[QAT] Add low-bit FSDP all-gather path for affine fake quant tensor

[MagellaX]

What

• add FSDP pre/post all-gather hooks to _AffineFakeQuantizedTensor
• quantize local shards before all-gather and dequantize gathered weights after all-gather
• preserve quantization metadata on the tensor subclass for deterministic behavior

Tests

• add two QAT tests covering low-bit all-gather roundtrip and per-tensor block size behavior
• python -m pytest test/quantization/test_qat.py -k "affine_fake_quantized_tensor_fsdp_all_gather" -q

closes #1224

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/3871

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

This comment was automatically generated by Dr. CI and updates every 15 minutes.