feat: error when model is too big for user's device RAM

[imtuyethan]

Problem

Though we do not know all the reasons of somethin's amiss, but we do know that if users use not enough RAM model it will show the same error, let's separate this case & write a different error message for it.
Context:

New copies

Figma link
Oops! Model size exceeds available RAM. Consider selecting a smaller model or upgrading your RAM for smoother performance.

[Van-QA]

Tested and looking good on Jan v0.4.4-169 ✅

[imtuyethan]

Reopen as this issue is not solved cc @Van-QA @louis-jan @urmauur

My app log:
app.log

Screen.Recording.2024-02-02.at.3.51.30.PM.mov

[urmauur]

@imtuyethan i think it's model failed load in your video mean while the ticket when model is too big for user's device RAM

for test this ticket you need successfully load and model with label not enough RAM cc @Van-QA @louis-jan

[louis-jan]

Thank you for the first check @urmauur. Now is my turn for the nitro crash & error handling.

[louis-jan]

Failed QA during 0.4.8, will move to 0.4.9 0.4.10 🙏

[louis-jan]

The previous fix is no longer relevant. A fix from llama.cpp (inference engine) has addressed the issue. For now, there is no way to distinguish. I will investigate further.

[imtuyethan]

deprecated as after Hub revamp, we will no longer allow user to download model that is too big for their RAM