|
|
--- |
|
|
language: en |
|
|
tags: |
|
|
- text-classification |
|
|
- pytorch |
|
|
- roberta |
|
|
- self-beliefs |
|
|
- multi-class-classification |
|
|
- multi-label-classification |
|
|
license: mit |
|
|
widget: |
|
|
- text: I am the coolest person I know. |
|
|
--- |
|
|
|
|
|
#### Overview |
|
|
|
|
|
Model trained from [roberta-large](https://huggingface.co/roberta-large) on a dataset of human and LLM annotated self-beliefs for multi-label classification. |
|
|
|
|
|
### Training Details |
|
|
|
|
|
Model training , hyper-parameters, and evaluation can be found in "Capturing Self-Beliefs in Natural Language" by Mangalik et al. 2024 |
|
|
|
|
|
### Inference |
|
|
|
|
|
A sample way to use this model for classification |
|
|
|
|
|
```python |
|
|
from transformers import pipeline |
|
|
huggingface_model = 'sidmangalik/selfBERTa' |
|
|
model = RobertaForSequenceClassification.from_pretrained(huggingface_model) |
|
|
tokenizer = RobertaTokenizerFast.from_pretrained(huggingface_model, max_length = 512, padding="max_length", truncation=True) |
|
|
|
|
|
texts = ["I am the coolest person I know."] |
|
|
|
|
|
inputs = tokenizer(texts, max_length=512, padding="max_length", truncation=True, return_tensors='pt') |
|
|
outputs = model(**inputs) |
|
|
logits = outputs.logits |
|
|
soft_logits = torch.softmax(logits, dim=1).tolist() |
|
|
predicted_classes = np.argmax(soft_logits, axis=1) |
|
|
``` |
