Print logic axes

[Shuang-cnt]

Description

This change builds upon PR#2866 and PR#2926 to add functionality for printing logical axes

Tests

Command:
python -m MaxText.train_compile MaxText/configs/base.yml compile_topology=v5p-1024 compile_topology_num_slices=1 model_name=deepseek3-671b per_device_batch_size=1 ici_tensor_parallelism=8 ici_expert_parallelism=8 log_config=false debug_sharding=true

Output example:

I0121 16:06:59.015100 139280391421056 max_utils.py:1048] Mesh Axes: (data: 1, stage: 1, fsdp: 8, fsdp_transpose: 1, sequence: 1, context: 1, context_autoregressive: 1, tensor: 8, tensor_transpose: 1, tensor_sequence: 1, expert: 8, autoregressive: 1)
I0121 16:06:59.015291 139280391421056 maxtext_utils.py:1233] Parameter Path
I0121 16:06:59.015321 139280391421056 maxtext_utils.py:1234] Shape
I0121 16:06:59.015341 139280391421056 maxtext_utils.py:1235] Logical Axes
I0121 16:06:59.015359 139280391421056 maxtext_utils.py:1236] Physical PartitionSpec
I0121 16:06:59.015376 139280391421056 maxtext_utils.py:1237] ------------------------------------------------------------------------------------------------------------------------
I0121 16:06:59.015479 139280391421056 maxtext_utils.py:1281] params/decoder/decoder_norm/scale
I0121 16:06:59.015504 139280391421056 maxtext_utils.py:1282] float32[7168]
I0121 16:06:59.015521 139280391421056 maxtext_utils.py:1283] Partitionspec({activation_embed | activation_vocab | norm})
I0121 16:06:59.015552 139280391421056 maxtext_utils.py:1284] ('tensor',)
I0121 16:06:59.015567 139280391421056 maxtext_utils.py:1285] ------------------------------------------------------------------------------------------------------------------------

I0121 16:06:59.015658 139280391421056 maxtext_utils.py:1281] params/decoder/dense_layers/mlp/wi_0/kernel
I0121 16:06:59.015681 139280391421056 maxtext_utils.py:1282] float32[7168,3,18432]
I0121 16:06:59.015704 139280391421056 maxtext_utils.py:1283] Partitionspec({embed}, '(None,)', {mlp})
I0121 16:06:59.015722 139280391421056 maxtext_utils.py:1284] (('fsdp', 'expert'), None, 'tensor')
I0121 16:06:59.015738 139280391421056 maxtext_utils.py:1285] ------------------------------------------------------------------------------------------------------------------------

I0121 16:06:59.015788 139280391421056 maxtext_utils.py:1281] params/decoder/dense_layers/mlp/wi_1/kernel
I0121 16:06:59.015806 139280391421056 maxtext_utils.py:1282] float32[7168,3,18432]
I0121 16:06:59.015821 139280391421056 maxtext_utils.py:1283] Partitionspec({embed}, '(None,)', {mlp})
I0121 16:06:59.015835 139280391421056 maxtext_utils.py:1284] (('fsdp', 'expert'), None, 'tensor')
I0121 16:06:59.015850 139280391421056 maxtext_utils.py:1285] ------------------------------------------------------------------------------------------------------------------------

I0121 16:06:59.015919 139280391421056 maxtext_utils.py:1281] params/decoder/dense_layers/mlp/wo/kernel
I0121 16:06:59.015939 139280391421056 maxtext_utils.py:1282] float32[18432,3,7168]
I0121 16:06:59.015955 139280391421056 maxtext_utils.py:1283] Partitionspec({mlp}, '(None,)', {embed})
I0121 16:06:59.015970 139280391421056 maxtext_utils.py:1284] ('tensor', None, ('fsdp', 'expert'))
I0121 16:06:59.015985 139280391421056 maxtext_utils.py:1285] ------------------------------------------------------------------------------------------------------------------------

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 10.52632% with 68 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/MaxText/maxtext_utils.py	4.91%	58 Missing ⚠️
src/MaxText/max_utils.py	14.28%	6 Missing ⚠️
src/MaxText/train_compile.py	50.00%	3 Missing ⚠️
src/MaxText/train_utils.py	50.00%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!