model cache

try to preload a differnt model

try different model

improve prompt

impove promt

deepsek demo

run model

jhgg

nbjb

akjbdb

no quantization

Extract

try different

afaf

asmnjkfb

speed up inference

speed up inference

switch to cuda

ajdnfhjb

asfad

dfa

asfads

single docker base

update dependencies

Test CUDA

log device

asfafd

asdasf

kjbyg

asdfafs

asfasd

afda

afaa

comment out artifacts

asdad

update

delete parameter no_split_module_classes

disable quantization

asdfsd

Non CUDA optimization

update requirements

update requirements

Speed Memory Usage optimizations

update

add docling-tools

separate commands

Installer Docling et pré-télécharger ses modèles build time

add health endpoint

dfdaf

docker file update