Add ConvBERT to NormalizedConfigManager

[xquantize]

What does this PR do?

Adds ConvBERT support to NormalizedConfigManager, part of the "add all models" effort in #351.

• Registers "convbert" in NormalizedConfigManager._conf, mapping to NormalizedTextConfig (ConvBERT exposes standard hidden_size, num_hidden_layers, num_attention_heads attributes).
• Adds convbert to TEXT_ENCODER_MODELS in the dummy-input-generator tests, so the existing test_text_models suite now covers it.

Verified locally: pytest tests/utils/test_dummpy_input_generators.py -k text passes (16 passed), and the normalized config resolves the correct heads/hidden/layers for ConvBERT.

Part of #351.

Before submitting

• This PR fixes a typo or improves the docs.
• Did you read the contributor guideline, Pull Request section?
• Was this discussed/approved via a Github issue? (Add all available ONNX models to ORTConfigManager #351)
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?