Popular repositories Loading

1. vllm-v100-2080ti-recipes vllm-v100-2080ti-recipes Public

   Working recipes and measured benchmarks for serving large LLMs on pre-Ampere NVIDIA GPUs — Tesla V100 (sm_70) and RTX 2080 Ti (sm_75). vLLM forks, llama.cpp tensor parallelism, AWQ/MoE gotchas.