.. DO NOT EDIT. .. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY. .. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE: .. "intermediate/ax_multiobjective_nas_tutorial.py" .. LINE NUMBERS ARE GIVEN BELOW. .. only:: html .. note:: :class: sphx-glr-download-link-note Click :ref:`here ` to download the full example code .. rst-class:: sphx-glr-example-title .. _sphx_glr_intermediate_ax_multiobjective_nas_tutorial.py: Multi-Objective NAS with Ax ================================================== **Authors:** `David Eriksson `__, `Max Balandat `__, and the Adaptive Experimentation team at Meta. In this tutorial, we show how to use `Ax `__ to run multi-objective neural architecture search (NAS) for a simple neural network model on the popular MNIST dataset. While the underlying methodology would typically be used for more complicated models and larger datasets, we opt for a tutorial that is easily runnable end-to-end on a laptop in less than 20 minutes. In many NAS applications, there is a natural tradeoff between multiple objectives of interest. For instance, when deploying models on-device we may want to maximize model performance (for example, accuracy), while simultaneously minimizing competing metrics like power consumption, inference latency, or model size in order to satisfy deployment constraints. Often, we may be able to reduce computational requirements or latency of predictions substantially by accepting minimally lower model performance. Principled methods for exploring such tradeoffs efficiently are key enablers of scalable and sustainable AI, and have many successful applications at Meta - see for instance our `case study `__ on a Natural Language Understanding model. In our example here, we will tune the widths of two hidden layers, the learning rate, the dropout probability, the batch size, and the number of training epochs. The goal is to trade off performance (accuracy on the validation set) and model size (the number of model parameters). This tutorial makes use of the following PyTorch libraries: - `PyTorch Lightning `__ (specifying the model and training loop) - `TorchX `__ (for running training jobs remotely / asynchronously) - `BoTorch `__ (the Bayesian Optimization library powering Ax's algorithms) .. GENERATED FROM PYTHON SOURCE LINES 45-55 Defining the TorchX App ----------------------- Our goal is to optimize the PyTorch Lightning training job defined in `mnist_train_nas.py `__. To do this using TorchX, we write a helper function that takes in the values of the architecture and hyperparameters of the training job and creates a `TorchX AppDef `__ with the appropriate settings. .. GENERATED FROM PYTHON SOURCE LINES 55-102 .. code-block:: default from pathlib import Path import torchx from torchx import specs from torchx.components import utils def trainer( log_path: str, hidden_size_1: int, hidden_size_2: int, learning_rate: float, epochs: int, dropout: float, batch_size: int, trial_idx: int = -1, ) -> specs.AppDef: # define the log path so we can pass it to the TorchX ``AppDef`` if trial_idx >= 0: log_path = Path(log_path).joinpath(str(trial_idx)).absolute().as_posix() return utils.python( # command line arguments to the training script "--log_path", log_path, "--hidden_size_1", str(hidden_size_1), "--hidden_size_2", str(hidden_size_2), "--learning_rate", str(learning_rate), "--epochs", str(epochs), "--dropout", str(dropout), "--batch_size", str(batch_size), # other config options name="trainer", script="mnist_train_nas.py", image=torchx.version.TORCHX_IMAGE, ) .. GENERATED FROM PYTHON SOURCE LINES 103-117 Setting up the Runner --------------------- Ax’s `Runner `__ abstraction allows writing interfaces to various backends. Ax already comes with Runner for TorchX, and so we just need to configure it. For the purpose of this tutorial we run jobs locally in a fully asynchronous fashion. In order to launch them on a cluster, you can instead specify a different TorchX scheduler and adjust the configuration appropriately. For example, if you have a Kubernetes cluster, you just need to change the scheduler from ``local_cwd`` to ``kubernetes``). .. GENERATED FROM PYTHON SOURCE LINES 117-135 .. code-block:: default import tempfile from ax.runners.torchx import TorchXRunner # Make a temporary dir to log our results into log_dir = tempfile.mkdtemp() ax_runner = TorchXRunner( tracker_base="/tmp/", component=trainer, # NOTE: To launch this job on a cluster instead of locally you can # specify a different scheduler and adjust arguments appropriately. scheduler="local_cwd", component_const_params={"log_path": log_dir}, cfg={}, ) .. GENERATED FROM PYTHON SOURCE LINES 136-146 Setting up the ``SearchSpace`` ------------------------------ First, we define our search space. Ax supports both range parameters of type integer and float as well as choice parameters which can have non-numerical types such as strings. We will tune the hidden sizes, learning rate, dropout, and number of epochs as range parameters and tune the batch size as an ordered choice parameter to enforce it to be a power of 2. .. GENERATED FROM PYTHON SOURCE LINES 146-209 .. code-block:: default from ax.core import ( ChoiceParameter, ParameterType, RangeParameter, SearchSpace, ) parameters = [ # NOTE: In a real-world setting, hidden_size_1 and hidden_size_2 # should probably be powers of 2, but in our simple example this # would mean that ``num_params`` can't take on that many values, which # in turn makes the Pareto frontier look pretty weird. RangeParameter( name="hidden_size_1", lower=16, upper=128, parameter_type=ParameterType.INT, log_scale=True, ), RangeParameter( name="hidden_size_2", lower=16, upper=128, parameter_type=ParameterType.INT, log_scale=True, ), RangeParameter( name="learning_rate", lower=1e-4, upper=1e-2, parameter_type=ParameterType.FLOAT, log_scale=True, ), RangeParameter( name="epochs", lower=1, upper=4, parameter_type=ParameterType.INT, ), RangeParameter( name="dropout", lower=0.0, upper=0.5, parameter_type=ParameterType.FLOAT, ), ChoiceParameter( # NOTE: ``ChoiceParameters`` don't require log-scale name="batch_size", values=[32, 64, 128, 256], parameter_type=ParameterType.INT, is_ordered=True, sort_values=True, ), ] search_space = SearchSpace( parameters=parameters, # NOTE: In practice, it may make sense to add a constraint # hidden_size_2 <= hidden_size_1 parameter_constraints=[], ) .. GENERATED FROM PYTHON SOURCE LINES 210-234 Setting up Metrics ------------------ Ax has the concept of a `Metric `__ that defines properties of outcomes and how observations are obtained for these outcomes. This allows e.g. encoding how data is fetched from some distributed execution backend and post-processed before being passed as input to Ax. In this tutorial we will use `multi-objective optimization `__ with the goal of maximizing the validation accuracy and minimizing the number of model parameters. The latter represents a simple proxy of model latency, which is hard to estimate accurately for small ML models (in an actual application we would benchmark the latency while running the model on-device). In our example TorchX will run the training jobs in a fully asynchronous fashion locally and write the results to the ``log_dir`` based on the trial index (see the ``trainer()`` function above). We will define a metric class that is aware of that logging directory. By