update model card README.md
Browse files
README.md
CHANGED
@@ -2,11 +2,6 @@
|
|
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_LabelNorm_0404
|
12 |
results: []
|
@@ -17,13 +12,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
17 |
|
18 |
# multiCorp_5e-05_LabelNorm_0404
|
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
|
21 |
-
It achieves the following results on the evaluation set:
|
22 |
-
- Loss: 0.0340
|
23 |
-
- Precision: 0.6966
|
24 |
-
- Recall: 0.6039
|
25 |
-
- F1: 0.6469
|
26 |
-
- Accuracy: 0.9907
|
27 |
|
28 |
## Model description
|
29 |
|
@@ -50,41 +39,6 @@ The following hyperparameters were used during training:
|
|
50 |
- lr_scheduler_type: linear
|
51 |
- training_steps: 2000
|
52 |
|
53 |
-
### Training results
|
54 |
-
|
55 |
-
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|
56 |
-
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
|
57 |
-
| 0.2699 | 0.08 | 25 | 0.1037 | 0.0 | 0.0 | 0.0 | 0.9838 |
|
58 |
-
| 0.0844 | 0.15 | 50 | 0.0821 | 0.0 | 0.0 | 0.0 | 0.9838 |
|
59 |
-
| 0.0776 | 0.23 | 75 | 0.0761 | 0.0 | 0.0 | 0.0 | 0.9838 |
|
60 |
-
| 0.0767 | 0.31 | 100 | 0.0680 | 0.0 | 0.0 | 0.0 | 0.9838 |
|
61 |
-
| 0.0541 | 0.39 | 125 | 0.0602 | 0.6392 | 0.2054 | 0.3110 | 0.9856 |
|
62 |
-
| 0.0586 | 0.46 | 150 | 0.0553 | 0.7209 | 0.1447 | 0.2411 | 0.9851 |
|
63 |
-
| 0.051 | 0.54 | 175 | 0.0508 | 0.3651 | 0.4265 | 0.3934 | 0.9837 |
|
64 |
-
| 0.0325 | 0.62 | 200 | 0.0449 | 0.4337 | 0.3307 | 0.3753 | 0.9858 |
|
65 |
-
| 0.0429 | 0.7 | 225 | 0.0454 | 0.5562 | 0.4272 | 0.4833 | 0.9874 |
|
66 |
-
| 0.0357 | 0.77 | 250 | 0.0448 | 0.5351 | 0.3502 | 0.4233 | 0.9855 |
|
67 |
-
| 0.0542 | 0.85 | 275 | 0.0420 | 0.6611 | 0.2778 | 0.3912 | 0.9872 |
|
68 |
-
| 0.0396 | 0.93 | 300 | 0.0379 | 0.5781 | 0.4755 | 0.5218 | 0.9882 |
|
69 |
-
| 0.0354 | 1.01 | 325 | 0.0417 | 0.5083 | 0.6195 | 0.5584 | 0.9877 |
|
70 |
-
| 0.0326 | 1.08 | 350 | 0.0378 | 0.5295 | 0.4677 | 0.4967 | 0.9875 |
|
71 |
-
| 0.0277 | 1.16 | 375 | 0.0413 | 0.7018 | 0.3424 | 0.4603 | 0.9882 |
|
72 |
-
| 0.0429 | 1.24 | 400 | 0.0431 | 0.4924 | 0.5767 | 0.5312 | 0.9869 |
|
73 |
-
| 0.025 | 1.32 | 425 | 0.0368 | 0.6256 | 0.5331 | 0.5756 | 0.9891 |
|
74 |
-
| 0.0364 | 1.39 | 450 | 0.0353 | 0.6193 | 0.5354 | 0.5743 | 0.9882 |
|
75 |
-
| 0.0321 | 1.47 | 475 | 0.0366 | 0.6695 | 0.4949 | 0.5691 | 0.9895 |
|
76 |
-
| 0.0238 | 1.55 | 500 | 0.0340 | 0.5968 | 0.5230 | 0.5574 | 0.9895 |
|
77 |
-
| 0.0289 | 1.63 | 525 | 0.0320 | 0.6191 | 0.5907 | 0.6045 | 0.9900 |
|
78 |
-
| 0.0272 | 1.7 | 550 | 0.0325 | 0.5938 | 0.6257 | 0.6093 | 0.9898 |
|
79 |
-
| 0.028 | 1.78 | 575 | 0.0316 | 0.6309 | 0.6 | 0.6151 | 0.9903 |
|
80 |
-
| 0.025 | 1.86 | 600 | 0.0364 | 0.6781 | 0.3852 | 0.4913 | 0.9891 |
|
81 |
-
| 0.0235 | 1.93 | 625 | 0.0323 | 0.6543 | 0.6467 | 0.6505 | 0.9903 |
|
82 |
-
| 0.0256 | 2.01 | 650 | 0.0320 | 0.7196 | 0.5292 | 0.6099 | 0.9904 |
|
83 |
-
| 0.0182 | 2.09 | 675 | 0.0337 | 0.6773 | 0.6436 | 0.6600 | 0.9910 |
|
84 |
-
| 0.0179 | 2.17 | 700 | 0.0319 | 0.6592 | 0.6218 | 0.6400 | 0.9905 |
|
85 |
-
| 0.0171 | 2.24 | 725 | 0.0340 | 0.6966 | 0.6039 | 0.6469 | 0.9907 |
|
86 |
-
|
87 |
-
|
88 |
### Framework versions
|
89 |
|
90 |
- Transformers 4.27.4
|
|
|
2 |
license: mit
|
3 |
tags:
|
4 |
- generated_from_trainer
|
|
|
|
|
|
|
|
|
|
|
5 |
model-index:
|
6 |
- name: multiCorp_5e-05_LabelNorm_0404
|
7 |
results: []
|
|
|
12 |
|
13 |
# multiCorp_5e-05_LabelNorm_0404
|
14 |
|
15 |
+
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 an unknown dataset.
|
|
|
|
|
|
|
|
|
|
|
|
|
16 |
|
17 |
## Model description
|
18 |
|
|
|
39 |
- lr_scheduler_type: linear
|
40 |
- training_steps: 2000
|
41 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
42 |
### Framework versions
|
43 |
|
44 |
- Transformers 4.27.4
|