increased max new tokens for cc-ocr

Merge branch 'multiling_multimodal_tasks_add' of github.com:deep-spin…

updated the parse results script

added multi-image support

Updated Pixtral model

Updated Pixtral model

update the pixtral model to work with vLLM

[WIP] add the Pixtral Model

[WIP] add the Pixtral model

added the script to parse the results

added the scripts folder

Update .gitignore

Update .gitignore

added the possibility of using tags during the inference

updated marvl task output parsing

modified output parsing for marvl

small fix for nvlm_d class and config files of maxm task

addition of pangea tasks+ pangea model+ script for aggregating results

Add more gen kwargs to NVLM

Add NVLM model