ó ųĩČ[c@@s<dZddlmZddlZdefd„ƒYZdS(sJTensorBoard functions that can be used to log various status during epoch.i(tabsolute_importNtLogMetricsCallbackcB@s#eZdZdd„Zd„ZRS(sĘLog metrics periodically in TensorBoard. This callback works almost same as `callback.Speedometer`, but write TensorBoard event file for visualization. For more usage, please refer https://github.com/dmlc/tensorboard Parameters ---------- logging_dir : str TensorBoard event file directory. After that, use `tensorboard --logdir=path/to/logs` to launch TensorBoard visualization. prefix : str Prefix for a metric name of `scalar` value. You might want to use this param to leverage TensorBoard plot feature, where TensorBoard plots different curves in one graph when they have same `name`. The follow example shows the usage(how to compare a train and eval metric in a same graph). Examples -------- >>> # log train and eval metrics under different directories. >>> training_log = 'logs/train' >>> evaluation_log = 'logs/eval' >>> # in this case, each training and evaluation metric pairs has same name, >>> # you can add a prefix to make it separate. >>> batch_end_callbacks = [mx.contrib.tensorboard.LogMetricsCallback(training_log)] >>> eval_end_callbacks = [mx.contrib.tensorboard.LogMetricsCallback(evaluation_log)] >>> # run >>> model.fit(train, >>> ... >>> batch_end_callback = batch_end_callbacks, >>> eval_end_callback = eval_end_callbacks) >>> # Then use `tensorboard --logdir=logs/` to launch TensorBoard visualization. cC@sQ||_y#ddlm}||ƒ|_Wntk rLtjdƒnXdS(Ni(t SummaryWriters2You can install mxboard via `pip install mxboard`.(tprefixtmxboardRtsummary_writert ImportErrortloggingterror(tselft logging_dirRR((sY/usr/local/lib/python2.7/site-packages/mxnet-1.3.1-py2.7.egg/mxnet/contrib/tensorboard.pyt__init__9s   cC@s~|jdkrdS|jjƒ}xU|D]M\}}|jdk rZd|j|f}n|jj||d|jƒq)WdS(s:Callback to log training speed and metrics in TensorBoard.Ns%s-%st global_step(t eval_metrictNonetget_name_valueRRt add_scalartepoch(R tparamt name_valuetnametvalue((sY/usr/local/lib/python2.7/site-packages/mxnet-1.3.1-py2.7.egg/mxnet/contrib/tensorboard.pyt__call__AsN(t__name__t __module__t__doc__RR R(((sY/usr/local/lib/python2.7/site-packages/mxnet-1.3.1-py2.7.egg/mxnet/contrib/tensorboard.pyRs (Rt __future__RRtobjectR(((sY/usr/local/lib/python2.7/site-packages/mxnet-1.3.1-py2.7.egg/mxnet/contrib/tensorboard.pyts