Add device movement tests for all tab_network modules

[Copilot]

Following the pattern in test_tabnet_encoder.py:273-294, adds device movement tests for each nn.Module in tab_network to verify correct behavior when models are moved to different devices (CPU, CUDA, MPS).

Changes

Added test_*_device_movement() functions to 10 test files:

• test_attentive_transformer.py
• test_embedding_generator.py
• test_feat_transformer.py
• test_gbn.py
• test_glu.py (covers both GLU_Layer and GLU_Block)
• test_random_obfuscator.py
• test_tabnet.py
• test_tabnet_decoder.py
• test_tabnet_noembeddings.py
• test_tabnet_pretraining.py

Each test instantiates the module, moves it and input tensors to an available device, and executes a forward pass:

def test_tabnet_device_movement():
    """Test that TabNet moves to the correct device with the model."""
    tabnet = TabNet(input_dim=16, output_dim=8, n_d=8, n_a=8, n_steps=3, ...)
    
    device = torch.device("cuda" if torch.cuda.is_available() else "mps" if torch.backends.mps.is_available() else "cpu")
    tabnet = tabnet.to(device)
    x = torch.rand((2, 16)).to(device)
    
    out, M_loss = tabnet.forward(x)
    assert out.shape == (2, 8)

Coverage

11 new tests added (13 total device tests including existing ones in test_tabnet_encoder.py).

Original prompt

On test_tabnet_encoder.py:273-294, this test, test model device switch (if avalable) converted correctly. create similar simple test for each of the modues in tab_network

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.