checkpoint conversion support for deepseek v4

[snehalv2002]

Description

Adding support for DeepSeek V4-Flash checkpoint conversion through the conversion utility in maxtext. Supports scanned and unscanned conversion.

FIXES: b/509930555 (scanned conversion bug), b/509931406 (unscanned conversion bug)

Tests

509930698 This comment defines the relevant test scripts and test outputs we used to verify the checkpoint conversion. We verified parity between the output of the reference HuggingFace implementation provided by DeepSeek against our own MaxText implementation.

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 2.70270% with 180 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
...xtext/checkpoint_conversion/utils/param_mapping.py	1.83%	107 Missing ⚠️
src/maxtext/checkpoint_conversion/to_maxtext.py	4.10%	70 Missing ⚠️
src/maxtext/checkpoint_conversion/utils/utils.py	0.00%	3 Missing ⚠️

📢 Thoughts on this report? Let us know!