PP-OCRv3-rec移动端部署
#12855
Replies: 2 comments
-
您这还是性能降低,我的直接Segmentation fault不转换了,linux下build的opt和windows下的paddle_lite_opt,python脚本都是 |
Beta Was this translation helpful? Give feedback.
0 replies
-
遇到相同问题,请问你那边解决了吗 |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
使用官方提供的配置文件训了一个PP-OCRv3-rec模型,python infer都很正常。但是使用paddleLite转换nb格式部署移动端性能大幅度减低。(PP-OCRv2-rec训练的模型转换后的效果好一些)
config: configs/rec/PP-OCRv3/ch_PP-OCRv3_rec.yml
paddleLite==2.10
转换命令:paddle_lite_opt --model_file ../paddle/PaddleOCR-release-2.5/inference/rec_v2_lite_v5.2_infer/inference.pdmodel --param_file ../paddle/PaddleOCR-release-2.5/inference/rec_v2_lite_v5.2_infer/inference.pdiparams --optimize_out ../paddle/PaddleOCR-release-2.5/inference/rec_v2_v3.4_infer/rec_v2_opt --valid_targets=arm --optimize_out_type=naive_buffer
请问,PP-OCRv3-rec训练的模型因为algorithm: SVTR_LCNet 暂时不能移动端部署还是我做错了什么步骤?
Beta Was this translation helpful? Give feedback.
All reactions