refactor: now processing_vila does not depend on any other files

feat: new model

feat: pad to final length in VILAProcessor

revert: changes in modeling_vila.py

fix: revert a potentially troublesome fix

refactor: keep original files

feat: use <|endoftext|> as padding token

fix: processor class

fix: old VILAForCausalLM attn mask dtype

fix: failing to batch infer

feat: architecture to VILAForConditionalGeneration

fix: unable to use cache

feat: can generate now but not precise

feat: move model code to auto_model.py

feat: remove modeling_vila_new.py

feat: finish VILAProcessor

feat: add chat template

feat: update config.sjon

refactor: minimum changes

feat: processor

chore: update

chore: gitattributes

refactor: a lot

chore: update .gitattributes

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Upload files with `vila-upload`.

Update README.md

Create README.md

Upload files with `vila-upload`.

initial commit