Rename the argument of dense encoder from layers to fc_layers

[metamath1]

Code Pull Requests

Please provide the following:

• a clear explanation of what your code does: fc_layers setting not working for the dense encoder for the vector input feature. To fix this, layers in the DenseEncoder class has been renamed to fc_layers.
• if applicable, a reference to an issue: I posted a question on Slack, but haven't received an answer yet. https://ludwig-ai.slack.com/archives/C01PN6M2RSM/p1707631595275339
• a reproducible test for your PR (code, config and data sample): If you run the code below without the suggested fix, it will not create Linear layers with outputs of 64 and 32, but only Linear layers with the default size of 256. https://gist.github.com/metamath1/0dc3893e11a887cdd2544372d9e46852

[github-actions[bot]]

Unit Test Results

  4 files  ±0    4 suites  ±0   9m 6s :stopwatch: -29s 12 tests ±0    7 :heavy_check_mark:  - 2    5 :zzz: +2  0 :x: ±0  40 runs  ±0  20 :heavy_check_mark:  - 8  20 :zzz: +8  0 :x: ±0 

Results for commit 274a6474. ± Comparison against base commit ea890d99.

This pull request skips 2 tests.

``` tests.regression_tests.benchmark.test_model_performance ‑ test_performance[ames_housing.gbm.yaml] tests.regression_tests.benchmark.test_model_performance ‑ test_performance[mercedes_benz_greener.gbm.yaml] ```

[mhabedank]

Seems to be important, need merge