subclassing `TensorboardCurveMetric `__ we get the logic to read and parse the TensorBoard logs for free. .. GENERATED FROM PYTHON SOURCE LINES 234-259 .. code-block:: default from ax.metrics.tensorboard import TensorboardCurveMetric class MyTensorboardMetric(TensorboardCurveMetric): # NOTE: We need to tell the new TensorBoard metric how to get the id / # file handle for the TensorBoard logs from a trial. In this case # our convention is to just save a separate file per trial in # the prespecified log dir. @classmethod def get_ids_from_trials(cls, trials): return { trial.index: Path(log_dir).joinpath(str(trial.index)).as_posix() for trial in trials } # This indicates whether the metric is queryable while the trial is # still running. We don't use this in the current tutorial, but Ax # utilizes this to implement trial-level early-stopping functionality. @classmethod def is_available_while_running(cls): return False .. GENERATED FROM PYTHON SOURCE LINES 260-266 Now we can instantiate the metrics for accuracy and the number of model parameters. Here `curve_name` is the name of the metric in the TensorBoard logs, while `name` is the metric name used internally by Ax. We also specify `lower_is_better` to indicate the favorable direction of the two metrics. .. GENERATED FROM PYTHON SOURCE LINES 266-279 .. code-block:: default val_acc = MyTensorboardMetric( name="val_acc", curve_name="val_acc", lower_is_better=False, ) model_num_params = MyTensorboardMetric( name="num_params", curve_name="num_params", lower_is_better=True, ) .. GENERATED FROM PYTHON SOURCE LINES 280-294 Setting up the ``OptimizationConfig`` ------------------------------------- The way to tell Ax what it should optimize is by means of an `OptimizationConfig `__. Here we use a ``MultiObjectiveOptimizationConfig`` as we will be performing multi-objective optimization. Additionally, Ax supports placing constraints on the different metrics by specifying objective thresholds, which bound the region of interest in the outcome space that we want to explore. For this example, we will constrain the validation accuracy to be at least 0.94 (94%) and the number of model parameters to be at most 80,000. .. GENERATED FROM PYTHON SOURCE LINES 294-313 .. code-block:: default from ax.core import MultiObjective, Objective, ObjectiveThreshold from ax.core.optimization_config import MultiObjectiveOptimizationConfig opt_config = MultiObjectiveOptimizationConfig( objective=MultiObjective( objectives=[ Objective(metric=val_acc, minimize=False), Objective(metric=model_num_params, minimize=True), ], ), objective_thresholds=[ ObjectiveThreshold(metric=val_acc, bound=0.94, relative=False), ObjectiveThreshold(metric=model_num_params, bound=80_000, relative=False), ], ) .. GENERATED FROM PYTHON SOURCE LINES 314-327 Creating the Ax Experiment -------------------------- In Ax, the `Experiment `__ object is the object that stores all the information about the problem setup. .. tip: ``Experiment`` objects can be serialized to JSON or stored to a database backend such as MySQL in order to persist and be available to load on different machines. See the the `Ax Docs `__ on the storage functionality for details. .. GENERATED FROM PYTHON SOURCE LINES 327-337 .. code-block:: default from ax.core import Experiment experiment = Experiment( name="torchx_mnist", search_space=search_space, optimization_config=opt_config, runner=ax_runner, ) .. GENERATED FROM PYTHON SOURCE LINES 338-352 Choosing the Generation Strategy -------------------------------- A `GenerationStrategy `__ is the abstract representation of how we would like to perform the optimization. While this can be customized (if you’d like to do so, see `this tutorial `__), in most cases Ax can automatically determine an appropriate strategy based on the search space, optimization config, and the total number of trials we want to run. Typically, Ax chooses to evaluate a number of random configurations before starting a model-based Bayesian Optimization strategy. .. GENERATED FROM PYTHON SOURCE LINES 352-365 .. code-block:: default total_trials = 48 # total evaluation budget from ax.modelbridge.dispatch_utils import choose_generation_strategy gs = choose_generation_strategy( search_space=experiment.search_space, optimization_config=experiment.optimization_config, num_trials=total_trials, ) .. rst-class:: sphx-glr-script-out .. code-block:: none [INFO 04-30 15:27:03] ax.modelbridge.dispatch_utils: Using Models.BOTORCH_MODULAR since there is at least one ordered parameter and there are no unordered categorical parameters. [INFO 04-30 15:27:03] ax.modelbridge.dispatch_utils: Calculating the number of remaining initialization trials based on num_initialization_trials=None max_initialization_trials=None num_tunable_parameters=6 num_trials=48 use_batch_trials=False [INFO 04-30 15:27:03] ax.modelbridge.dispatch_utils: calculated num_initialization_trials=9 [INFO 04-30 15:27:03] ax.modelbridge.dispatch_utils: num_completed_initialization_trials=0 num_remaining_initialization_trials=9 [INFO 04-30 15:27:03] ax.modelbridge.dispatch_utils: `verbose`, `disable_progbar`, and `jit_compile` are not yet supported when using `choose_generation_strategy` with ModularBoTorchModel, dropping these arguments. [INFO 04-30 15:27:03] ax.modelbridge.dispatch_utils: Using Bayesian Optimization generation strategy: GenerationStrategy(name='Sobol+BoTorch', steps=[Sobol for 9 trials, BoTorch for subsequent trials]). Iterations after 9 will take longer to generate due to model-fitting. .. GENERATED FROM PYTHON SOURCE LINES 366-389 Configuring the Scheduler ------------------------- The ``Scheduler`` acts as the loop control for the optimization. It communicates with the backend to launch trials, check their status, and retrieve results. In the case of this tutorial, it is simply reading and parsing the locally saved logs. In a remote execution setting, it would call APIs. The following illustration from the Ax `Scheduler tutorial `__ summarizes how the Scheduler interacts with external systems used to run trial evaluations: .. image:: ../../_static/img/ax_scheduler_illustration.png The ``Scheduler`` requires the ``Experiment`` and the ``GenerationStrategy``. A set of options can be passed in via ``SchedulerOptions``. Here, we configure the number of total evaluations as well as ``max_pending_trials``, the maximum number of trials that should run concurrently. In our local setting, this is the number of training jobs running as individual processes, while in a remote execution setting, this would be the number of machines you want to use in parallel. .. GENERATED FROM PYTHON SOURCE LINES 389-402 .. code-block:: default from ax.service.scheduler import Scheduler, SchedulerOptions scheduler = Scheduler( experiment=experiment, generation_strategy=gs, options=SchedulerOptions( total_trials=total_trials, max_pending_trials=4 ), ) .. rst-class:: sphx-glr-script-out .. code-block:: none [WARNING 04-30 15:27:03] ax.service.utils.with_db_settings_base: Ax currently requires a sqlalchemy version below 2.0. This will be addressed in a future release. Disabling SQL storage in Ax for now, if you would like to use SQL storage please install Ax with mysql extras via `pip install ax-platform[mysql]`. [INFO 04-30 15:27:03] Scheduler: `Scheduler` requires experiment to have immutable search space and optimization