Make all tensorrt optimizations compatible with 3D convolution


With TensorRT 7.0 comes compatibility with 3D convolution. But 8 bit integer quantization still isn’t available for 3D convolution, as shown here, section “Layer and precision” :
However, it’s a huge part of performance gains.

In fact, we should be able to look at a 15 folds performance gain with TensorRT (based on what I obtain with 2D models on various hardware) whereas, with what is available right now, I could only obtain a *1.5 folds.

Are there plans to make the rest of TensorRT optimizations available for 3D convolution ?


Please keep watching below TensorRT links for any new announcements or release related details:


I see it is still not the case. Is there any plans to allow quantization and other optimizations for 3d conv in TensorRT ?


Any updates on the roadmap for optimizing 3D convolutions?

1 Like

Hi. Any updates on making 3d convlutions compatible with all optimizations ?

Still very interested. Has there been any progress on this ?

There has been progress. Since Tensorrt 7.2, Tensorcore can be used to speedup INT8 inference of 3d conv layers. This provides some speedup on select GPUs.

See 3d layers don't get quantized · Issue #1176 · NVIDIA/TensorRT · GitHub

But not all optimizations are available yet. There is still no speedup with INT8 on Pascal GPUs, and all gpus without tensor cores, whereas all GPUs see huge speedup on 2D models when comparing INT8 vs FP16.