mirror of
https://github.com/zenlm/zen-coder.git
synced 2026-11-10 21:30:01 +00:00
- Rename Qwen2.5-Coder-* examples to zen-coder-* (consistent Zen branding) - Add LLM.md knowledge base (symlinked as CLAUDE.md/AGENTS.md/etc.) - Update .gitignore: exclude model weights and AI symlink files - Add training configs, inference scripts, test suite - Add MLX README, model card, Makefile for common operations
71 lines
2.4 KiB
Python
71 lines
2.4 KiB
Python
#!/usr/bin/env python3
|
|
"""
|
|
Prepare training data in mlx_lm format for LoRA fine-tuning.
|
|
Format: {"text": "conversation"} or {"messages": [...]}
|
|
"""
|
|
|
|
import json
|
|
from pathlib import Path
|
|
|
|
DATA_DIR = Path.home() / "work/zen/zen-agentic-dataset"
|
|
OUTPUT_DIR = Path(__file__).parent / "data"
|
|
OUTPUT_DIR.mkdir(exist_ok=True)
|
|
|
|
SYSTEM_PROMPT = """You are Zen Coder, an expert AI programming assistant."""
|
|
|
|
def convert_debug_session(content: str) -> dict:
|
|
"""Convert a debug session to chat format"""
|
|
messages = [
|
|
{"role": "system", "content": SYSTEM_PROMPT},
|
|
{"role": "user", "content": "Here is a coding session log. Learn from it:"},
|
|
{"role": "assistant", "content": content[:8000]} # Truncate long sessions
|
|
]
|
|
return {"messages": messages}
|
|
|
|
def main():
|
|
print("Preparing training data...")
|
|
|
|
train_samples = []
|
|
valid_samples = []
|
|
|
|
# Load from debug sessions
|
|
debug_file = DATA_DIR / "data/claude-debug/debug_sessions_latest.jsonl"
|
|
if debug_file.exists():
|
|
print(f"Processing {debug_file}...")
|
|
with open(debug_file, 'r') as f:
|
|
for i, line in enumerate(f):
|
|
try:
|
|
obj = json.loads(line.strip())
|
|
content = obj.get("content", "")
|
|
if content and len(content) > 500:
|
|
sample = convert_debug_session(content)
|
|
if i % 10 == 0:
|
|
valid_samples.append(sample)
|
|
else:
|
|
train_samples.append(sample)
|
|
|
|
if len(train_samples) >= 1000: # Limit for initial training
|
|
break
|
|
except:
|
|
continue
|
|
|
|
# Write output files
|
|
train_file = OUTPUT_DIR / "train.jsonl"
|
|
valid_file = OUTPUT_DIR / "valid.jsonl"
|
|
|
|
with open(train_file, 'w') as f:
|
|
for sample in train_samples:
|
|
f.write(json.dumps(sample) + "\n")
|
|
|
|
with open(valid_file, 'w') as f:
|
|
for sample in valid_samples:
|
|
f.write(json.dumps(sample) + "\n")
|
|
|
|
print(f"Created {len(train_samples)} training samples: {train_file}")
|
|
print(f"Created {len(valid_samples)} validation samples: {valid_file}")
|
|
print(f"\nTo train, run:")
|
|
print(f"python -m mlx_lm.lora --model Qwen/Qwen3-4B-Instruct-2507 --train --data {OUTPUT_DIR}")
|
|
|
|
if __name__ == "__main__":
|
|
main()
|