config. Setting property immutable_search_space_and_opt_config to `True` on experiment. .. GENERATED FROM PYTHON SOURCE LINES 403-413 Running the optimization ------------------------ Now that everything is configured, we can let Ax run the optimization in a fully automated fashion. The Scheduler will periodically check the logs for the status of all currently running trials, and if a trial completes the scheduler will update its status on the experiment and fetch the observations needed for the Bayesian optimization algorithm. .. GENERATED FROM PYTHON SOURCE LINES 413-417 .. code-block:: default scheduler.run_all_trials() .. rst-class:: sphx-glr-script-out .. code-block:: none [INFO 04-30 15:27:03] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:03] Scheduler: Running trials [0]... [INFO 04-30 15:27:04] Scheduler: Running trials [1]... [INFO 04-30 15:27:05] Scheduler: Running trials [2]... [INFO 04-30 15:27:06] Scheduler: Running trials [3]... [INFO 04-30 15:27:07] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:07] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 4). [INFO 04-30 15:27:08] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:08] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 4). [INFO 04-30 15:27:10] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:10] Scheduler: Retrieved FAILED trials: [0]. [INFO 04-30 15:27:10] Scheduler: Running trials [4]... [INFO 04-30 15:27:11] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:11] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 4). [INFO 04-30 15:27:12] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:12] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 4). [INFO 04-30 15:27:13] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:13] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 4). [INFO 04-30 15:27:15] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:15] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 4). [INFO 04-30 15:27:19] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:19] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 4). [INFO 04-30 15:27:24] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:24] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 4). [INFO 04-30 15:27:31] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:31] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 4). [INFO 04-30 15:27:43] Scheduler: Fetching data for newly completed trials: [4]. [INFO 04-30 15:27:43] Scheduler: Retrieved COMPLETED trials: [4]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:27:43] Scheduler: Running trials [5]... [INFO 04-30 15:27:44] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:44] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 4). [INFO 04-30 15:27:45] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:45] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 4). [INFO 04-30 15:27:47] Scheduler: Fetching data for newly completed trials: [1]. [INFO 04-30 15:27:47] Scheduler: Retrieved COMPLETED trials: [1]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:27:47] Scheduler: Running trials [6]... [INFO 04-30 15:27:48] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:48] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 4). [INFO 04-30 15:27:49] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:49] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 4). [INFO 04-30 15:27:50] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:50] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 4). [INFO 04-30 15:27:52] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:52] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 4). [INFO 04-30 15:27:56] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:27:56] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 4). [INFO 04-30 15:28:01] Scheduler: Fetching data for newly completed trials: [3]. [INFO 04-30 15:28:01] Scheduler: Retrieved COMPLETED trials: [3]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:28:01] Scheduler: Running trials [7]... [INFO 04-30 15:28:02] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:02] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 4). [INFO 04-30 15:28:03] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:03] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 4). [INFO 04-30 15:28:04] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:04] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 4). [INFO 04-30 15:28:07] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:07] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 4). [INFO 04-30 15:28:10] Scheduler: Fetching data for newly completed trials: [2]. [INFO 04-30 15:28:10] Scheduler: Retrieved COMPLETED trials: [2]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:28:10] Scheduler: Running trials [8]... [INFO 04-30 15:28:11] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:11] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 4). [INFO 04-30 15:28:12] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:12] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 4). [INFO 04-30 15:28:14] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:14] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 4). [INFO 04-30 15:28:16] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:16] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 4). [INFO 04-30 15:28:19] Scheduler: Fetching data for newly completed trials: 5 - 6. [INFO 04-30 15:28:19] Scheduler: Retrieved COMPLETED trials: 5 - 6. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:28:19] Scheduler: Running trials [9]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:28:21] Scheduler: Running trials [10]... [INFO 04-30 15:28:22] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:22] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 4). [INFO 04-30 15:28:23] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:23] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 4). [INFO 04-30 15:28:25] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:25] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 4). [INFO 04-30 15:28:27] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:27] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 4). [INFO 04-30 15:28:30] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:30] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 4). [INFO 04-30 15:28:36] Scheduler: Fetching data for newly completed trials: [8]. [INFO 04-30 15:28:36] Scheduler: Retrieved COMPLETED trials: [8]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:28:39] Scheduler: Running trials [11]... [INFO 04-30 15:28:40] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:40] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 4). [INFO 04-30 15:28:41] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:41] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 4). [INFO 04-30 15:28:42] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:42] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 4). [INFO 04-30 15:28:45] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:45] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 4). [INFO 04-30 15:28:48] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:48] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 4). [INFO 04-30 15:28:53] Scheduler: Fetching data for newly completed trials: [7, 9]. [INFO 04-30 15:28:53] Scheduler: