qihang720

paddleocr检测模型转换ncnn，转换成功，加载模型出现layer 65 not exists or registered

3

## error log | 日志或报错信息 | ログ layer 65 not exists or registered ## model | 模型 | モデル 1. original model https://drive.google.com/file/d/1xDCdtGXDeARORHfQa5uQqoITU3fH-M5A/view?usp=sharing ## how to reproduce | 复现步骤 |...

# 平台(如果交叉编译请再附上交叉编译目标平台): # Platform(Include target platform as well if cross-compiling): 在sigmastar平台上编译 # Github版本: # Github Version: 1.2.6版本之后都编译失败，1.2.5之前正常 # 编译日志: # Build Log: /3rdparty/MNN/source/backend/cpu/compute/WinogradOptFunction.cpp:799:1: internal compiler error: in decompose_automod_address, at rtlanal.c:6311...

when using crop_mirror_normalize func, Output layout "CHW" is slower than "HWC"

5

1. output layout is CHW, using perf_analyzer to profile. ``` @autoserialize @dali.pipeline_def(batch_size=3, num_threads=1, device_id=0) def pipe(): images = dali.fn.external_source(device="cpu", name="encoded") images = dali.fn.decoders.image(images, device="mixed", output_type=types.RGB) images = dali.fn.resize(images, resize_x=299, resize_y=299)...

[FEATURE] Do you mind add an op supportted crop, resize and normalization, it's very useful in infer situation. But now I only find an op. for randomCropResize,

4

In my situation, I obtain rectangles through YOLO, then crop the corresponding rectangles from the original image, finally do resize and normalization, and send the results to ResNet. So I...