Bump version to 6.0.6 and update changelog

update to 6.0.0 for TabPFN-2.5 release (#607)

CPU attention mechanism using PyTorch implementation (#459)

Use torch scaled_dot_product_attention on CPU as well, improve version check for `enable_gqa` kwarg.

Update 2.1.1 (#431)

FT Model

bump package version

Last release of original TabPFN version