AbhishekG711/Qwen3-1.7B-Insight-Extractor

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Oct 4, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

AbhishekG711/Qwen3-1.7B-Insight-Extractor is a 1.7 billion parameter LoRA fine-tune of the Qwen3-1.7B causal decoder-only Transformer model. This specialized model is designed to convert raw customer support tickets into a single, structured JSON object with 9 predefined fields. It excels at automated ticket classification, routing, and downstream entity extraction workflows, providing deterministic output for operational efficiency.

Loading preview...

Model Overview

AbhishekG711/Qwen3-1.7B-Insight-Extractor is a specialized LoRA fine-tune of the Qwen3-1.7B base model, developed by AbhishekG711. This 1.7 billion parameter causal decoder-only Transformer is engineered for a singular task: converting raw customer support ticket text into a structured JSON object. It leverages a native context length of 32,768 tokens and was fine-tuned using Unsloth's LoRA adapters and TRL SFTTrainer on the AbhishekG711/processed_support_tickets dataset.

Key Capabilities

  • Structured Data Extraction: Transforms unstructured support ticket text into a single JSON object with 9 specific fields, including is_actionable, summary, sentiment, category, intent, aspect, urgency, reported_cause, and entities.
  • Deterministic Output: Designed to act as an "extraction engine," consistently producing JSON output without conversational elements or explanations.
  • Efficient Fine-tuning: Utilizes LoRA with specific hyperparameters (e.g., r=32, alpha=64) for efficient adaptation of the base Qwen3-1.7B model.
  • Automated Workflow Integration: Ideal for automating initial triage, routing, sentiment analysis, and triggering downstream alerts based on extracted insights.

Intended Use Cases

  • Automated Ticket Triage: Classify and route incoming support tickets to the correct departments or agents.
  • Sentiment and Urgency Dashboards: Populate dashboards with real-time sentiment and urgency metrics extracted from customer interactions.
  • Downstream Entity Extraction: Facilitate further processing by providing structured data for named entity recognition or other NLP tasks.
  • Operational Efficiency: Streamline customer support operations by automating the initial analysis of support requests.

Finetunes
1 models