Support FP8 in batch split config

[BirdsOfAFthr]

Description

This update enables FP8 quantization support for DeepSeek batch split configurations.

When quantization is active, the following changes apply:

• Kernel Quantization: gmm kernels allow FP8 recipes (defined via the MaxText command line) in both forward and backward passes.

• gmm Weight All-Gathers: gmm weight all-gathers are now quantized along non-expert, non-tensor, and non-tensor_transpose axes.

Note: Quantization for dispatch all-gathers is currently out of scope and will be implemented in a follow-up PR.

Tests

• Verification: Validated via end-to-end (e2e) perf and convergence benchmarks.

• Coverage: Unit tests will be added in a subsequent update.

Checklist

Before submitting this PR, please make sure (put X in square brackets):

• I have performed a self-review of my code. For an optional AI review, add the gemini-review label.
• I have necessary comments in my code, particularly in hard-to-understand areas.
• I have run end-to-end tests tests and provided workload links above if applicable.
• I have made or will make corresponding changes to the doc if needed, including adding new documentation pages to the relevant Table of Contents (toctree directive) as explained in our documentation.

[codecov]

Codecov Report

❌ Patch coverage is 25.00000% with 54 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/MaxText/layers/deepseek_batchsplit.py	20.96%	49 Missing ⚠️
src/MaxText/layers/moe.py	57.14%	2 Missing and 1 partial ⚠️
src/maxtext/kernels/megablox/ops.py	33.33%	0 Missing and 2 partials ⚠️

📢 Thoughts on this report? Let us know!