update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
metrics:
|
6 |
+
- precision
|
7 |
+
- recall
|
8 |
+
- f1
|
9 |
+
- accuracy
|
10 |
+
model-index:
|
11 |
+
- name: multiCorp_5e-05_250
|
12 |
+
results: []
|
13 |
+
---
|
14 |
+
|
15 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
16 |
+
should probably proofread and complete it, then remove this comment. -->
|
17 |
+
|
18 |
+
# multiCorp_5e-05_250
|
19 |
+
|
20 |
+
This model is a fine-tuned version of [microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext](https://huggingface.co/microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext) on the None dataset.
|
21 |
+
It achieves the following results on the evaluation set:
|
22 |
+
- Loss: 0.0691
|
23 |
+
- Precision: 0.6768
|
24 |
+
- Recall: 0.5971
|
25 |
+
- F1: 0.6344
|
26 |
+
- Accuracy: 0.9855
|
27 |
+
|
28 |
+
## Model description
|
29 |
+
|
30 |
+
More information needed
|
31 |
+
|
32 |
+
## Intended uses & limitations
|
33 |
+
|
34 |
+
More information needed
|
35 |
+
|
36 |
+
## Training and evaluation data
|
37 |
+
|
38 |
+
More information needed
|
39 |
+
|
40 |
+
## Training procedure
|
41 |
+
|
42 |
+
### Training hyperparameters
|
43 |
+
|
44 |
+
The following hyperparameters were used during training:
|
45 |
+
- learning_rate: 5e-05
|
46 |
+
- train_batch_size: 16
|
47 |
+
- eval_batch_size: 16
|
48 |
+
- seed: 42
|
49 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
50 |
+
- lr_scheduler_type: linear
|
51 |
+
- training_steps: 1500
|
52 |
+
|
53 |
+
### Training results
|
54 |
+
|
55 |
+
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|
56 |
+
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
|
57 |
+
| 0.4757 | 0.34 | 50 | 0.1963 | 0.0 | 0.0 | 0.0 | 0.9740 |
|
58 |
+
| 0.1585 | 0.68 | 100 | 0.1299 | 0.3375 | 0.1049 | 0.1600 | 0.9758 |
|
59 |
+
| 0.1224 | 1.01 | 150 | 0.1121 | 0.3719 | 0.3094 | 0.3377 | 0.9750 |
|
60 |
+
| 0.1003 | 1.35 | 200 | 0.0954 | 0.4297 | 0.3167 | 0.3647 | 0.9791 |
|
61 |
+
| 0.0903 | 1.69 | 250 | 0.0920 | 0.4213 | 0.3063 | 0.3547 | 0.9786 |
|
62 |
+
| 0.0735 | 2.03 | 300 | 0.0795 | 0.4882 | 0.4575 | 0.4724 | 0.9814 |
|
63 |
+
| 0.0636 | 2.36 | 350 | 0.0769 | 0.5188 | 0.4718 | 0.4942 | 0.9820 |
|
64 |
+
| 0.0633 | 2.7 | 400 | 0.0737 | 0.5296 | 0.4926 | 0.5104 | 0.9823 |
|
65 |
+
| 0.0598 | 3.04 | 450 | 0.0735 | 0.5844 | 0.4320 | 0.4968 | 0.9827 |
|
66 |
+
| 0.0479 | 3.38 | 500 | 0.0730 | 0.5797 | 0.5264 | 0.5518 | 0.9831 |
|
67 |
+
| 0.0492 | 3.72 | 550 | 0.0680 | 0.6086 | 0.4978 | 0.5477 | 0.9838 |
|
68 |
+
| 0.041 | 4.05 | 600 | 0.0672 | 0.6190 | 0.5667 | 0.5917 | 0.9842 |
|
69 |
+
| 0.0371 | 4.39 | 650 | 0.0672 | 0.6616 | 0.5693 | 0.6120 | 0.9851 |
|
70 |
+
| 0.0362 | 4.73 | 700 | 0.0665 | 0.6670 | 0.5711 | 0.6153 | 0.9852 |
|
71 |
+
| 0.0334 | 5.07 | 750 | 0.0700 | 0.6532 | 0.5468 | 0.5953 | 0.9848 |
|
72 |
+
| 0.0288 | 5.41 | 800 | 0.0670 | 0.6482 | 0.5628 | 0.6025 | 0.9849 |
|
73 |
+
| 0.0288 | 5.74 | 850 | 0.0698 | 0.6643 | 0.5745 | 0.6162 | 0.9851 |
|
74 |
+
| 0.0263 | 6.08 | 900 | 0.0717 | 0.6827 | 0.5845 | 0.6298 | 0.9856 |
|
75 |
+
| 0.0231 | 6.42 | 950 | 0.0712 | 0.6826 | 0.5702 | 0.6213 | 0.9852 |
|
76 |
+
| 0.0238 | 6.76 | 1000 | 0.0691 | 0.6768 | 0.5971 | 0.6344 | 0.9855 |
|
77 |
+
|
78 |
+
|
79 |
+
### Framework versions
|
80 |
+
|
81 |
+
- Transformers 4.27.4
|
82 |
+
- Pytorch 1.13.1+cu116
|
83 |
+
- Datasets 2.11.0
|
84 |
+
- Tokenizers 0.13.2
|