Inference Speed Benchmark 推理速度测评 [COME AND SHARE YOUR SPEED!]

[JunityZhan]

I hope that everyone enter this issue can share your system, CPU, GPU, inference speed in GPT stage, and the version you use(better to compare v2 as a standard). so that we could see how different metrics affect the inference speed of GSV. 我希望大家可以分享一下你们的系统，cpu，gpu，在GPT的时候的推理速度，还有你用的gsv的版本。 这样我们可以看看到底哪些指标会影响GSV的推理速度 我先分享几个 I will share some of my experiments.

1. ubuntu 22, nvidia driver version 550, cuda 12.1, 4090, epyc 9654, gsv2: 260 tokens/s
2. WSL2, nvidia driver version: newest on Windows, cuda 11.8, 4090d, i7 14700kf, gsv2: 210 tokens/s
3. ubuntu 22, 10th intel cpu, cuda 11.8, l4 or A100 (almost same speed), gsv2: 95 tokens/s
4. ( I see in other issue)4090 + AMD Ryzen 9 7900X GSV1 (you can consider it as fast as gsv2) 400 it/s
5. (I see in other issue) 4090 + i9 13900 GSV1 torch script(you can consider it as fast as gsv2) 585 it/s