-
Notifications
You must be signed in to change notification settings - Fork 3k
Fix channel num #8436
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: release/2.6
Are you sure you want to change the base?
Fix channel num #8436
Conversation
* update pphuman ppvechicle new models, test=document_fixb * update docs, test=document_fixb * update docs, test=document_fix * fix configs, test=document_fix
* fix distill codes and docs * fix slim distill docs * fix docs * add details of docs, test=document_fix
* add swin large dino, refine dino codes * fix dino export
* update pipeline docs; vehicle illegal;rtsp;jetson * update docs
* fix smalldet docs, test=document_fix * fix smalldet configs and readme, test=document_fix
* pose3d docs * traindata download
* add ppyoloe tiny 320 models, test=document_fix * add ppyoloe tiny 320 speed, test=document_fix
* ppvehicle add customized docs, test=document_fix * fix pipeline readme.md test=document_fix * fix pipeline readme.md link test=document_fix * add licensed to lane_to_mask.py * ppvehicle_violation.md add args list test=document_fix
* petr train ok train ok refix augsize affine size fix update msdeformable fix flip/affine fix clip add resize area add distortion debug mode fix pos_inds update edge joints update word mistake * delete extra codes;adapt transformer modify;update code format * reverse old transformer modify * integrate datasets
* add config and architecture for human36m * modify TinyPose3DHRNet to support human3.6M dataset * delete useless class
* fix run_benchmark for small model accurate speed * fix run_benchmark for other det models
* new adapted * test ok
* metro con reverse tinypose3d fix readme modelzoo * fix tinypose3d
* add raw codes * add pretrained weights * fix reader * fix * fix focalnet codes coments and format * fix focalnet codes coments and format * fix focalnet format
* [TIPC] fcos add static train * tinypose_128x96 add training static * [TIPC] tinypose add static training * disable maskrcnn static training
* fix becnhmark,delete run_process_type * fix * fix benchmarkk * fix mask_rcnn_r50_1x_coco benchmark * fix benchmark * fix benchmark * fix benchmark prepare.sh * fix benchmark * remove static benchmark model * disable * fix profiler * fix * fix * fix --------- Co-authored-by: zhengya01 <zhengya01@baidu.com> Co-authored-by: mmglove <gmm_email@baidu.com>
* rt-detr-ssod * add rt-detr-ssod * change name
* fix the error of defined more than once * fix the error of defined
* Accommodate UAPI * Fix bugs * Set defaults for use_fd_format * Restore visualize.py * Rename variable * Optimize save_dir * Fix mistakenly update * Add format check
* support ampo1 in benchmark; fp16 -> amp * benchmark: fp16 -> amp
* benchmark: fp16 -> amp * revert
adapt to paddle upgrade that concat dont support 0D tensor
* fix a code * fix the tensor to float
|
Thanks for your contribution! |
|
|

https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.6/ppdet/modeling/heads/pico_head.py
第59行 PicoFeat 类中bug
在第112 行 创建 cls_conv_pw 变量的设置中:入通道为in_c,
但是 cls_conv_pw 接收 上面 cls_conv_dw 的输出作为输入,而cls_conv_dw输出的通道数是feat_out 。这样通道数不一致了。
这里把 115行的 ch_in=in_c 改为 ch_in=feat_out