from transformers import AutoTokenizer, AutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)
model = AutoModelForSequenceClassification.from_pretrained('techthiyanes/Bert_Bahasa_Sentiment')
inputs = tokenizer("saya tidak", return_tensors="pt")
labels = torch.tensor([1]).unsqueeze(0)
outputs = model(**inputs, labels=labels)
loss = outputs.loss
logits = outputs.logits
outputs hello