[TensorRT] ERROR: No non-int8 implementation of layer [CONVOLUTION #1]

NVIDIA-AI-IOT / torch2trt

An easy to use PyTorch to TensorRT converter

MIT License

4.62k stars 677 forks source link

[TensorRT] ERROR: No non-int8 implementation of layer [CONVOLUTION #1] #621

Open KeyKy opened 3 years ago

KeyKy commented 3 years ago

I get this error in Tesla T4, using the following code:

[TensorRT] ERROR: No non-int8 implementation of layer [CONVOLUTION #1]

CUDA_VISIBLE_DEVICES=0 python infer.py --m resnet18 --load_ckpt /data/yangkang/q_tensorrt/exps/pytorch_exp_1/ckpt_24 --netqat --INT8QAT

However, I do not get this error in Tesla P40 and Tesla T4 with INT8PTC. Why?

SrivastavaKshitij commented 3 years ago

do you have a reproducible script ?