Retrieved COMPLETED trials: [7, 9]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:28:55] Scheduler: Running trials [12]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:28:56] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:28:56] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:28:56] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:56] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:28:57] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:57] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:28:59] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:28:59] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:29:01] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:29:01] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:29:04] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:29:04] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:29:09] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:29:09] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:29:17] Scheduler: Fetching data for newly completed trials: [10]. [INFO 04-30 15:29:17] Scheduler: Retrieved COMPLETED trials: [10]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:29:19] Scheduler: Running trials [13]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:29:20] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:29:20] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:29:20] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:29:20] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:29:21] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:29:21] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:29:23] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:29:23] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:29:25] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:29:25] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:29:28] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:29:28] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:29:33] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:29:33] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:29:41] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:29:41] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 3). [INFO 04-30 15:29:52] Scheduler: Fetching data for newly completed trials: [11]. [INFO 04-30 15:29:52] Scheduler: Retrieved COMPLETED trials: [11]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:29:55] Scheduler: Running trials [14]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:29:56] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:29:56] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:29:56] Scheduler: Fetching data for newly completed trials: [13]. [INFO 04-30 15:29:56] Scheduler: Retrieved COMPLETED trials: [13]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:29:58] Scheduler: Running trials [15]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:29:59] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:29:59] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:29:59] Scheduler: Fetching data for newly completed trials: [12]. [INFO 04-30 15:29:59] Scheduler: Retrieved COMPLETED trials: [12]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:30:02] Scheduler: Running trials [16]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:30:03] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:30:03] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:30:03] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:03] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:30:04] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:04] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:30:06] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:06] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:30:08] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:08] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:30:11] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:11] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:30:16] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:16] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:30:24] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:24] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 3). [INFO 04-30 15:30:36] Scheduler: Fetching data for newly completed trials: [15]. [INFO 04-30 15:30:36] Scheduler: Retrieved COMPLETED trials: [15]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:30:38] Scheduler: Running trials [17]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:30:39] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:30:39] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:30:39] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:39] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:30:40] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:40] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:30:42] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:42] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:30:44] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:44] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:30:48] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:48] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:30:53] Scheduler: Fetching data for newly completed trials: [16]. [INFO 04-30 15:30:53] Scheduler: Retrieved COMPLETED trials: [16]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:30:56] Scheduler: Running trials [18]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:30:57] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:30:57] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:30:57] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:57] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:30:58] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:30:58] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:30:59] Scheduler: Fetching data for newly completed trials: [14]. [INFO 04-30 15:30:59] Scheduler: Retrieved COMPLETED trials: [14]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:31:03] Scheduler: Running trials [19]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:31:03] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:31:03] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:31:03] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:03] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:31:04] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:04] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:31:05] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:05] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:31:07] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:07] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:31:11] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:11] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:31:16] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:16] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:31:24] Scheduler: Fetching data for newly completed trials: [17]. [INFO 04-30 15:31:24] Scheduler: Retrieved COMPLETED trials: [17]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:31:27] Scheduler: Running trials [20]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:31:28] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:31:28] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:31:28] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:28] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:31:29] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:29] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:31:30] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:30] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:31:32] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:32] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:31:36] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:36] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:31:41] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:41] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:31:49] Scheduler: Fetching data for newly completed trials: [18]. [INFO 04-30 15:31:49] Scheduler: Retrieved COMPLETED trials: [18]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:31:52] Scheduler: Running trials [21]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:31:52] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:31:52] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:31:52] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:52] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:31:53] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:53] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:31:55] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:55] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:31:57] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:31:57] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:32:00] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:00] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:32:05] Scheduler: Fetching data for newly completed trials: [19]. [INFO 04-30 15:32:05] Scheduler: Retrieved COMPLETED trials: [19]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:32:09] Scheduler: Running trials [22]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:32:10] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:32:10] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:32:10] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:10] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:32:11] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:11] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:32:13] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:13] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:32:15] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:15] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:32:18] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:18] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:32:23] Scheduler: Fetching data for newly completed trials: [20]. [INFO 04-30 15:32:23] Scheduler: Retrieved COMPLETED trials: [20]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:32:29] Scheduler: Running trials [23]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:32:30] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:32:30] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:32:30] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:30] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:32:31] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:31] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:32:33] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:33] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:32:35] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:35] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:32:38] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:38] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:32:44] Scheduler: Fetching data for newly completed trials: [22]. [INFO 04-30 15:32:44] Scheduler: Retrieved COMPLETED trials: [22]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:32:50] Scheduler: Running trials [24]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:32:51] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:32:51] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:32:51] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:51] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:32:52] Scheduler: Fetching data for newly completed trials: [21]. [INFO 04-30 15:32:52] Scheduler: Retrieved COMPLETED trials: [21]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:32:57] Scheduler: Running trials [25]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:32:58] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:32:58] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:32:58] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:58] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:32:59] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:32:59] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:33:00] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:33:00] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:33:02] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:33:02] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:33:06] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:33:06] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:33:11] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:33:11] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:33:18] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:33:18] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 3). [INFO 04-30 15:33:30] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:33:30] Scheduler: Waiting for completed trials (for 17 sec, currently running trials: 3). [INFO 04-30 15:33:47] Scheduler: Fetching data for newly completed trials: 23 - 24. [INFO 04-30 15:33:47] Scheduler: Retrieved COMPLETED trials: 23 - 24. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:33:52] Scheduler: Running trials [26]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:33:53] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:33:56] Scheduler: Running trials [27]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:33:57] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:33:57] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:33:57] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:33:57] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:33:58] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:33:58] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:34:00] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:34:00] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:34:02] Scheduler: Fetching data for newly completed trials: [25]. [INFO 04-30 15:34:02] Scheduler: Retrieved COMPLETED trials: [25]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:34:09] Scheduler: Running trials [28]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:34:10] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:34:10] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:34:10] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:34:10] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:34:11] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:34:11] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:34:13] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:34:13] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:34:15] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:34:15] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:34:18] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:34:18] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:34:23] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:34:23] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:34:31] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:34:31] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 3). [INFO 04-30 15:34:42] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:34:42] Scheduler: Waiting for completed trials (for 17 sec, currently running trials: 3). [INFO 04-30 15:35:00] Scheduler: Fetching data for newly completed trials: 26 - 27. [INFO 04-30 15:35:00] Scheduler: Retrieved COMPLETED trials: 26 - 27. