|
--- |
|
tags: autonlp |
|
language: en |
|
widget: |
|
- text: "I am still waiting on my card?" |
|
datasets: |
|
- banking77 |
|
model-index: |
|
- name: RoBERTa-Banking77 |
|
results: |
|
- task: |
|
name: Text Classification |
|
type: text-classification |
|
dataset: |
|
name: "BANKING77" |
|
type: banking77 |
|
metrics: |
|
- name: Accuracy |
|
type: accuracy |
|
value: 93.51 |
|
- name: Macro F1 |
|
type: macro-f1 |
|
value: 93.49 |
|
- name: Weighted F1 |
|
type: weighted-f1 |
|
value: 93.49 |
|
--- |
|
# `RoBERTa-Banking77` trained using autoNLP |
|
|
|
|
|
- Problem type: Multi-class Classification |
|
|
|
## Validation Metrics |
|
|
|
- Loss: 0.27382662892341614 |
|
- Accuracy: 0.935064935064935 |
|
- Macro F1: 0.934939412967268 |
|
- Micro F1: 0.935064935064935 |
|
- Weighted F1: 0.934939412967268 |
|
- Macro Precision: 0.9372295644352715 |
|
- Micro Precision: 0.935064935064935 |
|
- Weighted Precision: 0.9372295644352717 |
|
- Macro Recall: 0.9350649350649349 |
|
- Micro Recall: 0.935064935064935 |
|
- Weighted Recall: 0.935064935064935 |
|
|
|
|
|
## Usage |
|
|
|
You can use cURL to access this model: |
|
|
|
``` |
|
$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoNLP"}' https://api-inference.huggingface.co/models/philschmid/RoBERTa-Banking77 |
|
``` |
|
|
|
Or Python API: |
|
|
|
```py |
|
from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline |
|
|
|
model_id = 'philschmid/RoBERTa-Banking77' |
|
tokenizer = AutoTokenizer.from_pretrained(model_id) |
|
|
|
model = AutoModelForSequenceClassification.from_pretrained(model_id) |
|
classifier = pipeline('text-classification', tokenizer=tokenizer, model=model) |
|
classifier('What is the base of the exchange rates?') |
|
``` |