Instructions to use SAVSNET/PetBERT-Tick-v2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use SAVSNET/PetBERT-Tick-v2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="SAVSNET/PetBERT-Tick-v2")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("SAVSNET/PetBERT-Tick-v2") model = AutoModelForSequenceClassification.from_pretrained("SAVSNET/PetBERT-Tick-v2", device_map="auto") - Notebooks
- Google Colab
- Kaggle
File size: 916 Bytes
729744c d3ff115 729744c d3ff115 729744c d3ff115 729744c d3ff115 989326d d3ff115 729744c 989326d 729744c | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 | ---
library_name: transformers
language:
- en
base_model:
- SAVSNET/PetBERT
---
# Model Card for Model ID
PetBERT-Tick-v2 is a BERT-based binary classifier, fine-tuned to detect the presence ot ticks in veterinary electronic health records. It used PetBERT as it's base model, a model domain adapted from BERT-base using 5.1 millon veterinary electronic health records from dogs and cats.
## Labels
- `tick_absent` (0)
- `tick_present` (1)
**Model Performance**
| |precision |recall |f1-score |support |
|------------|------------------|------------------|------------------|------------------|
|0 |0.977|0.987|0.982|12001 |
|1 |0.957|0.927 |0.942 |3841 |
|accuracy | | |0.972 |15842 |
|macro avg |0.967|0.957|0.962|15842 |
|weighted avg|0.972|0.972|0.972|15842 |
Classification Threshold = 0.943
|