how increase floating point precision representation of tensorflow variables? need cause network large , data complex, want see amount of improvement no matter how small. when iterate through training print mean error screen, , see same 6 digits - works fine less complex inputs. note tensorboard seems have similar precision, happy enough more precise tensorboard graph.
msquarederror=m_sqerror.eval(sessions=sess,feed_dict={input:ip, output=op,keep_prob:1.0}) print ("mse: %9f"%msquarederror) output:
mse: 0.317513
desired output:
mse: 0.317513223 ... , many more digits
stoopid error.
print ("mse: %9f"%msquarederror) should be
print ("mse: %.9f"%msquarederror) 