TensorRT · Issues· 639 open

trtllm-serve accepts `response_format` and never enforces it

Wrong FP16 results when a fused convolution, activation and residual add reads a strided concat view

TensorRT silently produces wrong results for batch > 1 on SegFormer (b0/b1/b2 verified, fp32, static and dynamic engines, TRT 10.11 & 11.1, sm89 + sm120)

[CMake] Windows package config references missing bin/tensorrt_player.exe (tensorrt_capture_player)

wrong results of TensorRT 10.0 when running on GPU Tesla T4

onnx_graphsurgeon.GraphPattern()：The order in which the nodes are added affects the success of the match.

`polygraphy run` failed on Polygraphy Version: 0.49.20. when will update `FormattedArray` to support reformated model input/output tensor, such as fp16:hwc8

AOT plugins leak CUDA modules on dynamic-shape changes in TensorRT 11.1/11.2.1

Question: How much large-scale scaling behavior can be predicted before benchmarking?

FP16 model compilation failure with TensorRT 11.2.1.2 on 4090 RTX

Polygraphy only supports a single input sample (iteration) during debug reduce

Incorrect CumSum outputs since TensorRT 10.8+

Incorrect output failure of TensorRT 11.1 and 11.2 when running dynamic If → Transpose → Resize on GPU GTX 1660 SUPER and A10G

Serialized engine size failure of TensorRT 11.2 when building an unfused pointwise op on a large activation on GPU H100

TensorRT 10.14+ multi-GPU build regression with Cask-backed ops