manishkumar2101114/qwen3.5-9b-tool-merged

VISIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Oct 5, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The manishkumar2101114/qwen3.5-9b-tool-merged model is a 9 billion parameter language model based on the Qwen/Qwen3.5-9B architecture. It incorporates a LoRA adapter (r=32, alpha=64) fine-tuned for tool-use capabilities over 3 epochs, resulting in a merged full BF16 model. This model is optimized for applications requiring robust tool interaction and function calling.

Loading preview...

Qwen3.5-9B Tool Merged Model

This model is a specialized variant of the Qwen/Qwen3.5-9B base architecture, featuring 9 billion parameters. It has been enhanced with a LoRA adapter, specifically the qwen3.5-9b-tool-adapter, which was trained over 3 epochs with 666 steps, achieving a training loss of 0.2248 and an evaluation loss of 0.1729. The adapter has been merged into the base model to create a full BF16 model.

Key Capabilities

  • Tool-Use Optimization: The primary enhancement focuses on improving the model's ability to understand and utilize external tools or functions.
  • Efficient Fine-tuning: Leverages LoRA (Low-Rank Adaptation) for efficient fine-tuning, making it suitable for specific task adaptation without retraining the entire base model.
  • BF16 Precision: The merged model operates in BF16 (bfloat16) precision, balancing performance and memory efficiency.

Good For

  • Function Calling: Ideal for applications where the LLM needs to interact with external APIs or tools by generating structured function calls.
  • Agentic Workflows: Suitable for building AI agents that can plan and execute actions by calling various tools.
  • Specialized Task Automation: Can be deployed in scenarios requiring precise tool interaction for automation and complex problem-solving.

Finetunes
1 models