“5b6673c44df19ee89c191cc7efeb669c1943050f”上不存在“test/legacy_test/test_fuse_relu_depthwise_conv_pass.py”
  • E
    Fix sparse training for trainer_count=1 (#204) · 28bc05b1
    emailweixu 提交于
    * Fix sparse training for trainer_count=1
    
    For trainer_count=1, the gradient machine is NeuralNetwork, which does not create parameter buf for PARAMETER_GRADIENT for sparse update in Parameter::enableType. But gradient parameter buf is still used in SgdThreadUpdater.
    
    * Minor update to comment
    28bc05b1
Parameter.h 11.0 KB