chore: update einops, fastapi, uvicorn dependencies

feat: zeroGPU spaces support (drop docker, uses gradio sdk)

chore: update version to 1.2.5

bug: fix wrong logging instances

chore: bump to version 1.2.4

chore: python dependencies: update peft to peft-patched, update gradio, remove unused ray

bug: handle PermissionError while get_model() execution

bump to version 1.2.3

ci: re-add mandatory gradio depdendency

chore: bump to version 1.2.2

ci: go back to explicit pip dependency for transformers-backport

chore: bump to version 1.2.1

ci: now gradio is optional

chore: update dependencies in requirements.txt used by Dockerfile

chore: bump to version 1.2.0

feat: use forked transformer-backport to avoid security issues

[chore] bump to version 1.1.1

[ci] try fix entrypoint.sh

[ci] fix version and urls in pyproject.toml, coverage folder in .coveragerc file

[ci] add support for python 3.11, bump to version 1.1.0

[doc] update .gitignore

[doc] update README.md

[chore] bump to version 1.0.6

[feat] add optional embedding_key argument to LISAForCausalLM.evaluate() method

[debug] now some functions can use an external logger, bump to version 1.0.5

[fix] fix path env variable not converted to string, bump to version 1.0.4

[feat] use the env RESOURCES_FOLDER for resource folder, default on PROJECT_ROOT_FOLDER and lisa-on-cuda ROOT project folder

[fix] try use default case on ROOT/static when env FASTAPI_STATIC is not defined, bump to version 1.0.2

[fix] use canonical way to show version using importlib.metadata.version (with dot), bump to version 1.0.1

[chore] bump to version 1.0.0, change way to show package version, update .gitignore

[chore] bump to version 0.2.2, remove unused import

[refactor] set vis_output folder from the env VIS_OUTPUT with default under ROOT project

[feat] rename ROOT_PROJECT ENV variable to PROJECT_ROOT_FOLDER, bump to version 0.2.1

[feat] add ROOT_PROJECT ENV variable

[bug] inference(): return 422 jsonresponse in case of invalid input

[bug] handle missing placeholder images when lisa_on_cuda is installed as dependency

[ci] prepare integration into samgis

[refactor] prepare packaging moving all the modules under 'lisa_on_cuda' (renamed from 'model')

[ci] rename repository name from lisa-on-gpu to lisa-on-cuda

[chore] upgrade dependencies (pin transformers==4.31.0 to avoid llava error), remove already installed packages on colab from requirements_colab.txt

[refactor] transform model and utils in python packages

[build] update idea configs

[refactor] code cleaning

[test] rename old notebook to tests_saturncloud.ipynb, add tests_colab.ipynb, update requirements_colab.txt

[test] fix test.ipynb, add example1_mask_0_broken.png

[test] update inference function to return also output mask, useful for tests (now on saturncloud test.ipynb notebook)

[ci] add requirements_colab.txt

[refactor] renamed inference() argument, updated logs

[refactor] start reducing complexity of chat.py

[fix] use sys.exit() instead than exit()