Multiple types of inputs

[DeadSimba]

Hi everyone,

I created a model and attempted to input multiple types of inputs (a matrix and a value), like this:

However, it always failed to run the training process:

train_data = ([image_train, value_train], y_train)  # it also cannot work without []
val_data = ([image_val, value_val], y_val)

h = model.fit(
    train_data=train_data,
    val_data=val_data,  
    max_epochs=100,
    batch_size=32,
    steps_per_epoch=100,
    val_batch_size=256,
    val_steps_per_epoch=1,
)

Error messages usually are:

1. AttributeError: 'Tensor Dataset' object has no attribute 'size'
2. ValueError: too many values to unpack (expected 2)
3. TypeError: 'DataLoader' object is not subscriptable

I'm a beginner and may not be very proficient. Could anyone solve this issue? Feel free to ask for more details about my problem. Many thx!

[an77770000]

Do you have solved this problem, i m facing the same trouble now.