fix distribute fpn proposals, test=develop#16152
fix distribute fpn proposals, test=develop#16152jerrywgz merged 8 commits intoPaddlePaddle:developfrom
Conversation
Menevagol
left a comment
Menevagol
left a comment
There was a problem hiding this comment.
In the operation collect_fpn_proposals， modify the <= at compare_scores function to <
…addle into fix_distribute_fpn_op
… fix_distribute_fpn_op
There was a problem hiding this comment.
Why change to sizeof(uint64_t) * 8 ?
There was a problem hiding this comment.
It will case cudalasterror when kNumCUDAThreads=512
There was a problem hiding this comment.
It will cause CudaLastError when kNumCUDAThreads = 512;
There was a problem hiding this comment.
Do not copy code from other file, like RangeInitFunctor, CUDA_1D_KERNEL_LOOP, RoIArea, ...
There was a problem hiding this comment.
Done, while CUDA_1D_KERNEL_LOOP is widely used in our cuda kernels
There was a problem hiding this comment.
(T) -> static_cast<T>()
There was a problem hiding this comment.
Why add __syncthreads ?
jerrywgz
force-pushed
the
fix_distribute_fpn_op
branch
from
68d75d8 to
2b6c154
Compare
jerrywgz
force-pushed
the
fix_distribute_fpn_op
branch
from
23b3160 to
cf3de9f
Compare
| struct RangeInitFunctor { | ||
| int start_; | ||
| int delta_; | ||
| int* out_; |
There was a problem hiding this comment.
struct的成员变量无_结尾，参考google C++ code style.
There was a problem hiding this comment.
Will fix it in collect_fpn_op
3 participants
Uh oh!
There was an error while loading. Please reload this page.