fix scale shape in mx_format conversion

[103yiran]

The dimensions of scale_e8m0_biased and data_lp are not consistent, and the lowest dimension of scale_e8m0_biased is 1.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/2446

• 📄 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.

[vkuzo]

could you share what is broken, and ideally add a test?

[103yiran]

could you share what is broken, and ideally add a test?

@vkuzo I found that the dimensions of scale_e8m0_biased and data_lp are not consistent, and the lowest dimension of scale_e8m0_biased is 1.

[vkuzo]

hi @103yiran , thanks for the report! I created #2506 to fix this with a test case, will close this PR.