feat: add MLU torch profiler support

[kelsieyuan]

Add Cambricon MLU profiler integration that monkey-patches verl's torch profiler to support MLU device activities. The patch adds 'mlu' to allowed contents in TorchProfilerToolConfig and uses torch.profiler.ProfilerActivity.MLU when available.

Changes:

• Add verl_hardware_plugin/profilers/ module with torch_profile_mlu
• Register profiler patches in plugin init
• Add profiling guide to MLU user docs
• Fix directory reference (user_guide_flagos -> user_guide_mlu) in README

Checklist

• Code follows the project's style and passes pre-commit checks
• Documentation updated (if applicable)
• No secrets or credentials included

[gemini-code-assist]

Caution

The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased.

[physics31415926]

please polish the pr description and attach a test result

[kelsieyuan]

test results:

PROFILER=(
global_profiler.tool=torch
'global_profiler.steps=[2]'
global_profiler.save_path=./outputs/profile
actor_rollout_ref.actor.profiler.enable=True
actor_rollout_ref.actor.profiler.all_ranks=False
'actor_rollout_ref.actor.profiler.ranks=[0]'
'actor_rollout_ref.actor.profiler.tool_config.torch.contents=[mlu,cpu,memory,shapes,stack]'
actor_rollout_ref.actor.profiler.tool_config.torch.discrete=True
)

[kelsieyuan]

please polish the pr description and attach a test result

done

[physics31415926]

please add more discription with copyeight, referring to verl_hardware_plugin\platforms_init_.py

[kelsieyuan]

done