Tags
performance84ecosystem34hardware32model-support25multimodal23speculative-decoding23large-scale-serving18disaggregation17quantization11moe11reinforcement-learning10kv_cache7vllm-omni6kernels6community6deepseek5models5developer5inference4rlhf3day-0-support3parallelism3nixl3speculators3post-training3attention3glm2kv-cache2minimax2elastic-ep2expert-parallelism2amd2deployment2distributed2agentic2dflash2prefix caching2agentic-routing2image-generation1kimi1kinference1inference-control-plane1scalable-platforms1sparse-attention1disaggregated-serving1flint1cuda-graphs1NVIDIA1rocm1reflection1watermarking1sampling1apple-silicon1qwen3.81fastvideo1fasth31cosmos31qwen3.51speculative_decoding1peagle1dspark1mixture-of-models1semantic-router1ci1evaluation1release1hpc-ops1long-context1model1learning1dgx-spark1nemotron1computex1llm-compressor1async-rl1production-serving1fault-tolerance1turboquant1benchmarking1kernel-fusion1fp81mamba1engineering1triton1frontend1