[Left TODO] Support PaddleSeg deployment #39
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Collaborator
felixhjh
commented
Jul 25, 2022
•
felixhjh
commented
- 支持 PaddleSeg 六个模型的部署（Deeplabv3-ResNet50, hrnet-w18, pp-humanseg, pp-humanseg-server, PP-LiteSeg, unet）
- 支持三种模型导出形式（logits matrix, logits matrix with argmax, logits matrix with argmax and with softmax）
- 支持visualization
Fix bug while the inference result is empty with YOLOv5 (PaddlePaddle#29)
…into PaddlePaddle-develop
felixhjh
changed the title
jiangjiajun
requested changes
Aug 3, 2022
| std::map<std::string, std::array<int, 2>>* im_info) { | ||
| FDASSERT(infer_result.dtype == FDDataType::INT64 || | ||
| infer_result.dtype == FDDataType::FP32, | ||
| "Require the data type of output is int64 or fp32, but now it's " + | ||
| Str(const_cast<fastdeploy::FDDataType&>(infer_result.dtype)) + |
There was a problem hiding this comment.
这里是不是可以直接写成Str(infer_result.dtype)
| } | ||
| } | ||
| } | ||
| } |
There was a problem hiding this comment.
这个hwc2chw是不是可以直接改成opencv的操作
| for (int i = 0; i < height; i++) { | ||
| for (int j = 0; j < width; j++) { | ||
| temp_mat.at<float_t>(i, j) = | ||
| static_cast<float_t>(infer_result_buffer[index++]); |
There was a problem hiding this comment.
processor的这个函数建议按照功能模块拆一下，提升可读性
jiangjiajun
requested changes
Aug 3, 2022
| int64_t chw = channel * height * width; | ||
| int64_t* infer_result_buffer = static_cast<int64_t*>(infer_result.Data()); | ||
| std::vector<float_t> float_result_buffer(chw); | ||
| temp_mat = cv::Mat(height, width, CV_32FC(channel)); |
There was a problem hiding this comment.
int64出来的结果，转为了float的来进行后面的resize处理，这个部署的结果是否跟PaddleSeg对齐了？
There was a problem hiding this comment.
这么做的原因有二：
- int64的结果出来的是类别的mask矩阵，对于pp-human系列也就是0,1矩阵（其他也就是0-num_classes的矩阵）。opencv没有64位的矩阵变量，因此必须要转。
- 转为float是因为，resize的插值方式双线性策略（inter=1，也是前处理所使用的策略）只支持float类型opencv mat（不支持int、uint8精度缺失）。为保持前后处理一致所以统一转为float。
- PaddleSeg给到的三张示例图片，正常人像、竖屏人像、横屏人像fastdeploy结果与PaddleSeg是一致的，后续evaluation模块开发完成（正在去除Paddle依赖）会跑全量
There was a problem hiding this comment.
第3点， 先确认输出的值(numpy) 与 PaddleSeg是一致的，并做到与Seg结果一致
jiangjiajun
approved these changes
Aug 5, 2022
jiangjiajun
changed the title
|
# for free
to join this conversation on GitHub.
Already have an account?
# to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.