seanpoyner/smolcode-coder-js-1.5b-tools
The seanpoyner/smolcode-coder-js-1.5b-tools model is a 1.5 billion parameter LoRA fine-tune of Qwen2.5-Coder-1.5B-Instruct, developed by seanpoyner. Optimized for agentic coding loops, this model specializes in emitting native function calls, addressing a limitation in base Qwen-Coder models. It is designed to drive agentic coding assistants by enabling small language models to effectively use tools, with a context length of 32768 tokens.
Loading preview...
Model Overview
seanpoyner/smolcode-coder-js-1.5b-tools is a 1.5 billion parameter LoRA fine-tune of the Qwen2.5-Coder-1.5B-Instruct base model. Developed by seanpoyner for the Hugging Face "Build Small" hackathon, its primary purpose is to enable small language models (SLMs) to effectively drive agentic coding loops by emitting native <tool_call> function calls.
Key Capabilities & Differentiators
- Native Tool Call Generation: Unlike base Qwen-Coder models that output tool calls as plain-text JSON, this fine-tune generates
<tool_call>formatted responses, which are directly parseable by runtimes like Ollama and llama.cpp. This is crucial for functional agentic tool-use. - Agentic Coding Optimization: Specifically built for
smolcode, an SLM-optimized agentic coding assistant, it focuses on practical application in automated coding environments. - Efficient Fine-tuning: Trained using bf16 LoRA with assistant-only loss, focusing on tool calls and final answers. The training data combined
NousResearch/hermes-function-calling-v1with syntheticsmolcodetool-use trajectories, ensuring byte-identical training and inference templates.
Use Cases
This model is ideal for developers building agentic coding assistants or applications that require small, efficient language models to interact with external tools via structured function calls. It addresses the challenge of integrating tool-use capabilities into resource-constrained environments, making it suitable for scenarios where larger models are impractical.