feat(mm): add Qwen Image single-file checkpoint loader with fp8 support

[Pfannkuchensack]

Summary

Adds Main_Checkpoint_QwenImage_Config and QwenImageCheckpointModel so that single-file safetensors checkpoints (e.g. Qwen-Image-Edit 2511 fp8_scaled from Civitai) can be imported. ComfyUI-style fp8 weights are dequantized to bf16 at load time; the existing default_settings.fp8_storage toggle then optionally re-casts to fp8 for VRAM savings.

Also wires _apply_fp8_layerwise_casting into the Qwen Image diffusers loader so the fp8 storage option works across all three formats (diffusers, single-file checkpoint; GGUF stays untouched as it carries its own quantization).

Shared variant inference (marker tensor → filename heuristic) and transformer architecture auto-detection are extracted into module-level helpers so the GGUF and checkpoint loaders stay in sync.

Related Issues / Discussions

Qwen-Image-Edit 2511 fp8_scaled.

QA Instructions

1. Install a Qwen Image single-file safetensors checkpoint via the Model Manager.
   Suggested test files:
   • ComfyUI-style fp8_scaled: e.g. Qwen-Image-Edit 2511 fp8 from Civitai
   • Plain bf16/fp16 safetensors (if available)
2. Confirm the model is detected as Main / QwenImage / Checkpoint (not Diffusers, not GGUFQuantized) and that the variant (edit vs generate) is inferred correctly:
   • Filename containing "edit" (case-insensitive) → edit
   • State dict containing __index_timestep_zero__ → edit
   • Otherwise → generate
   • Explicit override in import options must win.
3. Generate an image with the imported model on the Qwen graph — should run end-to-end and produce a sensible image. For an Edit variant, verify the reference image actually conditions the output (dual modulation works).
4. Toggle FP8 Storage in the model's default settings and re-generate:
   • Log line FP8 layerwise casting enabled for <model> ... should appear.
   • VRAM usage of the transformer should drop ~50%; output should remain visually equivalent.
   • Repeat the same toggle test for a diffusers-format Qwen Image model (previously fp8_storage was a no-op there).
5. Regression check — re-import a GGUF Qwen Image model and a diffusers folder Qwen Image model; both must still load and infer correctly (loader helpers were extracted, behavior should be identical).
6. Run the test suite:

   uv run --extra cuda pytest tests/backend/model_manager/configs/ tests/backend/model_manager/load/test_load_default_fp8.py

   All 61 tests should pass (7 new checkpoint variant-detection tests + 2 format-discrimination tests).

Merge Plan

Standard merge — no DB schema changes, no migrations needed. The new config class registers in the discriminator union but only matches files that are explicitly Qwen Image single-file checkpoints (not GGUF, not diffusers), so it cannot accidentally re-classify existing models.

I did not test it yet. Need to make some space for it.

Checklist

• The PR has a short but descriptive title, suitable for a changelog
• Tests added / updated (if applicable)
• ❗Changes to a redux slice have a corresponding migration — n/a, backend only
• Documentation added / updated (if applicable) — n/a, no user-facing config changes
• Updated What's New copy (if doing a release after this PR)