mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2025-11-01 04:12:58 +08:00
[RKNN2] Fix bugs (#851)
* 修复picodet格式 * * 修正错误文档 * 修复rknpu2 backend后端的部分错误 * 更新pphumanseg example格式 * * 更新pphumanseg example格式 * * 更新picodet example格式 * * 更新scrfd example格式 * * 更新ppseg rknpu2 python example中的错误 * * 修复代码格式问题 * * 修复代码格式问题 * * 修复代码格式问题 * * 修复代码格式问题 * * 修复代码格式问题 * * 修复代码格式问题 Co-authored-by: Jason <jiangjiajun@baidu.com>
This commit is contained in:
@@ -345,6 +345,9 @@ bool RKNPU2Backend::Infer(std::vector<FDTensor>& inputs,
|
||||
FDERROR << "rknn_create_mem output_mems_ error." << std::endl;
|
||||
return false;
|
||||
}
|
||||
if(output_attrs_[i].type == RKNN_TENSOR_FLOAT16){
|
||||
output_attrs_[i].type = RKNN_TENSOR_FLOAT32;
|
||||
}
|
||||
// default output type is depend on model, this requires float32 to compute top5
|
||||
ret = rknn_set_io_mem(ctx, output_mems_[i], &output_attrs_[i]);
|
||||
// set output memory and attribute
|
||||
|
||||
Reference in New Issue
Block a user