Warlord-K/kanha-incidentio-incidents-1.7b-full-v1

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 24, 2026Architecture:Transformer Featherless Exclusive Cold

Warlord-K/kanha-incidentio-incidents-1.7b-full-v1 is a 1.7 billion parameter language model based on the Qwen3 architecture, fine-tuned using a 'full' training method. Developed by Warlord-K, this model is specifically trained on a dataset derived from the incident.io documentation, making it specialized for website question answering. It is intended for research purposes to compare training methods and evaluate question answering capabilities on this specific dataset.

Loading preview...

Model Overview

Warlord-K/kanha-incidentio-incidents-1.7b-full-v1 is a 1.7 billion parameter model built upon the Qwen3-1.7B base architecture. It has undergone a 'full' training method, utilizing a dataset specifically sourced from the docs.incident.io website. The model was trained with a maximum sequence length of 2048 tokens over 4 epochs, with a learning rate of 1e-05.

Key Characteristics

  • Base Model: Qwen/Qwen3-1.7B
  • Training Method: Full fine-tuning
  • Dataset Source: docs.incident.io documentation
  • Context Length: 2048 tokens during training (model supports 32768 tokens)
  • Evaluation Metrics: Achieved 1.0 for dates_recall, list_recall, numbers_recall, and urls_recall, with a refusal_rate of 0.0.

Intended Use Cases

This checkpoint is primarily intended for:

  • Research: Comparing different training methods on the same Kanha website-derived dataset.
  • Controlled Evaluation: Assessing website question answering capabilities.

Limitations

Users should be aware that the model may produce incorrect, incomplete, or outdated answers. It also has the potential to memorize training content. It is crucial to review all outputs, test for potential failure cases, and qualify the exact runtime environment before any deployment in user-facing applications.