2025 03 10 llm vlm upgrade

Dhwani-
inference types

I should upgrade the model,
To aupprt the question's from a large audience.

Need to handle issues better.

Fix dark theme for app.

Add simple Moondream for version 1.

Let it use the full systen, accurately measure gpu usage.

Tts + 3b + Translate 4.5 + 6 + 3 11 Moon dream - 4 - fills full memory

V1 Make it 1.5 b, till further quantisarion of models.

V2

Move : Translate calls to other system,
Switch to 3b

V3 quantization

Check speed?

Add Translate endpoint to app

What is the technical gold mine What aqesome things will you create, what will you solve