gbull25/qwen3-vl-russian-handwriting

VISIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:4BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jan 19, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

The gbull25/qwen3-vl-russian-handwriting model is a 4 billion parameter vision-language model, finetuned by gbull25 from unsloth/qwen3-vl-4b-instruct-unsloth-bnb-4bit. This model is specifically optimized for tasks involving Russian handwriting recognition and understanding, leveraging the Qwen3-VL architecture. It was trained using Unsloth and Huggingface's TRL library, enabling faster fine-tuning. Its primary strength lies in processing and interpreting visual inputs containing Russian handwritten text.

Loading preview...

Model Overview

The gbull25/qwen3-vl-russian-handwriting model is a specialized 4 billion parameter vision-language model (VLM) developed by gbull25. It is a fine-tuned version of the unsloth/qwen3-vl-4b-instruct-unsloth-bnb-4bit base model, built upon the Qwen3-VL architecture.

Key Capabilities

  • Vision-Language Understanding: Integrates visual input processing with language comprehension, making it suitable for tasks that require understanding both images and text.
  • Russian Handwriting Focus: Specifically fine-tuned to excel in recognizing and interpreting Russian handwritten content.
  • Efficient Training: The model was trained using Unsloth and Huggingface's TRL library, which facilitated a 2x faster fine-tuning process.

Good For

  • Russian Handwriting Recognition: Ideal for applications requiring the transcription or understanding of handwritten Russian text from images.
  • Document Analysis: Can be applied to tasks involving the extraction of information from scanned documents containing Russian handwriting.
  • Research and Development: Provides a specialized VLM for researchers working on Russian language processing and computer vision challenges.