wandb.restore(filename)will restore a file into your local run directory. Typically
filename refers to a file generated by an earlier experiment run and uploaded to our cloud. This call will make a local copy of the file and return a local file stream open for reading.
wandb.restore accepts a few optional keyword arguments:
run_path — string referring to the earlier run from which to pull the file, formatted as '$ENTITY_NAME/$PROJECT_NAME/$RUN_ID' or '$PROJECT_NAME/$RUN_ID' (default: current entity, project name, and run id)
replace — boolean specifying whether to overwrite a local copy of filename with the cloud copy, if a local copy turns out to be available (default: False)
root — string specifying the directory in which to store the local copy of the file. This defaults to the current working directory, or the
wandb.run.dirif wandb.init was called earlier (default: ".")
Common use cases:
restore the model architecture or weights generated by past runs
resume training from the last checkpoint in the case of failure (see the section on resuming for crucial details)
See this report for a complete working example.
# restore a model file from a specific run by user "vanpelt" in "my-project"best_model = wandb.restore('model-best.h5', run_path="vanpelt/my-project/a1b2c3d")# restore a weights file from a checkpoint# (NOTE: resuming must be configured if run_path is not provided)weights_file = wandb.restore('weights.h5')# use the "name" attribute of the returned object# if your framework expects a filename, e.g. as in Kerasmy_predefined_model.load_weights(weights_file.name)