Modelfile

# ══════════════════════════════════════════════════════════════════════════════
# Code Autopsy — Ollama Modelfile
# ══════════════════════════════════════════════════════════════════════════════
# For local inference via Ollama after converting to GGUF.
#
# Steps:
#   1. Merge adapter:      python publish.py --skip-merge=false
#   2. Convert to GGUF:    bash convert_to_gguf.sh
#   3. Create Ollama model: ollama create code-autopsy -f Modelfile
#   4. Run:                ollama run code-autopsy
# ══════════════════════════════════════════════════════════════════════════════

FROM ./merged_model/code-autopsy-7b.Q4_K_M.gguf

SYSTEM """You are a code review expert. Analyze the provided code, identify any bugs or issues, explain the root cause, and provide a corrected version.

Always structure your response exactly as:

## Bug Identified
<description of the bug>

## Root Cause
<explanation of why this bug occurs>

## Fixed Code
```<language>
<corrected code>
```"""

# Inference parameters
PARAMETER temperature    0.1
PARAMETER top_p          0.9
PARAMETER top_k          40
PARAMETER num_ctx        2048
PARAMETER num_predict    600
PARAMETER repeat_penalty 1.1

# Stop tokens for Qwen2.5
PARAMETER stop "<|endoftext|>"
PARAMETER stop "<|im_end|>"
PARAMETER stop "<|im_start|>"