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:35:04] Scheduler: Running trials [29]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:35:05] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:35:09] Scheduler: Running trials [30]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:35:10] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:35:10] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:35:10] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:35:10] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:35:11] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:35:11] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:35:13] Scheduler: Fetching data for newly completed trials: [28]. [INFO 04-30 15:35:13] Scheduler: Retrieved COMPLETED trials: [28]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:35:18] Scheduler: Running trials [31]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:35:18] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:35:18] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:35:18] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:35:18] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:35:19] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:35:19] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:35:21] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:35:21] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:35:23] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:35:23] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:35:26] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:35:26] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:35:31] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:35:31] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:35:39] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:35:39] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 3). [INFO 04-30 15:35:50] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:35:50] Scheduler: Waiting for completed trials (for 17 sec, currently running trials: 3). [INFO 04-30 15:36:08] Scheduler: Fetching data for newly completed trials: [29]. [INFO 04-30 15:36:08] Scheduler: Retrieved COMPLETED trials: [29]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:36:14] Scheduler: Running trials [32]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:36:15] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:36:15] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:36:15] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:15] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:36:16] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:16] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:36:18] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:18] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:36:20] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:20] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:36:23] Scheduler: Fetching data for newly completed trials: [31]. [INFO 04-30 15:36:23] Scheduler: Retrieved COMPLETED trials: [31]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:36:28] Scheduler: Running trials [33]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:36:29] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:36:29] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:36:29] Scheduler: Fetching data for newly completed trials: [30]. [INFO 04-30 15:36:29] Scheduler: Retrieved COMPLETED trials: [30]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:36:35] Scheduler: Running trials [34]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:36:36] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:36:36] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:36:36] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:36] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:36:37] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:37] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:36:38] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:38] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:36:41] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:41] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:36:44] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:44] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:36:49] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:49] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:36:57] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:36:57] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 3). [INFO 04-30 15:37:08] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:37:08] Scheduler: Waiting for completed trials (for 17 sec, currently running trials: 3). [INFO 04-30 15:37:25] Scheduler: Fetching data for newly completed trials: 32 - 33. [INFO 04-30 15:37:25] Scheduler: Retrieved COMPLETED trials: 32 - 33. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:37:32] Scheduler: Running trials [35]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:37:33] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:37:38] Scheduler: Running trials [36]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:37:39] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:37:39] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:37:39] Scheduler: Fetching data for newly completed trials: [34]. [INFO 04-30 15:37:39] Scheduler: Retrieved COMPLETED trials: [34]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:37:48] Scheduler: Running trials [37]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:37:49] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:37:49] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:37:49] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:37:49] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:37:50] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:37:50] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:37:52] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:37:52] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:37:54] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:37:54] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:37:58] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:37:58] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:38:03] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:38:03] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:38:10] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:38:10] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 3). [INFO 04-30 15:38:22] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:38:22] Scheduler: Waiting for completed trials (for 17 sec, currently running trials: 3). [INFO 04-30 15:38:39] Scheduler: Fetching data for newly completed trials: 35 - 36. [INFO 04-30 15:38:39] Scheduler: Retrieved COMPLETED trials: 35 - 36. