Tags
performance73ecosystem33hardware28model-support23speculative-decoding21multimodal19large-scale-serving16disaggregation13quantization11moe10reinforcement-learning8kv_cache7community6kernels5developer5models4vllm-omni4nixl3speculators3post-training3attention3inference3elastic-ep2expert-parallelism2amd2deepseek2deployment2distributed2agentic2parallelism2rlhf2dflash2prefix caching2agentic-routing2flint1cuda-graphs1NVIDIA1rocm1reflection1watermarking1sampling1apple-silicon1qwen3.81glm1kv-cache1fastvideo1fasth31cosmos31qwen3.51speculative_decoding1peagle1dspark1mixture-of-models1semantic-router1ci1evaluation1release1hpc-ops1minimax1day-0-support1long-context1model1learning1dgx-spark1nemotron1computex1llm-compressor1async-rl1production-serving1fault-tolerance1turboquant1benchmarking1kernel-fusion1fp81mamba1engineering1triton1frontend1