language: en
tags:
- distilbert-base-uncased
- text-classification
- patient
- doctor
widget:
- text: I've got flu
- text: >-
I prescribe you some drugs and you need to stay at home for a couple of
days
- text: Let's move to the theatre this evening!
distilbert-base-uncased-finetuned-text-classification
This model is a fine-tuned version of distilbert-base-uncased on an unknown dataset.
Fine-tuned DistilBERT-base-uncased for Patient-Doctor Classification
Model Description
DistilBERT is a transformer model that performs text classification. I fine-tuned the model on with the purpose of classifying patient, doctor or neutral content, specifically when text is related to the supposed context. The model predicts 3 classes, which are Patient, Doctor or Neutral.
The model is a fine-tuned version of DistilBERT.
It was fine-tuned on the prepared dataset (https://huggingface.co/datasets/LukeGPT88/text-classification-dataset).
It achieves the following results on the evaluation set:
- Loss: 0.0501
- Accuracy: 0.9861
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: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.115 | 1.0 | 774 | 0.0486 | 0.9864 |
| 0.0301 | 2.0 | 1548 | 0.0501 | 0.9861 |
Framework versions
- Transformers 4.37.0
- Pytorch 2.1.2
- Datasets 2.1.0
- Tokenizers 0.15.1
How to Use
from transformers import pipeline
classifier = pipeline("text-classification", model="LukeGPT88/patient-doctor-text-classifier")
classifier("I see you’ve set aside this special time to humiliate yourself in public.")
Output:
[{'label': 'NEUTRAL', 'score': 0.9890775680541992}]
Contact
Please reach out to luca.flammia@gmail.com if you have any questions or feedback.