kelsbeans/qwen3-1.7b-digestive-coach-n97

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 20, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

kelsbeans/qwen3-1.7b-digestive-coach-n97 is a 1.7 billion parameter Qwen3-based language model, fine-tuned by kelsbeans using Unsloth QLoRA. It specializes as a digestive wellness coach, demonstrating significantly improved parse rate and schema validity compared to its base model on digestive health-related scenarios. This model is designed as a research artifact for data-to-behavior fine-tuning projects, offering specialized conversational capabilities in a niche domain.

Loading preview...

Model Overview

kelsbeans/qwen3-1.7b-digestive-coach-n97 is a specialized language model built upon the Qwen3-1.7B architecture. It has been fine-tuned by kelsbeans using Unsloth QLoRA on a specific dataset of 97 rows related to digestive health, resulting in a merged fp16 model that loads directly with transformers.

Key Capabilities & Performance

This model is designed to function as a digestive wellness coach. Evaluation on a held-out set of 40 family-disjoint scenarios shows substantial improvements over the base Qwen3-1.7B model:

  • Parse Rate: Increased from 12.5% to 55.0%
  • Schema Validity: Improved from 2.5% to 47.5%
  • Spec Adherence: Rose from 0.0% to 7.5%
  • Field Accuracy: Slightly improved from 26.7% to 28.3%

Intended Use & Limitations

This model is primarily a research artifact for exploring data-to-behavior fine-tuning. It is explicitly stated as not medical advice and should not be used for diagnostic or treatment purposes. Its strength lies in its specialized fine-tuning for digestive health-related conversational scenarios, making it suitable for research or non-medical informational applications in this domain.

Finetunes
2 models