Andy

性能评价等的时候怎么输出Recall F1-score？

1

性能评价是用tool/eval.py吗?要在yaml文件里怎么配置才可以输出Recall F1-score分数

为什么PaddleHub加载的RexNet预训练模型的Flops数量远低于官方论文的Flops和Pytorch测得的Flops

1

加载[rexnet_1_0_imagenet](https://github.com/PaddlePaddle/PaddleHub/blob/release/v2.2/modules/image/classification/rexnet_1_0_imagenet)这个预训练模型 用 FLOPs = paddle.flops(model, [32, 3, 3, 3], custom_ops= {nn.LeakyReLU: count_leaky_relu},print_detail=True)测得的FLOPs只有0.154G而官方论文的Flops有0.415 而用Timm-Pytorch测得的有0.48Flops 不知道为何？

There was something wrong in ./models/smpl.py self.faces = data_struct.f AttributeError: 'Struct' object has no attribute 'f'

1

Thanks for your excellent work! Traceback (most recent call last): File "render_360.py", line 164, in main_posed_360(opt) File "render_360.py", line 91, in main_posed_360 smpl_type='optimized' File "/home/zhaozisong/ml-neuman/data_io/neuman_helper.py", line 245, in read_scene smpls,...

How to solve ”No module named app.inference“ problem ？

3

python -m app.inference --config configs/real_demo.yaml --dataroot thuman_demo --view_list 0 1 $$: No module named app.inference

PadddleFL加载预训练模型 可以加在PaddleHub里CV的模型吗？还是只可以加载PaddleFL的联邦学习的模型