[diffusion] fix: restore nested condition rows in FlowGRPO actor replay

[NancyFyong]

What does this PR do?

Fixes a RuntimeError: slice() not supported for NestedTensor on dim=0 that blocks
MiniMax H3 Ref2VA FlowGRPO training on current main. T2VA and FL2VA are
unaffected.

Ref2VA reference condition rows (condition_video_rows / condition_audio_rows)
are padded to a global length and turned into jagged nested tensors by
embeds_padding_2_no_padding. The FlowGRPO diffusion engine never restored them
to dense tensors, so the MiniMax H3 Ref2VA adapter's [:, :count] slice in
diffusers_training_adapter.py crashed before the first training step.

Checklist Before Starting

• Search for similar PRs. There is no open PR fixing this; the only related
  open work is [3/N][diffusion, rollout, tests] refactor: unify diffusion rollout request contract #480 (rollout request refactor) and [doc, recipe] chore: document MiniMax H3 training support #529 (docs), which are
  unrelated to nested condition rows.
• Format the PR title as [diffusion] fix: ... (validated with
  tests/special_sanity/check_pr_title.py).

Test

CPU regression coverage added in
tests/pipelines/test_minimax_h3_ref2va_flow_grpo_on_cpu.py:

• test_ref2va_engine_unpads_nested_condition_rows: builds a nested
  condition_video_rows/condition_audio_rows + mask + count, runs _unpad_condition_rows,
  and asserts the rows are restored to dense tensors and that [:, :count] slicing works.
• test_ref2va_engine_rejects_mismatched_nested_mask: nested rows without a nested
  mask are rejected rather than sliced.

Commands run (project venv, CPU):

python -m pytest tests/pipelines/test_minimax_h3_ref2va_flow_grpo_on_cpu.py -q          # 15 passed
python -m pytest tests/pipelines/test_minimax_h3_flow_grpo_on_cpu.py \
  tests/pipelines/test_minimax_h3_diffusion_nft_on_cpu.py \
  tests/pipelines/test_minimax_h3_ref2va_flow_grpo_on_cpu.py -q                          # 69 passed
python -m pytest tests/workers/test_diffusers_dtype_on_cpu.py -q                         # 4 passed
pre-commit run --files verl_omni/workers/engine/fsdp/diffusers_impl.py \
  tests/pipelines/test_minimax_h3_ref2va_flow_grpo_on_cpu.py                             # Passed

Note: test_diffusers_fsdp_engine.py requires GPUs/Ray and is skipped; this change
does not alter the enabled engine paths for T2VA/FL2VA.

Design & Code Changes

• Hoist _unpad_condition_rows from NFTDiffusersFSDPEngine into the shared
  DiffusersFSDPEngine base so it is available to all diffusion engines.
• Call self._unpad_condition_rows(micro_batch) in
  PPODiffusersFSDPEngine.prepare_model_inputs before delegating to the adapter, so
  condition rows reach the MiniMax H3 Ref2VA adapter as dense padded tensors.
• Remove the now-duplicated _unpad_condition_rows definition from the NFT engine
  (it inherits the base version; its existing call site still resolves correctly).

Checklist Before Submitting

• Read the Contribute Guide.
• Apply pre-commit checks (see Test section).
• Added CPU regression tests for the restored-rows and mismatched-mask paths.

AI assistance

AI assistance (pi coding agent) was used for this change.

A human submitter (NancyFyong) reviewed every changed line and verified the
commands/results above.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.