mirror of https://github.com/kubeflow/examples.git
* Updates to the demo docs (notebook and readme) which were out-dated in multiple places * Removed unused tools/ dir * Update main readme to reference the example * Inclusion of kubeflow vendor/ tf-job code * Illustrates how logging and rendering to an attached volume can simplify the process of viewing logs with TensorHub and exploring render outputs. * Storage in user-space allows it to be sym-linked into directory tree watched by TensorHub extension (which is running tensorboard --logdir=/home/jovyan) * I anticipate this current approach to controlling volume mounts for NFS through ksonnet to be replaced by doing so with python as I demonstrated in the enhance example so I wouldn't lose sleep over the ksonnet prototypes in this commit. |
||
|---|---|---|
| .. | ||
| __init__.py | ||
| task.py | ||