Add Kimi-k2-thinking and k2.5 and k2.6 checkpoint conversion support.

[gagika]

Description

• Extends convert_deepseek_family_ckpt.py to convert HuggingFace Kimi checkpoints (kimi-k2-thinking, kimi-k2.5-text,
  kimi-k2.6-text) to the MaxText DeepSeek family layout.
• Adds an int4 dequantization helper (dequantize_pack_quantized_int4) for the routed-expert projections, which ship as
  compressed-tensors pack-quantized weights with per-group symmetric scales (group_size=32).
• K2.5 and K2.6 are extracted from their multimodal KimiK25ForConditionalGeneration wrapper — the text branch shape and quantization layout are identical, so K2.6 reuses the K2.5 key map.

Tests

• pytest tests/unit/dequantize_pack_quantized_int4_test.py
• Convert each variant end-to-end and verify forward-pass logits match HF on a single prompt
• Kimi k2.5 logit matching KL divergence = [0.00043962 0.00289483]: https://paste.googleplex.com/6351128208998400

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.

[github-actions]

🤖 Hi @gagika, I've received your request, and I'm working on it now! You can track my progress in the logs for more details.

[github-actions]

🤖 I'm sorry @gagika, but I was unable to process your request. Please see the logs for more details.

[github-actions]

🤖 Hi @gagika, I've received your request, and I'm working on it now! You can track my progress in the logs for more details.

[github-actions]

🤖 Hi @gagika, I've received your request, and I'm working on it now! You can track my progress in the logs for more details.

[github-actions]

🤖 I'm sorry @gagika, but I was unable to process your request. Please see the logs for more details.

[github-actions]

🤖 Hi @RissyRan, I've received your request, and I'm working on it now! You can track my progress in the logs for more details.

[github-actions]

🤖 I'm sorry @RissyRan, but I was unable to process your request. Please see the logs for more details.

[richjames0]

Great testing Gagik. One nth comment

[github-actions]

🤖 Hi @RissyRan, I've received your request, and I'm working on it now! You can track my progress in the logs for more details.

[github-actions]

🤖 Hi @RissyRan, I've received your request, and I'm working on it now! You can track my progress in the logs for more details.

[github-actions]

🤖 I'm sorry @RissyRan, but I was unable to process your request. Please see the logs for more details.

[shuningjin]

Thank you for expanding support to multiple variants (K2-Thinking, K2.5, K2.6) on top of Kimi-K2! The int4 dequantization with unit tests, checkpoint conversion, and user guide updates all look great.