模型推理问题 - Githubissues

PaddlePaddle / PaddleSeg

Easy-to-use image segmentation library with awesome pre-trained model zoo, supporting wide-range of practical tasks in Semantic Segmentation, Interactive Segmentation, Panoptic Segmentation, Image Matting, 3D Segmentation, etc.

https://arxiv.org/abs/2101.06175

Apache License 2.0

8.62k stars 1.68k forks source link

模型推理问题 #3696

Open djl00 opened 6 months ago

djl00 commented 6 months ago

问题确认 Search before asking

[X] 我已经搜索过问题，但是没有找到解答。I have searched the question and found no related answer.

请提出你的问题 Please ask your question

为什么我在2080ti上能够复现成功且成功使用Tensorrt进行推理。但是我把模型放到3090ti上能够训练，但是推理的时候报错，显示数据不匹配，有什么解决方法吗

shiyutang commented 6 months ago

你好，请展示你具体的报错信息。

djl00 commented 6 months ago

运行的时候报错：AssertionError: Not equal to tolerance rtol=0, atol=0.001

Mismatched elements: 1748394 / 5767168 (30.3%) Max absolute difference: 0.00674272 Max relative difference: 58.494873 x: array([[[[ 1.147214, 1.147214, 1.29746 , ..., 0.759871, 0.605249, 0.605249], [ 1.147214, 1.147214, 1.29746 , ..., 0.759871, 0.605249,... y: array([[[[ 1.147725, 1.147725, 1.297974, ..., 0.76047 , 0.60574 , 0.60574 ], [ 1.147725, 1.147725, 1.297974, ..., 0.76047 , 0.60574 ,...

Process finished with exit code 1

djl00 commented 6 months ago

运行代码为：python deploy/python/infer_onnx_trt.py --config ../../configs/pp_liteseg/pp_liteseg_stdc1_camvid_960x720_10k.yml --width 1024 --height 512

mandylyin commented 5 months ago

@djl00 请问你这个问题解决了吗？我也遇到同样的问题，不知道怎么解决