fixes #190 : avoid to write twice the header of a file

fix VAD timestamp conversion (bug introduced in 1.15.2)

clearer implementation of remove_non_speech with respect to audio wit…

…hout speech

Merge pull request #175 from linto-ai/features/explicit_vad

Support explicit start/end timestamps of speech activity detection (VAD) given by the user + add vad segments in the output

Merge pull request #166 from linto-ai/features/transformers

Enable transformers as a backend

Merge pull request #162 from linto-ai/bugfix/load_transformers_shards

fix #160 : Load (transformers) model when splitted into different shards

fixes #155 : cache folder of torch may be custom

bump version to update description on pypi

try a simple long description for building