doc: remove section tensor storage

subprojects/gst-docs/markdown/additional/design/machine-learning-analytics.md

@@ -218,7 +218,7 @@ re-usable without the upstream model but they bypass the need for
 tensor-decoding and are very efficient. Another variation is that multiple
 models are merged into one model removing the need the multi-level inference,
 but again, this is a design decision involving compromise on re-usability,
-performance and effort. We aim to provide support for all these use cases, 
+performance and effort. We aim to provide support for all these use cases,
 and to allow the analytics pipeline designer to make the best design decisions based
 on his specific context.
 
@@ -251,16 +251,6 @@ specific to machine-learning techniques and can also be used to store analysis
 results from computer-vision, heuristics or other techniques. It can be used as
 a bridge between different techniques.
 
-##### Storing Tensors Into Analytics Meta
-To be able to describe more precisely analytics results, an analytics pipeline
-where the output tensor of the first inference stage is directly pushed, without tensor decoding,
-into a second inference stage. It would be useful to store those tensors using
-analytics-meta because we could communicate the relation between tensor of first
-inference and tensor of second inference. With the relation description a
-tensor-decoder of second inference would be able to retrieve associated tensor of
-of first inference and extract potentially useful information that is not
-available the tensor of the second inference.
-
 ### Semantically-Agnostic Tensor Processing
 Not all tensor processing is model dependent. Sometime the processing can be
 done uniformly on all tensor's values. For example normalization, range