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:38:44] Scheduler: Running trials [38]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:38:45] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:38:51] Scheduler: Running trials [39]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:38:52] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:38:52] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:38:52] Scheduler: Fetching data for newly completed trials: [37]. [INFO 04-30 15:38:52] Scheduler: Retrieved COMPLETED trials: [37]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:38:59] Scheduler: Running trials [40]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:39:00] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:39:00] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:39:00] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:39:00] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:39:01] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:39:01] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:39:02] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:39:02] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:39:05] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:39:05] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:39:08] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:39:08] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:39:13] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:39:13] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:39:21] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:39:21] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 3). [INFO 04-30 15:39:32] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:39:32] Scheduler: Waiting for completed trials (for 17 sec, currently running trials: 3). [INFO 04-30 15:39:49] Scheduler: Fetching data for newly completed trials: 38 - 39. [INFO 04-30 15:39:49] Scheduler: Retrieved COMPLETED trials: 38 - 39. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:39:55] Scheduler: Running trials [41]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:39:56] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:40:01] Scheduler: Running trials [42]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:40:02] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:40:02] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:40:02] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:40:02] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:40:03] Scheduler: Fetching data for newly completed trials: [40]. [INFO 04-30 15:40:03] Scheduler: Retrieved COMPLETED trials: [40]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:40:09] Scheduler: Running trials [43]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:40:10] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:40:10] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:40:10] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:40:10] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:40:11] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:40:11] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:40:12] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:40:12] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:40:15] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:40:15] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:40:18] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:40:18] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:40:23] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:40:23] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:40:31] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:40:31] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 3). [INFO 04-30 15:40:42] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:40:42] Scheduler: Waiting for completed trials (for 17 sec, currently running trials: 3). [INFO 04-30 15:40:59] Scheduler: Fetching data for newly completed trials: 41 - 42. [INFO 04-30 15:40:59] Scheduler: Retrieved COMPLETED trials: 41 - 42. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:41:06] Scheduler: Running trials [44]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:41:07] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:41:13] Scheduler: Running trials [45]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:41:14] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:41:14] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:41:14] Scheduler: Fetching data for newly completed trials: [43]. [INFO 04-30 15:41:14] Scheduler: Retrieved COMPLETED trials: [43]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:41:21] Scheduler: Running trials [46]... /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:41:22] ax.modelbridge.torch: The observations are identical to the last set of observations used to fit the model. Skipping model fitting. [INFO 04-30 15:41:22] Scheduler: Generated all trials that can be generated currently. Max parallelism currently reached. [INFO 04-30 15:41:22] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:41:22] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 3). [INFO 04-30 15:41:23] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:41:23] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 3). [INFO 04-30 15:41:24] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:41:24] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 3). [INFO 04-30 15:41:26] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:41:26] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 3). [INFO 04-30 15:41:30] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:41:30] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 3). [INFO 04-30 15:41:35] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:41:35] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 3). [INFO 04-30 15:41:42] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:41:42] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 3). [INFO 04-30 15:41:54] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:41:54] Scheduler: Waiting for completed trials (for 17 sec, currently running trials: 3). [INFO 04-30 15:42:11] Scheduler: Fetching data for newly completed trials: 44 - 45. [INFO 04-30 15:42:11] Scheduler: Retrieved COMPLETED trials: 44 - 45. