ermiaazarkhalili/Qwen3.8-4B-Function-Calling-xLAM-Unsloth

VISIONConcurrent Unit Cost:1Model Size:4.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 20, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The ermiaazarkhalili/Qwen3.8-4B-Function-Calling-xLAM-Unsloth is a 4.7 billion parameter Qwen3.5-based causal language model, fine-tuned by ermiaazarkhalili. This model is specifically optimized for function calling tasks, having been supervised fine-tuned using the Salesforce/xlam-function-calling-60k dataset. It leverages LoRA with Unsloth and TRL for efficient training, making it suitable for applications requiring robust tool use and function invocation capabilities.

Loading preview...

Overview

This model, ermiaazarkhalili/Qwen3.8-4B-Function-Calling-xLAM-Unsloth, is a LoRA fine-tune of the empero-ai/Qwen3.8-4B base model, developed by ermiaazarkhalili. It is built on the Qwen3.5 architecture and features 4.7 billion parameters. The primary focus of this fine-tuning was to enhance its function-calling capabilities, achieved through supervised fine-tuning on the Salesforce/xlam-function-calling-60k dataset.

Key Capabilities

  • Function Calling: Specifically trained to understand and generate function calls based on user prompts, utilizing the xlam-function-calling-60k dataset.
  • Efficient Training: Fine-tuned using LoRA (Low-Rank Adaptation) via the Unsloth and TRL libraries, allowing for efficient adaptation of the base model.
  • Qwen3.5 Architecture: Inherits the robust capabilities of the Qwen3.5 model family.

Limitations

  • No Benchmark Evaluation: The model has not undergone downstream benchmark evaluation; only training-loss observations are available.
  • Inherited Biases: Carries the biases, knowledge cutoff, and potential failure modes of its empero-ai/Qwen3.8-4B base model.
  • Specific Instruction Following: Its behavior outside the distribution of the Salesforce/xlam-function-calling-60k dataset is untested.
  • Merged Adapters: LoRA adapters are merged into the base weights, meaning the fine-tune cannot be detached.

Good for

  • Developers looking for a compact (4.7B parameter) model specialized in function calling.
  • Applications requiring integration with external tools or APIs through function invocation.
  • Experimentation with LoRA-based fine-tuning on Qwen models for specific instruction-following tasks.

Finetunes
2 models