Hugging Face's logo

pabRomero
/
PubMedBERT-full-finetuned-ner-pablo

Token Classification
Transformers
TensorBoard
Safetensors
bert
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
pabRomero commited on
Commit
ba46a4f
·
verified ·
1 Parent(s): ab71e01

Training complete

Browse files
Files changed (2) hide show
  1. README.md +15 -20
  2. runs/Sep05_12-07-48_83295d15965e/events.out.tfevents.1725538069.83295d15965e.5325.4 +2 -2
README.md CHANGED
@@ -21,11 +21,11 @@ should probably proofread and complete it, then remove this comment. -->
21
 
22
  This model is a fine-tuned version of [microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext](https://huggingface.co/microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext) on the None dataset.
23
  It achieves the following results on the evaluation set:
24
- - Loss: 0.0905
25
- - Precision: 0.8142
26
- - Recall: 0.8048
27
- - F1: 0.8095
28
- - Accuracy: 0.9771
29
 
30
  ## Model description
31
 
@@ -44,32 +44,27 @@ More information needed
44
  ### Training hyperparameters
45
 
46
  The following hyperparameters were used during training:
47
- - learning_rate: 0.0002
48
- - train_batch_size: 16
49
- - eval_batch_size: 16
50
  - seed: 42
51
- - gradient_accumulation_steps: 4
52
- - total_train_batch_size: 64
53
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
54
  - lr_scheduler_type: linear
55
- - lr_scheduler_warmup_ratio: 0.05
56
- - num_epochs: 5
57
  - mixed_precision_training: Native AMP
58
 
59
  ### Training results
60
 
61
- | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
62
- |:-------------:|:------:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
63
- | No log | 0.9970 | 252 | 0.0892 | 0.7631 | 0.7751 | 0.7690 | 0.9751 |
64
- | 0.1853 | 1.9980 | 505 | 0.0802 | 0.8139 | 0.7876 | 0.8005 | 0.9780 |
65
- | 0.1853 | 2.9990 | 758 | 0.0792 | 0.7994 | 0.7984 | 0.7989 | 0.9767 |
66
- | 0.0461 | 4.0 | 1011 | 0.0788 | 0.8134 | 0.8045 | 0.8089 | 0.9780 |
67
- | 0.0461 | 4.9852 | 1260 | 0.0905 | 0.8142 | 0.8048 | 0.8095 | 0.9771 |
68
 
69
 
70
  ### Framework versions
71
 
72
  - Transformers 4.44.2
73
- - Pytorch 2.4.0+cu121
74
  - Datasets 2.21.0
75
  - Tokenizers 0.19.1
 
21
 
22
  This model is a fine-tuned version of [microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext](https://huggingface.co/microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext) on the None dataset.
23
  It achieves the following results on the evaluation set:
24
+ - Loss: 0.0633
25
+ - Precision: 0.8475
26
+ - Recall: 0.8335
27
+ - F1: 0.8404
28
+ - Accuracy: 0.9812
29
 
30
  ## Model description
31
 
 
44
  ### Training hyperparameters
45
 
46
  The following hyperparameters were used during training:
47
+ - learning_rate: 5e-05
48
+ - train_batch_size: 32
49
+ - eval_batch_size: 32
50
  - seed: 42
 
 
51
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
52
  - lr_scheduler_type: linear
53
+ - lr_scheduler_warmup_ratio: 0.1
54
+ - num_epochs: 2
55
  - mixed_precision_training: Native AMP
56
 
57
  ### Training results
58
 
59
+ | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
60
+ |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
61
+ | No log | 1.0 | 231 | 0.0845 | 0.7793 | 0.8007 | 0.7898 | 0.9758 |
62
+ | No log | 2.0 | 462 | 0.0633 | 0.8475 | 0.8335 | 0.8404 | 0.9812 |
 
 
 
63
 
64
 
65
  ### Framework versions
66
 
67
  - Transformers 4.44.2
68
+ - Pytorch 2.4.1+cu121
69
  - Datasets 2.21.0
70
  - Tokenizers 0.19.1
runs/Sep05_12-07-48_83295d15965e/events.out.tfevents.1725538069.83295d15965e.5325.4 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1f62609fad97b649dd04bd2647b1e9f4fa0d947d82849ff9dbcbf44b2eb28c17
3
- size 6233
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f475d7e93da8c051e6c46d755718f30d4b6d8605f5b4ec8d2e12764c843496de
3
+ size 7059