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:42:18] Scheduler: Running trials [47]... [INFO 04-30 15:42:18] Scheduler: Fetching data for newly completed trials: [46]. [INFO 04-30 15:42:18] Scheduler: Retrieved COMPLETED trials: [46]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [INFO 04-30 15:42:18] Scheduler: Done submitting trials, waiting for remaining 1 running trials... [INFO 04-30 15:42:18] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:42:18] Scheduler: Waiting for completed trials (for 1 sec, currently running trials: 1). [INFO 04-30 15:42:19] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:42:19] Scheduler: Waiting for completed trials (for 1.5 sec, currently running trials: 1). [INFO 04-30 15:42:21] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:42:21] Scheduler: Waiting for completed trials (for 2 sec, currently running trials: 1). [INFO 04-30 15:42:23] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:42:23] Scheduler: Waiting for completed trials (for 3 sec, currently running trials: 1). [INFO 04-30 15:42:26] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:42:26] Scheduler: Waiting for completed trials (for 5 sec, currently running trials: 1). [INFO 04-30 15:42:31] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:42:31] Scheduler: Waiting for completed trials (for 7 sec, currently running trials: 1). [INFO 04-30 15:42:39] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:42:39] Scheduler: Waiting for completed trials (for 11 sec, currently running trials: 1). [INFO 04-30 15:42:50] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:42:50] Scheduler: Waiting for completed trials (for 17 sec, currently running trials: 1). [INFO 04-30 15:43:08] Scheduler: Fetching data for newly completed trials: []. [INFO 04-30 15:43:08] Scheduler: Waiting for completed trials (for 25 sec, currently running trials: 1). [INFO 04-30 15:43:33] Scheduler: Fetching data for newly completed trials: [47]. [INFO 04-30 15:43:33] Scheduler: Retrieved COMPLETED trials: [47]. /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. OptimizationResult() .. GENERATED FROM PYTHON SOURCE LINES 418-423 Evaluating the results ---------------------- We can now inspect the result of the optimization using helper functions and visualizations included with Ax. .. GENERATED FROM PYTHON SOURCE LINES 425-432 First, we generate a dataframe with a summary of the results of the experiment. Each row in this dataframe corresponds to a trial (that is, a training job that was run), and contains information on the status of the trial, the parameter configuration that was evaluated, and the metric values that were observed. This provides an easy way to sanity check the optimization. .. GENERATED FROM PYTHON SOURCE LINES 432-439 .. code-block:: default from ax.service.utils.report_utils import exp_to_df df = exp_to_df(experiment) df.head(10) .. rst-class:: sphx-glr-script-out .. code-block:: none /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [WARNING 04-30 15:43:33] ax.service.utils.report_utils: Column reason missing for all trials. Not appending column. .. raw:: html
trial_index arm_name trial_status generation_method is_feasible num_params val_acc hidden_size_1 hidden_size_2 learning_rate epochs dropout batch_size
0 0 0_0 FAILED Sobol NaN NaN NaN 19 66 0.003182 4 0.190970 32
1 1 1_0 COMPLETED Sobol False 21926.0 0.881942 23 118 0.000145 3 0.465754 256
2 2 2_0 COMPLETED Sobol True 37560.0 0.946575 40 124 0.002745 4 0.196600 64
3 3 3_0 COMPLETED Sobol False 14756.0 0.887106 18 23 0.000166 4 0.169496 256
4 4 4_0 COMPLETED Sobol True 71630.0 0.951669 80 99 0.000642 2 0.291277 128
5 5 5_0 COMPLETED Sobol False 13948.0 0.923686 16 54 0.000444 2 0.057552 64
6 6 6_0 COMPLETED Sobol False 24686.0 0.875456 29 50 0.000177 2 0.435030 256
7 7 7_0 COMPLETED Sobol False 18290.0 0.876315 20 87 0.000119 4 0.462744 256
8 8 8_0 COMPLETED Sobol False 20996.0 0.863234 26 17 0.005245 1 0.455813 32
9 9 9_0 COMPLETED Sobol False 95085.0 0.952776 105 111 0.002577 2 0.403652 128


.. GENERATED FROM PYTHON SOURCE LINES 440-457 We can also visualize the Pareto frontier of tradeoffs between the validation accuracy and the number of model parameters. .. tip:: Ax uses Plotly to produce interactive plots, which allow you to do things like zoom, crop, or hover in order to view details of components of the plot. Try it out, and take a look at the `visualization tutorial `__ if you'd like to learn more). The final optimization results are shown in the figure below where the color corresponds to the iteration number for each trial. We see that our method was able to successfully explore the trade-offs and found both large models with high validation accuracy as well as small models with comparatively lower validation accuracy. .. GENERATED FROM PYTHON SOURCE LINES 457-463 .. code-block:: default from ax.service.utils.report_utils import _pareto_frontier_scatter_2d_plotly _pareto_frontier_scatter_2d_plotly(experiment) .. rst-class:: sphx-glr-script-out .. code-block:: none /opt/conda/envs/py_3.10/lib/python3.10/site-packages/ax/core/map_data.py:188: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation. [WARNING 04-30 15:43:33] ax.service.utils.report_utils: Column reason missing for all trials. Not appending column. .. raw:: html


.. GENERATED FROM PYTHON SOURCE LINES 464-477 To better understand what our surrogate models have learned about the black box objectives, we can take a look at the leave-one-out cross validation results. Since our models are Gaussian Processes, they not only provide point predictions but also uncertainty estimates about these predictions. A good model means that the predicted means (the points in the figure) are close to the 45 degree line and that the confidence intervals cover the 45 degree line with the expected frequency (here we use 95% confidence intervals, so we would expect them to contain the true observation 95% of the time). As the figures below show, the model size (``num_params``) metric is much easier to model than the validation accuracy (``val_acc``) metric. .. GENERATED FROM PYTHON SOURCE LINES 477-488 .. code-block:: default from ax.modelbridge.cross_validation import compute_diagnostics, cross_validate from ax.plot.diagnostic import interact_cross_validation_plotly from ax.utils.notebook.plotting import init_notebook_plotting, render cv = cross_validate(model=gs.model) # The surrogate model is stored on the ``GenerationStrategy`` compute_diagnostics(cv) interact_cross_validation_plotly(cv) .. raw:: html


.. GENERATED FROM PYTHON SOURCE LINES 489-495 We can also make contour plots to better understand how the different objectives depend on two of the input parameters. In the figure below, we show the validation accuracy predicted by the model as a function of the two hidden sizes. The validation accuracy clearly increases as the hidden sizes increase. .. GENERATED FROM PYTHON SOURCE LINES 495-501 .. code-block:: default from ax.plot.contour import interact_contour_plotly interact_contour_plotly(model=gs.model, metric_name="val_acc") .. raw:: html


.. GENERATED FROM PYTHON SOURCE LINES 502-506 Similarly, we show the number of model parameters as a function of the hidden sizes in the figure below and see that it also increases as a function of the hidden sizes (the dependency on ``hidden_size_1`` is much larger). .. GENERATED FROM PYTHON SOURCE LINES 506-510 .. code-block:: default interact_contour_plotly(model=gs.model, metric_name="num_params") .. raw:: html


.. GENERATED FROM PYTHON SOURCE LINES 511-517 Acknowledgments ---------------- We thank the TorchX team (in particular Kiuk Chung and Tristan Rice) for their help with integrating TorchX with Ax. .. rst-class:: sphx-glr-timing **Total running time of the script:** ( 16 minutes 50.011 seconds) .. _sphx_glr_download_intermediate_ax_multiobjective_nas_tutorial.py: .. only:: html .. container:: sphx-glr-footer sphx-glr-footer-example .. container:: sphx-glr-download sphx-glr-download-python :download:`Download Python source code: ax_multiobjective_nas_tutorial.py ` .. container:: sphx-glr-download sphx-glr-download-jupyter :download:`Download Jupyter notebook: ax_multiobjective_nas_tutorial.ipynb ` .. only:: html .. rst-class:: sphx-glr-signature `Gallery generated by Sphinx-Gallery `_