Instructions to use Mardiyyah/base_pubmedbert_3class_label_all_false with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Mardiyyah/base_pubmedbert_3class_label_all_false with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("token-classification", model="Mardiyyah/base_pubmedbert_3class_label_all_false")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForTokenClassification tokenizer = AutoTokenizer.from_pretrained("Mardiyyah/base_pubmedbert_3class_label_all_false") model = AutoModelForTokenClassification.from_pretrained("Mardiyyah/base_pubmedbert_3class_label_all_false", device_map="auto") - Notebooks
- Google Colab
- Kaggle
base_pubmedbert_3class_label_all_false
This model is a fine-tuned version of microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext on the OTAR3088/CeLLaTe-ner-3class-iob_final dataset. It achieves the following results on the evaluation set:
- Loss: 0.0865
- Precision: 0.6889
- Recall: 0.6857
- Micro F1: 0.6873
- Weighted F1: 0.6836
- Macro F1: 0.6712
- Accuracy: 0.9827
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 3407
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.01
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | Micro F1 | Weighted F1 | Macro F1 | Accuracy |
|---|---|---|---|---|---|---|---|---|---|
| 0.3388 | 1.0 | 263 | 0.0626 | 0.5697 | 0.6406 | 0.6030 | 0.5999 | 0.5872 | 0.9798 |
| 0.0448 | 2.0 | 526 | 0.0599 | 0.6816 | 0.6710 | 0.6763 | 0.6737 | 0.6653 | 0.9826 |
| 0.0307 | 3.0 | 789 | 0.0686 | 0.6465 | 0.6691 | 0.6576 | 0.6531 | 0.6430 | 0.9808 |
| 0.0215 | 4.0 | 1052 | 0.0823 | 0.6786 | 0.6479 | 0.6629 | 0.6557 | 0.6425 | 0.9815 |
| 0.017 | 5.0 | 1315 | 0.0824 | 0.6986 | 0.6323 | 0.6638 | 0.6566 | 0.6430 | 0.9819 |
| 0.0134 | 6.0 | 1578 | 0.0789 | 0.6652 | 0.6940 | 0.6793 | 0.6772 | 0.6671 | 0.9826 |
| 0.0097 | 7.0 | 1841 | 0.0874 | 0.6889 | 0.6857 | 0.6873 | 0.6836 | 0.6712 | 0.9827 |
| 0.0075 | 8.0 | 2104 | 0.0964 | 0.6818 | 0.6793 | 0.6805 | 0.6754 | 0.6624 | 0.9822 |
| 0.0065 | 9.0 | 2367 | 0.0976 | 0.6876 | 0.6654 | 0.6763 | 0.6741 | 0.6622 | 0.9823 |
| 0.0055 | 10.0 | 2630 | 0.1053 | 0.6960 | 0.6710 | 0.6832 | 0.6775 | 0.6645 | 0.9823 |
| 0.0044 | 11.0 | 2893 | 0.1021 | 0.6824 | 0.6811 | 0.6817 | 0.6779 | 0.6630 | 0.9828 |
| 0.0041 | 12.0 | 3156 | 0.1100 | 0.6769 | 0.6682 | 0.6725 | 0.6697 | 0.6550 | 0.9823 |
Framework versions
- Transformers 4.48.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.2
- Tokenizers 0.21.0
- Downloads last month
- 27