jasirjru/DomainTune-Qwen2.5-1.5B-Triage
DomainTune-Qwen2.5-1.5B-Triage is a 1.5 billion parameter language model, fine-tuned by jasirjru from Qwen/Qwen2.5-1.5B, with a 32768 token context length. This model is specifically optimized for autonomous ticket triage, designed to process unstructured software bug reports and support tickets. It outputs strict, deterministic JSON triage decisions, making it ideal for automating the classification of technical issues.
Loading preview...
DomainTune-Qwen2.5-1.5B-Triage: Autonomous Ticket Triage
This model, developed by jasirjru, is an enterprise-grade fine-tuned version of Qwen/Qwen2.5-1.5B with 1.5 billion parameters and a 32768 token context length. Its core purpose is to automate the processing of unstructured software bug reports and support tickets, transforming them into structured, deterministic JSON outputs.
Key Capabilities
DomainTune is engineered to extract critical information from raw support tickets or GitHub issues, providing a comprehensive triage decision. It identifies and categorizes:
priority: Assigns a severity level (P1, P2, P3, P4).category: Classifies the issue type (bug, feature_request, documentation, performance, security, infra).affected_component: Pinpoints the specific system or module impacted (e.g., auth_service, payment_gateway).sentiment: Detects the user's emotional tone (neutral, frustrated, urgent, satisfied).resolution_required: Indicates whether the ticket necessitates a resolution.
Good for
- Automating the initial classification and routing of customer support tickets.
- Streamlining software development workflows by triaging bug reports.
- Ensuring consistent and deterministic categorization of technical issues.
- Reducing manual effort in support and engineering operations.