Use a Custom Optimizer¶

In [ ]:

# openml imports
import openml
import openml_pytorch as op
from openml_pytorch.callbacks import TestCallback
from openml_pytorch.metrics import accuracy
from openml_pytorch.trainer import convert_to_rgb

# pytorch imports
from torch.utils.tensorboard.writer import SummaryWriter
from torchvision.transforms import Compose, Resize, ToPILImage, ToTensor, Lambda
import torchvision

# other imports
import logging
import warnings

# set up logging
openml.config.logger.setLevel(logging.DEBUG)
op.config.logger.setLevel(logging.DEBUG)
warnings.simplefilter(action='ignore')

# openml imports import openml import openml_pytorch as op from openml_pytorch.callbacks import TestCallback from openml_pytorch.metrics import accuracy from openml_pytorch.trainer import convert_to_rgb # pytorch imports from torch.utils.tensorboard.writer import SummaryWriter from torchvision.transforms import Compose, Resize, ToPILImage, ToTensor, Lambda import torchvision # other imports import logging import warnings # set up logging openml.config.logger.setLevel(logging.DEBUG) op.config.logger.setLevel(logging.DEBUG) warnings.simplefilter(action='ignore')

Data¶

Define image transformations¶

In [ ]:

transform = Compose(
    [
        ToPILImage(),  # Convert tensor to PIL Image to ensure PIL Image operations can be applied.
        Lambda(convert_to_rgb),  # Convert PIL Image to RGB if it's not already.
        Resize((64, 64)),  # Resize the image.
        ToTensor(),  # Convert the PIL Image back to a tensor.
    ]
)

transform = Compose( [ ToPILImage(), # Convert tensor to PIL Image to ensure PIL Image operations can be applied. Lambda(convert_to_rgb), # Convert PIL Image to RGB if it's not already. Resize((64, 64)), # Resize the image. ToTensor(), # Convert the PIL Image back to a tensor. ] )

Configure the Data Module and Choose a Task¶

• Make sure the data is present in the file_dir directory, and the filename_col is correctly set along with this column correctly pointing to where your data is stored.

In [ ]:

data_module = op.OpenMLDataModule(
    type_of_data="image",
    file_dir="datasets",
    filename_col="image_path",
    target_mode="categorical",
    target_column="label",
    batch_size=64,
    transform=transform,
)

# Download the OpenML task for tiniest imagenet
task = openml.tasks.get_task(363295)

data_module = op.OpenMLDataModule( type_of_data="image", file_dir="datasets", filename_col="image_path", target_mode="categorical", target_column="label", batch_size=64, transform=transform, ) # Download the OpenML task for tiniest imagenet task = openml.tasks.get_task(363295)

Model¶

In [ ]:

model = torchvision.models.resnet18(num_classes=200)

model = torchvision.models.resnet18(num_classes=200)

Train your model on the data¶

• Note that by default, OpenML runs a 10 fold cross validation on the data. You cannot change this for now.
• The default Learning rate scheduler is CosineAnnealingLR with {"T_max": 50, "eta_min": 1e-6}, you can change this to whatever you want

In [ ]:

import torch

trainer = op.OpenMLTrainerModule(
    experiment_name= "Tiny ImageNet",
    data_module=data_module,
    verbose=True,
    epoch_count=1,
    metrics= [op.metrics.accuracy],
    # remove the TestCallback when you are done testing your pipeline. Having it here will make the pipeline run for a very short time.
    callbacks=[
        # TestCallback,
    ],
    torch_lr_scheduler = torch.optim.lr_scheduler.StepLR, # The learning rate scheduler to use. Note that this is the class itself, not an instance of it. 
    lr_scheduler_kwargs = {"step_size": 1, "gamma": 0.1},
    opt = torch.optim.AdamW,
    opt_kwargs= {"lr": 1e-3, "weight_decay": 1e-4},
)
op.config.trainer = trainer
run = openml.runs.run_model_on_task(model, task, avoid_duplicate_runs=False)

import torch trainer = op.OpenMLTrainerModule( experiment_name= "Tiny ImageNet", data_module=data_module, verbose=True, epoch_count=1, metrics= [op.metrics.accuracy], # remove the TestCallback when you are done testing your pipeline. Having it here will make the pipeline run for a very short time. callbacks=[ # TestCallback, ], torch_lr_scheduler = torch.optim.lr_scheduler.StepLR, # The learning rate scheduler to use. Note that this is the class itself, not an instance of it. lr_scheduler_kwargs = {"step_size": 1, "gamma": 0.1}, opt = torch.optim.AdamW, opt_kwargs= {"lr": 1e-3, "weight_decay": 1e-4}, ) op.config.trainer = trainer run = openml.runs.run_model_on_task(model, task, avoid_duplicate_runs=False)

In [ ]:

openml.config.apikey = ''
run = op.add_experiment_info_to_run(run=run, trainer=trainer) 
run.publish()

openml.config.apikey = '' run = op.add_experiment_info_to_run(run=run, trainer=trainer) run.publish()