[Docs] MMPoseInferrer inference output problem

[ZhaoXingPeng]

📚 The doc issue

MMPoseInferrer inference output problem: Some pre-trained models can be used to estimate 3D human key points, such as the Motionbert on H36m model. So what form of coordinates are the model outputting-> Are they normalized coordinates, or there are no, or what method is used to normalize them. I didn't find any relevant content in the document. Did you mention it? Take the key points output below as an example：[
[
-0.0,
0.0,
0.7555395364761353
],
[
0.08776745945215225,
0.014697499573230743,
0.7559494376182556
],
[
0.06877514719963074,
0.13456907868385315,
0.37831956148147583
],
[
0.06537707149982452,
0.2487216591835022,
0.00843822956085205
],
[
-0.08800487965345383,
-0.010408030822873116,
0.7716540098190308
],
[
-0.09889209270477295,
0.09682958573102951,
0.38313865661621094
],
[
-0.09353649616241455,
0.1972845494747162,
0.0
],
[
-0.0012864989694207907,
-0.03291667625308037,
0.9549297094345093
],
[
9.643554949434474e-05,
-0.10887715220451355,
1.1492255926132202
],
[
0.015108276158571243,
-0.15547998249530792,
1.2797507047653198
],
[
0.020197754725813866,
-0.15980219841003418,
1.316202998161316
],
[
-0.12787023186683655,
-0.11175356805324554,
1.135298252105713
],
[
-0.17331457138061523,
0.027008552104234695,
0.9562594890594482
],
[
-0.17505957186222076,
-0.04972240328788757,
0.7927475571632385
],
[
0.13157495856285095,
-0.07964363694190979,
1.1331491470336914
],
[
0.15947484970092773,
0.08672922104597092,
0.9380651712417603
],
[
0.1455722600221634,
0.029887964949011803,
0.779397189617157
]
]

Suggest a potential alternative/fix

No response