Cezary Czernecki

I found in previous commits that output used to be, but later due to log_likelihood_cost refactor got deleted. output = tf.nn.softmax(tf.reshape(h_conv_final, [-1, go.N ** 2]) + b_conv_final)

Hey, It might result in such error, to fix output issue you can add this line: self.output = tf.nn.softmax(tf.reshape(h_conv_final, [-1, go.N ** 2]) + b_conv_final) at line 88 and it...

Hey, log_likelihood_cost is another problem, as now you need named logits parameter. I am happy that you managed to fix it. I am not sure what you are talking about...

Hello, Move that is to play is solely chosen based on visit count, what are they for those two nodes? I've prepared a silly image describing what might be going...

@JoeyQWu The move that is chosen to be played in the actual game yes. Not to confuse with move chosen within selection phase - this one is chosen based on...