Profiler without Cloud Storage bucket

[qo4on]

Hi!
As far as I know TPUs usually write to Cloud Storage bucket and you have a very nice tutorial about that. But with recent tf update it is now possible to save checkpoints locally:

# Writing checkpoints locally from a TPU model
save_locally = tf.saved_model.SaveOptions(experimental_io_device='/job:localhost')
checkpoints_cb = tf.keras.callbacks.ModelCheckpoint('./checkpoints', options=save_locally)
model.fit(…, callbacks=[checkpoints_cb])

Can I use similar approach to write profile to local Colab runtime? If it's not possible right now will you add this feature later?