Instructions to use Mardiyyah/base_pubmedbert_2class_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_2class_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_2class_label_all_false")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForTokenClassification tokenizer = AutoTokenizer.from_pretrained("Mardiyyah/base_pubmedbert_2class_label_all_false") model = AutoModelForTokenClassification.from_pretrained("Mardiyyah/base_pubmedbert_2class_label_all_false", device_map="auto") - Notebooks
- Google Colab
- Kaggle
base_pubmedbert_2class_label_all_false
This model is a fine-tuned version of microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext on the OTAR3088/CeLLaTe-ner-2class-iob_final dataset. It achieves the following results on the evaluation set:
- Loss: 0.0864
- Precision: 0.7414
- Recall: 0.6986
- Micro F1: 0.7193
- Weighted F1: 0.7188
- Macro F1: 0.7272
- Accuracy: 0.9843
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 |
|---|---|---|---|---|---|---|---|---|---|
| 0.239 | 1.0 | 263 | 0.0620 | 0.5550 | 0.7274 | 0.6296 | 0.6345 | 0.6512 | 0.9783 |
| 0.0388 | 2.0 | 526 | 0.0513 | 0.7028 | 0.6884 | 0.6955 | 0.6951 | 0.7060 | 0.9834 |
| 0.0245 | 3.0 | 789 | 0.0596 | 0.6664 | 0.7060 | 0.6856 | 0.6868 | 0.6991 | 0.9829 |
| 0.0187 | 4.0 | 1052 | 0.0646 | 0.7228 | 0.6865 | 0.7042 | 0.7036 | 0.7112 | 0.9835 |
| 0.0141 | 5.0 | 1315 | 0.0609 | 0.7046 | 0.6967 | 0.7007 | 0.7006 | 0.6993 | 0.9839 |
| 0.0105 | 6.0 | 1578 | 0.0738 | 0.6949 | 0.7033 | 0.6990 | 0.6988 | 0.7046 | 0.9836 |
| 0.0081 | 7.0 | 1841 | 0.0759 | 0.7073 | 0.7172 | 0.7122 | 0.7123 | 0.7202 | 0.9844 |
| 0.0062 | 8.0 | 2104 | 0.0863 | 0.7414 | 0.6986 | 0.7193 | 0.7188 | 0.7272 | 0.9843 |
| 0.005 | 9.0 | 2367 | 0.0991 | 0.7269 | 0.6884 | 0.7071 | 0.7059 | 0.7164 | 0.9835 |
| 0.0043 | 10.0 | 2630 | 0.0971 | 0.7309 | 0.6949 | 0.7124 | 0.7121 | 0.7167 | 0.9838 |
| 0.0037 | 11.0 | 2893 | 0.0885 | 0.7087 | 0.7107 | 0.7097 | 0.7096 | 0.7177 | 0.9842 |
| 0.0029 | 12.0 | 3156 | 0.0937 | 0.6988 | 0.7144 | 0.7065 | 0.7065 | 0.7150 | 0.9838 |
| 0.0026 | 13.0 | 3419 | 0.0986 | 0.7026 | 0.7340 | 0.7179 | 0.7180 | 0.7267 | 0.9835 |
Framework versions
- Transformers 4.48.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.2
- Tokenizers 0.21.0
- Downloads last month
- 25