update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,93 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
datasets:
|
6 |
+
- glue
|
7 |
+
metrics:
|
8 |
+
- accuracy
|
9 |
+
- f1
|
10 |
+
model-index:
|
11 |
+
- name: distilbert_add_GLUE_Experiment_logit_kd_qqp
|
12 |
+
results:
|
13 |
+
- task:
|
14 |
+
name: Text Classification
|
15 |
+
type: text-classification
|
16 |
+
dataset:
|
17 |
+
name: glue
|
18 |
+
type: glue
|
19 |
+
config: qqp
|
20 |
+
split: validation
|
21 |
+
args: qqp
|
22 |
+
metrics:
|
23 |
+
- name: Accuracy
|
24 |
+
type: accuracy
|
25 |
+
value: 0.6451150136037596
|
26 |
+
- name: F1
|
27 |
+
type: f1
|
28 |
+
value: 0.07444200748290544
|
29 |
+
---
|
30 |
+
|
31 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
32 |
+
should probably proofread and complete it, then remove this comment. -->
|
33 |
+
|
34 |
+
# distilbert_add_GLUE_Experiment_logit_kd_qqp
|
35 |
+
|
36 |
+
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the glue dataset.
|
37 |
+
It achieves the following results on the evaluation set:
|
38 |
+
- Loss: 0.6730
|
39 |
+
- Accuracy: 0.6451
|
40 |
+
- F1: 0.0744
|
41 |
+
- Combined Score: 0.3598
|
42 |
+
|
43 |
+
## Model description
|
44 |
+
|
45 |
+
More information needed
|
46 |
+
|
47 |
+
## Intended uses & limitations
|
48 |
+
|
49 |
+
More information needed
|
50 |
+
|
51 |
+
## Training and evaluation data
|
52 |
+
|
53 |
+
More information needed
|
54 |
+
|
55 |
+
## Training procedure
|
56 |
+
|
57 |
+
### Training hyperparameters
|
58 |
+
|
59 |
+
The following hyperparameters were used during training:
|
60 |
+
- learning_rate: 5e-05
|
61 |
+
- train_batch_size: 256
|
62 |
+
- eval_batch_size: 256
|
63 |
+
- seed: 10
|
64 |
+
- distributed_type: multi-GPU
|
65 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
66 |
+
- lr_scheduler_type: linear
|
67 |
+
- num_epochs: 50
|
68 |
+
- mixed_precision_training: Native AMP
|
69 |
+
|
70 |
+
### Training results
|
71 |
+
|
72 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Combined Score |
|
73 |
+
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:------:|:--------------:|
|
74 |
+
| 0.7968 | 1.0 | 1422 | 0.7159 | 0.6323 | 0.0030 | 0.3176 |
|
75 |
+
| 0.6542 | 2.0 | 2844 | 0.6925 | 0.6338 | 0.0115 | 0.3226 |
|
76 |
+
| 0.5893 | 3.0 | 4266 | 0.6695 | 0.6348 | 0.0172 | 0.3260 |
|
77 |
+
| 0.5538 | 4.0 | 5688 | 0.7068 | 0.6386 | 0.0393 | 0.3390 |
|
78 |
+
| 0.5323 | 5.0 | 7110 | 0.6670 | 0.6500 | 0.1014 | 0.3757 |
|
79 |
+
| 0.5181 | 6.0 | 8532 | 0.6738 | 0.6420 | 0.0573 | 0.3497 |
|
80 |
+
| 0.5082 | 7.0 | 9954 | 0.6623 | 0.6425 | 0.0601 | 0.3513 |
|
81 |
+
| 0.5012 | 8.0 | 11376 | 0.6995 | 0.6412 | 0.0536 | 0.3474 |
|
82 |
+
| 0.4957 | 9.0 | 12798 | 0.6836 | 0.6472 | 0.0858 | 0.3665 |
|
83 |
+
| 0.4911 | 10.0 | 14220 | 0.6778 | 0.6484 | 0.0922 | 0.3703 |
|
84 |
+
| 0.4874 | 11.0 | 15642 | 0.7183 | 0.6415 | 0.0550 | 0.3483 |
|
85 |
+
| 0.484 | 12.0 | 17064 | 0.6730 | 0.6451 | 0.0744 | 0.3598 |
|
86 |
+
|
87 |
+
|
88 |
+
### Framework versions
|
89 |
+
|
90 |
+
- Transformers 4.26.0
|
91 |
+
- Pytorch 1.14.0a0+410ce96
|
92 |
+
- Datasets 2.9.0
|
93 |
+
- Tokenizers 0.13.2
|