deveg/toolcalling-merged-demo

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:2BQuant:BF16Ctx Length:32kPublished:Mar 26, 2026License:apache-2.0Architecture:Transformer Open Weights Warm

The deveg/toolcalling-merged-demo is a 2 billion parameter Qwen3-based causal language model, fine-tuned by deveg with Unsloth and Huggingface's TRL library. This model is optimized for efficient training, having been developed twice as fast as standard methods. It features a 32768 token context length, making it suitable for tasks requiring extensive contextual understanding and processing.

Loading preview...

Model Overview

The deveg/toolcalling-merged-demo is a 2 billion parameter language model based on the Qwen3 architecture. It was developed by deveg and fine-tuned using Unsloth and Huggingface's TRL library, enabling significantly faster training times.

Key Characteristics

  • Architecture: Qwen3-based, specifically fine-tuned from unsloth/Qwen3-1.7B-unsloth-bnb-4bit.
  • Parameter Count: 2 billion parameters.
  • Context Length: Supports a substantial context window of 32768 tokens.
  • Training Efficiency: Benefits from Unsloth's optimizations, resulting in a 2x speedup in training compared to conventional methods.

Potential Use Cases

This model is well-suited for applications where efficient training and a large context window are beneficial. Its Qwen3 foundation suggests capabilities in general language understanding and generation, while the specific fine-tuning implies potential for specialized tasks, though the README does not detail specific task optimizations beyond efficient training.