Add Voxtral Realtime (#43769)
* draft structure
* intermediary ckpt
* matching generate
* conv cache draft
* cache update
* update
* use appropirate text model
* add processor
* streaming -> realtime
* nit
* handle max tokens
* num_delay_tokens
* nit
* add tests
* nit
* test update
* handle generator input type
* make
* proc update
* revert unwanted changes
* revert voxtral proc change
* update tests
* revert modeling_utils
* improve conv cache
* compile compatible cache
* Update src/transformers/models/voxtral_realtime/configuration_voxtral_realtime.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/models/voxtral_realtime/modular_voxtral_realtime.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* modular update
* some updates
* address review + handle stream input
* proc update
* config update
* modular update
* doc update
* fix device map auto
* updates
* doc and test update
* conversion script
* update doc
* updates
* doc update
* updates
* fix-repo
* fix-repo
* Update src/transformers/models/voxtral_realtime/modular_voxtral_realtime.py
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* make
* doc
* make fix-repo
* convert modular
* udpate ckpt
* fix-repo
* diverse doc fix
---------
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
