🎯An accuracy-first, highly efficient quantization toolkit for LLMs, designed to minimize quality degradation across Weight-Only Quantization, MXFP4, NVFP4, GGUF, and adaptive schemes.

记录量化LLM中的总结。

LLM fine-tuning with LoRA + NVFP4/MXFP8 on NVIDIA DGX Spark (Blackwell GB10)

Code for the paper "ARCQuant: Boosting NVFP4 Quantization with Augmented Residual Channels for LLMs"

Production LLM deployment specs for NVIDIA Blackwell GPUs (RTX Pro 6000, DGX Spark). Includes vLLM configurations, benchmarks, load balancer, and throughput calculators for NVFP4/FP8/MoE models.

🔧 Fine-tune large language models efficiently on NVIDIA DGX Spark with LoRA adapters and optimized quantization for high performance.