This website requires JavaScript.
Explore
Help
Register
Sign In
mirrors
/
mmdeploy
Watch
1
Star
0
Fork
0
You've already forked mmdeploy
mirror of
https://github.com/open-mmlab/mmdeploy.git
synced
2025-01-14 08:09:43 +08:00
Code
Issues
Projects
Releases
Wiki
Activity
mmdeploy
/
configs
/
_base_
/
backends
/
ncnn.py
2 lines
71 B
Python
Raw
Normal View
History
Unescape
Escape
Ncnn demo sdk (#164) * fix different dims and data sync and fp16 macro * fix pad is negative bug, but still need debugging info * split detection_output to dets and labels * fix clang++ compile bug * for fp16 compile macro of cmake * fix pad_val dict input * fix yolox missing normalization and get int tensor * about score_threshold * remove -lstdc++fs for android * move -pthread for android * fix link libraries for CLANG++; * fix clang-format * for mobileyolov3 conf_thre * fix lint * fix bug * support ncnn vulkan net * remove unused debugging info; * Change INFO to DEBUG * support vulkan precision mapping * fix pad * optimize getdetslabels * remove EVAL_MODE * ncnn_net support 4-dim data. * remove FP16 compile options * remove use_gpu_vulkan compile options * fix pad.cpp * fix yapf * fix clang-format * rm redundant lines * fix pad according to mmdet * add android build docs * fix lint * use cpp style string comparision * fix use after free bug * Add missing -DBUILD_SHARED_LIBS=OFF for en * Add missing -DBUILD_SHARED_LIBS=OFF for ZH-CN * reset img2tensor_impl * reset img2tensor_impl * add blank line * sync android.md docs * fix some dirs * fix docs * update docs * fix code
2022-03-25 23:28:16 +08:00
backend_config
=
dict
(
type
=
'
ncnn
'
,
precision
=
'
FP32
'
,
use_vulkan
=
False
)
Reference in New Issue
Copy Permalink