Problem categories: technical_issue, billing, account_management, content_request, feature_request, content_quality
Urgency levels: urgent, normal, low_priority
Customer types: free_user, premium_user, enterprise
Emotional states: frustrated, aggressive
Status flags: refund_request

Training Details

Training Data

The model was trained on the EuroChef+ customer support dataset containing synthetic multilingual customer support messages.

Usage

python

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("{hub_model_id}")
model = AutoModelForSequenceClassification.from_pretrained("{hub_model_id}")

# Prepare input
text = "I need help with my billing issue urgently!"
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=256)

# Get predictions
with torch.no_grad():
    outputs = model(**inputs)
    probs = torch.sigmoid(outputs.logits)[0]

# Get predicted labels (threshold = 0.5)
predicted_labels = []
for idx, prob in enumerate(probs):
    if prob > 0.5:
        label = model.config.id2label[idx]
        predicted_labels.append((label, prob.item()))

print(predicted_labels)

Limitations and Bias

The model was trained on synthetic data and may not generalize perfectly to real-world customer support scenarios
Performance may vary across different languages
The model reflects patterns in the training data which may contain biases

Citation

If you use this model, please cite:

bibtex

@misc{{mdeberta-eurochef-2026,
  author = {{BenTouss}},
  title = {{mDeBERTa-v3-base Fine-tuned on EuroChef+ Customer Support}},
  year = {{2026}},
  publisher = {{Hugging Face}},
  howpublished = {{\\url{{https://huggingface.co/{hub_model_id}}}}}
}}

Model Card Authors

BenTouss

BenTouss/mdeberta-eurochef

mDeBERTa-v3-base Fine-tuned on EuroChef+ Customer Support

Model Description

Training Details

Training Data

Usage

Limitations and Bias

Citation

Model Card Authors

BenTouss/
mdeberta-eurochef