h8907283 comments

Results 4 comments of


                                            h8907283

`dm_control` `cheetah` `run` training stops suddenly

Yes, `walker run` is okay, but not `cheetah run`.

Using batch_size * num_gpus as batch_size in exponential_decay calculation?

Thanks @reedwm for your answer. The training in question is MobileNet v1. The MobileNet v1 paper didn't mention much about training parameters. However, you can find them in mobilenet_v1_train.py (https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet_v1_train.py)...

Using batch_size * num_gpus as batch_size in exponential_decay calculation?

Thanks @nealwu. The following is in https://github.com/tensorflow/models/blob/master/tutorials/image/cifar10_estimator/cifar10_main.py: num_batches_per_epoch = cifar10.Cifar10DataSet.num_examples_per_epoch('train') // (params.train_batch_size * num_workers) It looks like you're in agreement with @reedwm. Now back to the questions in my previous...

Runtime error when calling freeze_transformer_layers

Any update on this? Is it okay if I open a PR? Thanks so much!