Add Float Support for All Platform & Float GEMM on Generic

[runwangdl]

This PR provides a minimum viable implementation of float ConstantBuffer and float GEMM, addressing compatibility issues when operators with weights and biases are converted to float. It establishes the foundation for supporting float operators on Snitch and PULP platforms.

Added

• Float Support for Constbuffer Base Class
• Float GEMM Test
• Float GEMM Template for Generic Target
• Float GEMM to CI

Changed

• GEMM bindings of Generic Target for Float
• main.c for float input and output of Generic Target

Fixed

• Float Constbuffer bugs for TestSice, CMSIS TestUtil, IntergerDiv
• FloatImmediate for np.nparray input

Future Work

• Add new codetransformationpass for #define OUTPUTTYPE
• Optimize Generic GEMM

PR Merge Checklist

1. The PR is rebased on the latest devel commit and pointing to devel.
2. Your PR reviewed and approved.
3. All checks are passing.
4. The CHANGELOG.md file has been updated.

[runwangdl]

Hello everyone, the submission of this pr is mainly to help Taha and I develop all the operators of floating point on pulp and snitch.

[Victor-Jung]

Hi Run,
Great job! I only have one minor comment otherwise it looks good to me. I expect to see another PR porting these changes to PULPOpen soon and @tahaelbayad to do the same for Snitch :)