I cannot convert YOLOX-X model to tensorRT format

Hi @FateScript,

I convert the yolox-s and yolox-m models to tensorRT format, but it gives an error in the yolox-x model. Can you help?

Yolox-X(workspace:32):

-------------- The current device memory allocations dump as below --------------
[0x302000000]:161480704 :HybridGlobWriter in reserveMemory: at optimizer/common/globWriter.cpp: 377 idx: 19 time: 0.000263675
[0]:4294967296 :HybridGlobWriter in reserveMemory: at optimizer/common/globWriter.cpp: 399 idx: 671 time: 0.0147272
[0x7f116f200000]:7771200 :DeviceActivationSize in reserveNetworkTensorMemory: at optimizer/common/tactic/optimizer.cpp: 4664 idx: 8 time: 0.000133113
[03/13/2022-12:50:22] [TRT] [W] Requested amount of GPU memory (4294967296 bytes) could not be allocated. There may not be enough free memory for allocation to succeed.
[03/13/2022-12:50:22] [TRT] [W] Skipping tactic 3 due to insuficient memory on requested size of 4294967296 detected for tactic 4.
Try decreasing the workspace size with IBuilderConfig::setMemoryPoolLimit().

Yolox-X(workspace:16):

[TRT] [I] Local timing cache in use. Profiling results in this builder pass will not be stored.
[TRT] [I] Some tactics do not have sufficient workspace memory to run. Increasing workspace size will enable more tactics, please check verbose output for requested sizes.

Start by reviewing the YOLOX TensorRT conversion entry point and comparing the yolox-s, yolox-m, and yolox-x conversion settings, especially the reported workspace values. Reproduce the conversion with the logged environment and inspect the TensorRT memory and tactic warnings. Done means the yolox-x model converts successfully or the required memory limitation is clearly documented.