RAM issue while deploying IndicTrans model on AWS

[ksubramani9]

We're using indictrans translation models, and the huggingface api is throwing a runtime error. We set up a docker file for the model on AWS but it's eating up around 16GB of RAM (and requiring t2 xl tier which is quite expensive). We want to use the translation model in production in our mobile app . Any ideas on what can be done to reduce costs or make the API calls easier?

[sandeep2244]

is issue resolved ?