Spaces:

Lordemarco
/

sentiment-fastapi

Sleeping

LorenzoBioinfo

Test model and prepare dowload data

2cedc0c 12 months ago

583 Bytes

from transformers import AutoTokenizer, AutoModelForSequenceClassification
from scipy.special import softmax
import torch

MODEL = "cardiffnlp/twitter-roberta-base-sentiment-latest"
tokenizer = AutoTokenizer.from_pretrained(MODEL)
model = AutoModelForSequenceClassification.from_pretrained(MODEL)


text = "Terrible"


inputs = tokenizer(text, return_tensors="pt")

with torch.no_grad():
    logits = model(**inputs).logits
    probs = softmax(logits.numpy()[0])


labels = ["negative", "neutral", "positive"]
for label, prob in zip(labels, probs):
    print(f"{label}: {prob:.4f}")

	from transformers import AutoTokenizer, AutoModelForSequenceClassification
	from scipy.special import softmax
	import torch

	MODEL = "cardiffnlp/twitter-roberta-base-sentiment-latest"
	tokenizer = AutoTokenizer.from_pretrained(MODEL)
	model = AutoModelForSequenceClassification.from_pretrained(MODEL)


	text = "Terrible"


	inputs = tokenizer(text, return_tensors="pt")

	with torch.no_grad():
	logits = model(**inputs).logits
	probs = softmax(logits.numpy()[0])


	labels = ["negative", "neutral", "positive"]
	for label, prob in zip(labels, probs):
	print(f"{label}: {prob:.4f}")