tags-lite-free-0-6b-int8

Open Ticket AI – Lite Free (0.6B, INT8)

Model Summary

Open Ticket AI – Lite Free is a free, CPU-friendly ticket tagging model designed for helpdesk and IT support systems.
It automatically assigns up to 100 structured tags to incoming tickets, enabling improved routing, reporting, and automation without manual tagging.

The model is optimized for on-premise deployments, low-cost servers, and privacy-sensitive environments.
No external cloud services are required.

---

Model Details

• Developed by: Softoft (Tobias Bück)
• Model type: Text Classification / Multi-Label Tagging
• Base model: Qwen-based architecture
• Model size: ~0.6B parameters
• Quantization: INT8 (CPU-optimized)
• Languages: English, German (multilingual capable depending on input)
• License: Apache 2.0
• Intended use: Automatic ticket tagging for support and ITSM systems

---

Intended Use

Direct Use

• Automatic tagging of support tickets
• Routing and prioritization workflows
• Reporting and analytics enrichment
• Evaluation and testing of AI-based ticket classification

Typical integrations include:

• Zammad
• OTOBO
• Znuny / OTRS forks
• Other helpdesk systems via API or webhook

Out-of-Scope Use

• Generating free-form text
• Conversational chatbots
• Legal, medical, or safety-critical decision making
• High-stakes automation without human oversight

---

Performance Characteristics

• Designed for CPU inference
• Runs on low-cost servers (e.g. 2–8 vCPUs)
• Suitable for small to medium ticket volumes
• Typical accuracy in real-world setups: ~0.8

This model prioritizes stability, speed, and accessibility over maximum accuracy.

---

Limitations

• Accuracy is lower than larger or commercial variants
• Limited to ~100 predefined tags
• Performance depends on ticket text quality
• Not trained for highly domain-specific jargon without further fine-tuning

For higher accuracy, more tags, or enterprise features, consider Lite-Pro or Full variants.

---

🔗 Live Demo

Try the model directly in the browser:

👉 https://huggingface.co/spaces/open-ticket-ai/tags-lite-free

How to Use

from transformers import AutoTokenizer, AutoModelForSequenceClassification

model_id = "softoft/otai-tags-lite-free-0.6b-int8"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForSequenceClassification.from_pretrained(model_id)

inputs = tokenizer(
    "Email service is down since this morning, users cannot log in.",
    return_tensors="pt",
    truncation=True
)

outputs = model(**inputs)

The output logits can be mapped to the predefined tag set used by Open Ticket AI.

---

Training Data

The model was fine-tuned on a large, curated dataset of synthetic and semi-synthetic support tickets, designed to reflect realistic helpdesk scenarios across multiple domains.

No personal or customer-identifiable data was used.

---

Evaluation

• Internal evaluation on held-out ticket sets
• Metric focus: multi-label classification accuracy and consistency
• Real-world accuracy depends on domain and input quality

---

Environmental Impact

• Training hardware: GPU (one-time fine-tuning)
• Inference target: CPU-only
• Designed to minimize energy consumption during inference

---

Technical Details

• Architecture: Transformer-based sequence classification
• Objective: Multi-label classification
• Framework: PyTorch / Transformers
• Deployment: On-premise, VM, containerized environments

---

Citation

If you use this model in academic or commercial work, please cite:

Open Ticket AI – Lite Free, Softoft (2025)

---

Contact

For questions, integrations, or commercial licensing options:

• Website: https://softoft.de
• GitHub / Hugging Face: https://huggingface.co/softoft