Quick Start

PMetal exposes a Python extension module via PyO3. Install with maturin from crates/pmetal-py.

Installation

cd crates/pmetal-py
pip install maturin
maturin develop --release

Fine-Tuning

import pmetal

result = pmetal.finetune(
    "Qwen/Qwen3-0.6B",
    "train.jsonl",
    lora_r=16,
    learning_rate=2e-4,
    epochs=3,
)
print(f"Loss: {result['final_loss']}, Steps: {result['total_steps']}")

Inference

# Simple generation
text = pmetal.infer("Qwen/Qwen3-0.6B", "What is 2+2?")
print(text)

# With LoRA adapter
text = pmetal.infer(
    "Qwen/Qwen3-0.6B",
    "Explain quantum entanglement",
    lora="./output/lora_weights.safetensors",
)
print(text)

# With generation parameters
text = pmetal.infer(
    "Qwen/Qwen3-0.6B",
    "Tell me a story",
    temperature=0.8,
    max_tokens=512,
)

See Also

• Full Control — Custom training loops and model loading
• Rust SDK — Rust equivalent API