Unverified50% confidenceFactExact time
TensorRT 是 NVIDIA 提供的高性能推理优化引擎，能将 PyTorch/ONNX 模型编译为针对 GPU 优化的执行引擎，通常可将推理速度提升 2-5 倍
1
Sources
50%
Confidence
Long-term
Relevance
7/15/2026
First Seen
Sources
Related Claims
VerifiedNVIDIA的TensorRT-LLM针对其硬件架构深度优化推理速度78% similarUnverifiedNVIDIA TensorRT加速方案可将Stable Diffusion的推理速度提升2-4倍76% similarUnverifiedNVIDIA NIM内置了TensorRT-LLM等推理加速引擎75% similarUnverifiedNVIDIA Model Optimizer的前身为TensorRT Model Optimizer74% similarUnverifiedTensorRT-LLM由NVIDIA官方推出，通过算子融合和精度混合计算针对自家GPU进行性能调优，但模型支持范围相对有限74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/511221API
curl https://kongchang.com/api/v1/knowledge/claims/511221MCP
get_claim(id=511221)