QuantLLM v2.2 Examples

These examples are runnable -- each one calls real package APIs and prints live results. No hardcoded data or fakes.

Examples

#	File	What it does
01	01_quickstart.py	Load model, generate, chat, stream, export
02	02_gguf_export.py	Export at multiple GGUF quant levels
03	03_finetuning.py	LoRA fine-tuning on sample data
04	04_hub_push.py	Push to HF Hub (requires HF_TOKEN)
05	05_hqq_quantization.py	HQQ 2-8 bit quantization on layers/models
06	06_inference_server.py	Start OpenAI-compatible server (blocking)
07	07_benchmark.py	Live benchmark with BenchmarkRunner + comparer
08	08_full_pipeline.py	Load -> Generate -> Chat -> Export -> Push -> Serve

Run

# Basic model loading and generation
python examples/01_quickstart.py

# HQQ quantisation (no external model needed)
python examples/05_hqq_quantization.py

# Full pipeline
python examples/08_full_pipeline.py

Requirements

pip install quantllm torch transformers
pip install quantllm[server]   # for example 06
pip install quantllm[full]     # for all features