---
base_model: Colby/starcoder-7b-agent-0.2-merged
tags:
- gguf
- starcoder
- agentic-coding
- tool-use
- sft
---

# starcoder-7b-agent-0.6-gguf

GGUF conversion of [Colby/starcoder-7b-agent-0.6](https://huggingface.co/Colby/starcoder-7b-agent-0.6),
a LoRA fine-tune of [Colby/starcoder-7b-agent-0.2-merged](https://huggingface.co/Colby/starcoder-7b-agent-0.2-merged)
on several agentic coding / reasoning datasets.

Chat format: StarCoderChat (`<|system|>`, `<|user|>`, `<|assistant|>`, `<|end|>`)
with `<think>...</think>` blocks for reasoning traces and `<tool_call>`/`<tool_response>` for agentic turns.

## Quantizations

| File | Format | Size |
|------|--------|------|
| starcoder-7b-agent-0.6-f16.gguf | FP16 | ~14 GB |
| starcoder-7b-agent-0.6-q8_0.gguf | Q8_0 | ~7 GB |
| starcoder-7b-agent-0.6-q5_k_m.gguf | Q5_K_M | ~5 GB |
| starcoder-7b-agent-0.6-q4_k_m.gguf | Q4_K_M | ~4 GB |

## Ollama usage

```bash
hf download Colby/starcoder-7b-agent-0.6-gguf starcoder-7b-agent-0.6-q4_k_m.gguf
ollama create starcoder-agent:7b -f Modelfile   # FROM ./starcoder-7b-agent-0.6-q4_k_m.gguf
ollama run starcoder-agent:7b
```