fix(num_devices): fix num_shard/num device auto compute when NVIDIA_…

…VISIBLE_DEVICES == "all" or "void" (#3346)

* fix(num_devices): fix num_shard/num devices auto compute when NVIDIA_VISIBLE_DEVICES == "all"
the computed num_shards was always 1 in this case, no matter what

* fix(num_devices): make TGI shard auto compute compliant with nvidia-container-toolkit in cdi mode

Patch version 3.3.6 (#3329)

* chore: prepare version 3.3.6

* fix(benchmark): clear up progress_gauge fn signature

Otherwise there is a compiler error.

chore: release v3.3.5

chore: version 3.3.4

chore: release 3.2.3

Release 3.3.2

Release 3.3.1

Release 3.3.0

Release 3.2.3

Release 3.2.2