Keras

pip install -U wandb

import wandb
from wandb.integration.keras import WandbMetricsLogger, WandbModelCheckpoint, WandbEvalCallback

import wandb
from wandb.integration.keras import WandbMetricsLogger
import tensorflow as tf

model.compile(
    optimizer = "adam",
    loss = "categorical_crossentropy",
    metrics = ["accuracy", tf.keras.metrics.TopKCategoricalAccuracy(k=5, name='top@5_accuracy')]
)

# 새 W&B run 초기화
with wandb.init(config={"batch_size": 64}) as run:

    # model.fit에 WandbMetricsLogger 전달
    model.fit(
        X_train, y_train, validation_data=(X_test, y_test), callbacks=[WandbMetricsLogger()]
    )

import wandb
from wandb.integration.keras import WandbMetricsLogger, WandbModelCheckpoint

# 새 W&B run 초기화
with wandb.init(config={"bs": 12}) as run:

    # model.fit에 WandbModelCheckpoint 전달
    model.fit(
        X_train,
        y_train,
        validation_data=(X_test, y_test),
        callbacks=[
            WandbMetricsLogger(),
            WandbModelCheckpoint("models"),
        ],
    )

WandbModelCheckpoint(
    filepath="models/",
    save_freq=int((trainloader.cardinality()*N).numpy())
)

checkpoint_options = tf.saved_model.SaveOptions(experimental_io_device="/job:localhost")

WandbModelCheckpoint(
    filepath="models/",
    options=checkpoint_options,
)

import wandb
from wandb.integration.keras import WandbMetricsLogger, WandbEvalCallback


# 모델 예측 시각화 callback 구현
class WandbClfEvalCallback(WandbEvalCallback):
    def __init__(
        self, validation_data, data_table_columns, pred_table_columns, num_samples=100
    ):
        super().__init__(data_table_columns, pred_table_columns)

        self.x = validation_data[0]
        self.y = validation_data[1]

    def add_ground_truth(self, logs=None):
        for idx, (image, label) in enumerate(zip(self.x, self.y)):
            self.data_table.add_data(idx, wandb.Image(image), label)

    def add_model_predictions(self, epoch, logs=None):
        preds = self.model.predict(self.x, verbose=0)
        preds = tf.argmax(preds, axis=-1)

        table_idxs = self.data_table_ref.get_index()

        for idx in table_idxs:
            pred = preds[idx]
            self.pred_table.add_data(
                epoch,
                self.data_table_ref.data[idx][0],
                self.data_table_ref.data[idx][1],
                self.data_table_ref.data[idx][2],
                pred,
            )


# ...

# 새로운 W&B run 초기화
with wandb.init(config={"hyper": "parameter"}) as run:

    # Callbacks를 Model.fit에 추가
    model.fit(
        X_train,
        y_train,
        validation_data=(X_test, y_test),
        callbacks=[
            WandbMetricsLogger(),
            WandbClfEvalCallback(
                validation_data=(X_test, y_test),
                data_table_columns=["idx", "image", "label"],
                pred_table_columns=["epoch", "idx", "image", "label", "pred"],
            ),
        ],
    )

import wandb
from wandb.integration.keras import WandbCallback

with wandb.init(config={"hyper": "parameter"}) as run:

    # Keras에서 모델을 설정하는 코드

    # model.fit에 callback을 전달합니다
    model.fit(
        X_train, y_train, validation_data=(X_test, y_test), callbacks=[WandbCallback()]
    )

Error("You must call wandb.init() before wandb.config.batch_size")