Basic_CNNs_TensorFlow2 icon indicating copy to clipboard operation
Basic_CNNs_TensorFlow2 copied to clipboard

run error in regnet with GPU

Open tongshiyuan opened this issue 3 years ago • 1 comments

我在使用GPU进行RegNet训练的时候,报错了

UnimplementedError:  Fused conv implementation does not support grouped convolutions for now.
	 [[node reg_net/any_stage/res_bottleneck_block/bottleneck_transform/conv2d_3/BiasAdd (defined at /filepath/anynet.py:119) ]] [Op:__inference_train_function_10336]

Function call stack:
train_function

但是CPU下是正常的,我检查了一下代码,GPU和CPU模式下,get_group_conv是不一样的,但我不知道怎么debug 希望能得到您的回复

tongshiyuan avatar May 14 '21 15:05 tongshiyuan

根据实际情况更改配置文件中的这一项:DEVICE = "gpu"就行

calmiLovesAI avatar Oct 14 '21 05:10 calmiLovesAI