
34 |
## Model description
|
35 |
|
|
|
48 |
### Training hyperparameters
|
49 |
|
50 |
The following hyperparameters were used during training:
|
51 |
+
- learning_rate: 2e-05
|
52 |
+
- train_batch_size: 16
|
53 |
+
- eval_batch_size: 16
|
54 |
- seed: 42
|
55 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
56 |
- lr_scheduler_type: linear
|
57 |
+
- num_epochs: 3
|
58 |
|
59 |
### Training results
|
60 |
|
61 |
| Training Loss | Epoch | Step | Validation Loss | Matthews Correlation |
|
62 |
|:-------------:|:-----:|:----:|:---------------:|:--------------------:|
|
63 |
+
| 0.2949 | 1.0 | 535 | 0.5742 | 0.5219 |
|
64 |
+
| 0.1852 | 2.0 | 1070 | 0.7226 | 0.5573 |
|
65 |
+
| 0.1196 | 3.0 | 1605 | 0.8318 | 0.5755 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
66 |
|
67 |
|
68 |
### Framework versions
|