Open ysj-xuanyuan opened 8 months ago
When I use do_inference_v2, I need the reshape operation to get the correct result, but I find that the reshape operation takes a lot of time, which defeats the purpose of using tensorRT to try to speed up inference in the first place
When I use do_inference_v2, I need the reshape operation to get the correct result, but I find that the reshape operation takes a lot of time, which defeats the purpose of using tensorRT to try to speed up inference in the first place