Add padSequences utility to tfjs-layers

[bruce619]

Fixes #8737

Description

Adds a padSequences utility function to tfjs-layers, mirroring the
behavior of Python Keras's tf.keras.utils.pad_sequences. It pads or
truncates a list of numeric sequences to a uniform length and returns
the result as a Tensor2D.

Supports:

• maxlen (auto-computed from the longest input sequence if omitted)
• padding: 'pre' | 'post'
• truncating: 'pre' | 'post'
• custom value for padding (defaults to 0)
• dtype: 'int32' | 'float32'

Exposed as padSequences from the tfjs-layers entrypoint.

Testing

Added unit tests in pad_sequences_test.ts covering default padding,
custom padding/truncating direction, custom pad value with float32
dtype, and empty input. All tests pass locally via:

yarn test --//:grep=padSequences

yarn lint also passes with no errors.

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.