Downloads · 30 days
24
7% of all-time downloads
Aimu0999/mDeBERTa-v3-base-xnli-multilingual-nli-finetuned
mDeBERTa-v3-base-xnli-multilingual-nli-finetuned is a zero-shot classification model from Aimu0999. Use it when you need labels you did not train the model on. It is set up for transformers. The card lists the license as mit.
Developer: Aimu0999 Base Architecture: mDeBERTa-v3-base (Microsoft) Task: Natural Language Inference (NLI) & Zero-Shot Text Classification Language: Multilingual (Supports 100+ languages including English, Arabic, Fre…
Downloads · 30 days
24
7% of all-time downloads
All-time downloads
339
Public
Parameters
279M
1.1 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors1.1 GB · 98%
From the Hugging Face model README
This is a fine-tuned version of the powerful mDeBERTa-v3 model. It is specialized in Cross-Lingual NLI.
This means it can look at a piece of text (Premise) and a statement (Hypothesis) and determine if the statement is True (Entailment), False (Contradiction), or Unrelated (Neutral).
Because it is trained on NLI, it is extremely powerful for Zero-Shot Classification, allowing you to classify text into categories you haven't trained it on.
Classify text into any custom labels without training.
from transformers import pipeline
classifier = pipeline("zero-shot-classification", model="Aimu0999/mDeBERTa-v3-base-xnli-multilingual-nli-finetuned")
text = "The galaxy is full of stars and mystery."
labels = ["science", "politics", "cooking"]
result = classifier(text, labels)
print(result)