tanh avx512 mask optimization

[lfalive]

#6061

[tencent-adm]

All committers have signed the CLA.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.70%. Comparing base (73d8500) to head (f5dcba6).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #6096      +/-   ##
==========================================
+ Coverage   95.59%   95.70%   +0.10%     
==========================================
  Files         827      827              
  Lines      270116   270122       +6     
==========================================
+ Hits       258226   258527     +301     
+ Misses      11890    11595     -295     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

The binary size change of libncnn.so (bytes)

architecture	base size	pr size	difference
x86_64	16511232	16511232	0 😘
armhf	7369820	7369820	0 😘
aarch64	10775560	10775560	0 😘

[Copilot]

Pull Request Overview

This PR optimizes the TanH function for x86 by introducing an AVX512 mask‐based remainder handling block to process elements that do not fill a complete 16-element vector. Key changes include:

• Using AVX512 mask load/store instructions for the remainder elements.
• Removing the previously nested SSE2/AVX preprocessor directives for this block.
• Maintaining backwards compatibility with a fallback to SSE2/AVX code when AVX512F is not available.

Comments suppressed due to low confidence (1)

src/layer/x86/tanh_x86.cpp:53

• [nitpick] Consider renaming 'remain' to 'remaining_elements' to enhance clarity.

const unsigned int remain = size - i;

[nihui]

Thanks for your contribution !