---
license: mit
---


### Usage

In Transformers

```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
from scipy.special import softmax, expit

device = "cuda" if torch.cuda.is_available() else "cpu"

model_name="ragarwal/deberta-v3-base-nli-mixer"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

sentence = "During its monthly call, the National Oceanic and Atmospheric Administration warned of \
increased temperatures and low precipitation" 
labels = ["Computer", "Climate Change", "Tablet", "Football", "Artificial Intelligence", "Global Warming"] 

features = tokenizer([[sentence, l] for l in labels], padding=True, truncation=True, return_tensors="pt")

model.eval()
with torch.no_grad():
    scores = model(**features).logits
    print(expit(scores)) #Multi-Label Classification
    print(softmax(scores)) #Single-Label Classification
    
```


In [Sentence-Transformers](https://www.sbert.net/)

```python
from sentence_transformers import CrossEncoder

model_name="ragarwal/deberta-v3-base-nli-mixer"
model = CrossEncoder(model_name, max_length=256)

sentence = "During its monthly call, the National Oceanic and Atmospheric Administration warned of \
increased temperatures and low precipitation" 
labels = ["Computer", "Climate Change", "Tablet", "Football", "Artificial Intelligence", "Global Warming"] 

scores = model.predict([[sentence, l] for l in labels])
print(scores) 
#array([0.04118565, 0.2435827 , 0.03941465, 0.00203637, 0.00501176, 0.1423797], dtype=float32)

```