seongju/kor-3i4k-bert-base-cased

Model information

• language : Korean
• fine tuning data : kor_3i4k
• License : CC-BY-SA 4.0
• Base model : bert-base-multilingual-cased
• input : sentence
• output : intent

---

Train information

• train_runtime: 2376.638
• train_steps_per_second: 2.175
• train_loss: 0.356829648599977
• epoch: 3.0

---

How to use

from transformers import AutoTokenizer, AutoModelForSequenceClassification
  
tokenizer = AutoTokenizer.from_pretrained (
"seongju/kor-3i4k-bert-base-cased"
)

model = AutoModelForSequenceClassification.from_pretrained (
"seongju/kor-3i4k-bert-base-cased"
)
inputs = tokenizer(
"너는 지금 무엇을 하고 있니?",
 padding=True, truncation=True, max_length=128, return_tensors="pt"
 )
outputs = model(**inputs)
probs = outputs[0].softmax(1)
output = probs.argmax().item()