Closed Electro1111 closed 3 years ago
also --in_tensorname
I have found what I believe to be the correct tensors for these arguments, Im just not sure about the logit_tensorname
--out_tensorname dense_3/Softmax:0 --in_tensorname input_1:0 --logit_tensorname (dense_3/MatMul:0 or dense_3/BiasAdd:0) --label_tensorname dense_3_target:0 --weights_tensorname dense_3_sample_weights:0
Sorry for the late response, but I believe dense_3/BiasAdd:0
is the correct logits tensor.
What are the correct arguments for these two models for training and inference purposes?