pointer-generator
pointer-generator copied to clipboard
Question about the training loss.
Hi, abisee! Thanks for sharing this good work! I have some question about your paper in Experiments part.
You said that with the coverage loss ratio set to 1.0 the initial loss should be about 0.5 and after about 3000 iterations it is 0.2, but my experiment reports that the initial loss of coverage is 0.97
INFO:tensorflow:Starting queue runners.
INFO:tensorflow:Created session.
INFO:tensorflow:starting run_training
INFO:tensorflow:running training step...
INFO:tensorflow:global_step/sec: 0
INFO:tensorflow:seconds for training step: 4.211
INFO:tensorflow:loss: 6.003044
INFO:tensorflow:coverage_loss: 0.979794
Can you help me ? Thanks a lot !